Sensor-based categorization of upper limb performance in daily life
Jessica Barth, Keith Lohse, Jeffrey Konrad, Marghuertta Bland, Catherine Lang
Washington University in St. Loui, St. Louis, USA
Abstract
Background: Wearable sensors (e.g. accelerometers) for tracking human physical activity have allowed for measurement of objective activity performance of the upper limb in daily life. Data extracted from accelerometers can be used to quantify multiple variables measuring different aspects of upper limb performance in one or both limbs. Work to date has focused on single variables, but upper limb performance is likely multidimensional. Here, we propose multivariate categories of upper limb performance, derived from wearable sensor data, as a potential solution for improving stroke rehabilitation care.
Methods: This study analyzed data extracted from bimanual, wrist-worn triaxial accelerometers in adults from three previous cohorts (N=211), two samples of persons with stroke and one sample from neurologically intact adult controls. Data used were upper limb performance variables calculated from accelerometer data, associated clinical measures, and participant demographics. A total of 12 cluster solutions (3-, 4-, or 5-clusters based with 12, 9, 7, or 5 input variables) were calculated to systematically evaluate the most parsimonious solution. Quality metrics and principal component analysis of each solution were calculated to arrive at a locally-optimal solution with respect to number of input variables and number of clusters. Data from earlier time points will be evaluated for their potential to predict eventual cluster membership.
Results/Anticipated Results: Across different numbers of input variables, two principal components consistently explained the most variance. Across the models with differing numbers of upper limb input performance variables, a 5-cluster solution explained the most overall total variance (79%) and had the best model-fit (AIC improvement of 184, compared to the next best model). The clusters are named by the amount of overall upper limb activity and integration of the upper limbs into daily activity. The category names in order of increasing upper limb performance are: Minimal Activity/Rare Integration, Minimal Activity/Limited Integration, Moderate Activity/Moderate Integration, Moderate Activity/Full Integration, and High Activity/Full Integration. People from the stroke cohorts ended up in all 5 categories while the adult controls ended up in the moderate to high categories. Variables that may predict eventual cluster membership of those in stroke cohort will be provided.
Discussion/Significance: We identified 5 categories of upper limb performance formed from 5 upper limb performance variables in cohorts with and without neurological upper limb deficits. Following validation on a larger, heterogeneous sample, these categories may be used as outcomes in upper limb stroke research and implemented into clinical rehabilitation practice.
Categories
Motor Rehabilitation
The Anatomical Tracings of Lesions After Stroke (ATLAS) Dataset – Release 2.0
Sook-Lei Liew1, Bethany Lo1, Miranda Donnelly1, Artemis Zavaliangos-Petropulu1, Jessica Jeong1, Giuseppe Barisano1, Alexandre Hutton1, Julia Simon1, Julia Juliano1, Anisha Suri2, Tyler Ard1, Nerisa Banaj3, Michael Borich4, Lara Boyd5, Amy Brodtmann6, Cathrin Buetefisch7, Lei Cao8, Jessica Cassidy9, Valentina Ciullo3, Adriana Conforto10,11, Steven Cramer12, Rosalia Dacosta-Aguayo13, Ezequiel de la Rosa14,15, Martin Domin16, Adrienne Dula17, Wuwei Feng18, Alexandre Franco8,19,20, Fatemeh Geranmayeh21, Alexandre Gramfort22, Chris Gregory23, Colleen Hanlon24, Brenton Hordacre25, Steven Kautz23,26, Mohamed Salah Khlif27, Hosung Kim1, Jan Kirschke28, Jingchun Liu29, Martin Lotze16, Bradley MacIntosh30,31, Maria Mataró13,32, Feroze Mohamed33, Jan Nordvik34,35, Gilsoon Park1, Amy Pienta36, Fabrizio Piras37, Shane Redman36, Kate Revill7, Mauricio Reyes38, Andrew Robertson39,40, Na Jin Seo41, Surjo Soekadar42, Gianfranco Spalletta3, Alison Sweet36, Maria Telenczuk22, Gregory Thielman43, Lars Westlye44,45, Carolee Winstein1, George Wittenberg46,2, Kristin Wong47, Chunshui Yu29,29
1University of Southern California, Los Angeles, USA. 2University of Pittsburgh, Pittsburgh, USA. 3IRCCS Santa Lucia Foundation, Rome, Italy. 4Emory University School of Medicine, Atlanta, USA. 5University of British Columbia, Vancouver, Canada. 6University of Melbourne, Melbourne, Australia. 7Emory University, Atlanta, USA. 8Child Mind Institute, New York, USA. 9University of North Carolina at Chapel Hill, Chapel Hill, USA. 10São Paulo University, Sao Paulo, Brazil. 11Hospital Israelita Albert Einstein, Sao Paulo, Brazil. 12University of California Los Angeles, Los Angeles, USA. 13University of Barcelona, Barcelona, Spain. 14icometrix, Leuven, Belgium. 15Technical University of Munich, Munich, Germany. 16University of Greifswald, Greifswald, Germany. 17The University of Texas Austin, Austin, USA. 18Duke University School of Medicine, Durham, USA. 19Nathan Kline Institute for Psychiatric Research, Orangeburg, USA. 20NYU Grossman School of Medicine, New York, USA. 21Imperial College London, London, United Kingdom. 22Université Paris-Saclay, Palaiseau, France. 23The Medical University of South Carolina, Charleston, USA. 24Wake Forest School of Medicine, Winston Salem, USA. 25University of South Australia, Adelaide, Australia. 26Ralph H Johnson VA Medical Center, Charleston, USA. 27The Florey Institute of Neuroscience and Mental Health, Heidelberg, Australia. 28Technical University Munich, Munich, Germany. 29Tianjin Medical University General Hospital, Tianjin, China. 30University of Toronto, Toronto, Canada. 31Hurvitz Brain Sciences Program, Toronto, Canada. 32Institut de Recerca Sant Joan de Déu, 08950 Esplugues de Llobregat, Spain. 33Jefferson Magnetic Resonance Imaging Center, Philadelphia, USA. 34CatoSenteret Rehabilitation Center, SON, Norway. 35Oslo Metropolitan University, Oslo, Norway. 36University of Michigan, Ann Arbor, USA. 37IRCCS Santa Lucia Foundation, Rome, USA. 38University of Bern, Bern, Switzerland. 39University of Waterloo, Waterloo, Canada. 40Sunnybrook Research Institute, Toronto, Canada. 41Medical University of South Carolina, Charleston, USA. 42Charité - Universitätsmedizin Berlin, Berlin, Germany. 43University of the Sciences, Philadelphia, USA. 44University of Oslo, Oslo, Norway. 45Oslo University Hospital, Oslo, Norway. 46Department of Veterans Affairs, Pittsburgh, USA. 47University of Texas at Austin, Austin, USA
Abstract
Stroke is a leading cause of disability. Predicting stroke outcomes is critical for therapeutic optimization and developing personalized rehabilitation strategies for patients. It has been shown that quantifying the extent of stroke damage through segmenting lesioned tissue and using these measures to capture injury to specific brain structures can be used as a biomarker to predict outcomes and recovery. Manual lesion segmentation remains the gold standard, but it is time-consuming, subjective and requires extensive neuroanatomical expertise, significantly limiting this approach for large-scale research. To address this, in 2018 we released a large, open-source dataset of stroke T1-weighted MRIs with manually segmented lesions (ATLAS v1.2, N=304). This dataset has been used extensively to develop and evaluate automated lesion segmentation methods, including non-stroke applications, demonstrating the utility of large, publicly available imaging datasets. Several promising automated segmentation methods were developed using this dataset, with reported Dice coefficients up to 0.670. Of the 17 methods published, however, most did not make their code public, limiting their value to the scientific community. A limitation of ATLAS v1.2 was the lack of separate, hidden test data, thereby allowing the possibility that algorithms were overfitted to the training data and reported performances were inflated.
To address these issues, we created ATLAS v2.0 (N=955), a larger, multi-site dataset of T1-weighted MRIs with corresponding manual lesion segmentations. It includes a public training dataset (n=655) and a hidden test dataset (n=300). The lesion segmentations in the test dataset are only available as part of lesion segmentation challenges for unbiased evaluation of algorithm performance. ATLAS v2.0 was created by lesion tracers who received standardized training in stroke neuroanatomy and lesion identification. A neuroradiologist was consulted for cases with complex pathology, and all segmentations were reviewed and approved by two separate raters. T1-weighted MRIs and lesion segmentations are available in native (raw) space on the Archive of Data on Disability to Enable Policy and research (ADDEP) at The Inter-university Consortium for Political and Social Research (ICPSR). In addition, a preprocessed dataset (intensity normalized, registered to an MNI-152 template, and defaced to remove potentially identifying facial features) was developed, visually inspected, and made publicly available on the International Neuroimaging Data-sharing Initiative (INDI) platform.
We anticipate that the release of this updated, multi-site dataset, with a larger sample size and hidden test data, will help both refine current methods and spur the development of new methods. In particular, the test data will allow for the unbiased evaluation of new automated stroke lesion segmentation methods via segmentation challenges. We expect that ATLAS v2.0 will lead to the development of open- source automated segmentation methods that perform on par with manual segmentation, thereby reducing impact of a key limitation surrounding large-scale stroke neuroimaging research.
Categories
Stroke
Motor overflow in the leg after stroke: minimal role for corticomotor pathways
Brice Cleland, Sangeetha Madhavan
University of Illinois at Chicago, Chicago, USA
Abstract
Introduction: Involuntary activation of a limb contralateral to a contracting limb (motor overflow) occurs after stroke, particularly in the non-paretic limbs. Motor overflow may occur because the contralesional hemisphere assumes greater motor control of the paretic limb or because inhibition from the ipsilesional to contralesional hemisphere is diminished. This study investigated motor overflow and its mechanisms in the legs of individuals with stroke. We hypothesized that motor overflow would be greater in the non-paretic than paretic leg and would be associated with greater excitability of contralesional pathways to the paretic leg and lesser inhibition from the ipsilesional to contralesional hemisphere.
Methods: Individuals with chronic stroke performed unilateral isometric and dynamic ankle dorsiflexion. Motor overflow was quantified with electromyography in the non-target tibialis anterior, which was normalized to rest (>1 reflects motor overflow). To characterize corticomotor excitability to the non- target leg, transcranial magnetic stimulation was applied, and motor evoked potentials (MEPs) in the non-target leg were assessed. To provide insight into motor overflow mechanisms, we measured MEPs in the target leg. Ipsilateral motor excitability was assessed as relative ipsilateral MEP amplitude.
Interhemispheric inhibition was assessed as the ipsilateral silent period duration. Correlations between motor overflow, ipsilateral excitability, and interhemispheric inhibition were assessed.
Results: 28 individuals with stroke were included [62 (7) years; 24 male; 5.8 (3.9) years poststroke]. Motor overflow in the non-target leg occurred during both conditions in both legs (p≤0.002) and was greater in the non-paretic than the paretic leg [4.1 (4.2) vs. 2.1 (1.8)]. Some participants (25%) had pronounced motor overflow in the non-paretic leg, but there was no difference between legs (p=0.24) in the remaining sample. In both legs, non-target MEPs were greater during isometric and dynamic conditions than rest (p≤0.01), but not when normalized to background muscle activity (p=0.10). Greater motor overflow in the non-paretic leg during the dynamic condition was associated with lesser inhibition from the ipsilesional to the contralesional hemisphere (p=0.01).
Discussion: Our findings suggest that there is motor overflow in both legs during isometric and dynamic ankle dorsiflexion after stroke. Motor overflow was pronounced in the non-paretic leg in a subset (25%), but similar between legs in the remaining sample. MEPs in the non-target leg suggest that corticomotor excitability is greater during isometric and dynamic conditions than rest, but not because of differences in cortical excitability. Instead, motor overflow may reflect increased excitability of non-corticomotor pathways. We found that motor overflow in the non-paretic limb was associated with lesser interhemispheric inhibition from the ipsilesional to contralesional hemisphere, which may contribute to motor overflow at a subcortical level. Understanding motor overflow after stroke is important because it may impair lower limb performance or be exploited to enhance functional gains.
Categories
Stroke
Improvement in capacity for activity vs. improvement in performance of activity in daily life during outpatient neurorehabilitation
Catherine E. Lang1, Carey L. Holleran1, Michael J Strube1, Terry D. Ellis2, Caitlin A. Newman3, Meghan Fahey3, Tamara R. DeAngelis2, Timothy Nordahl2, Darcy S. Reisman4, Gammon M. Earhart1, Keith R. Lohse1, Marghuretta D. Bland1
1Washington University School of Medicine, Saint Louis, USA. 2Boston University, Boston, USA. 3Shirley Ryan Ability Lab, Chicago, USA. 4University of Delaware, Newark, USA
Abstract
An assumption amongst rehabilitation providers has been that as capacity for activity improves during an episode of neurorehabilitation service, performance of activity in daily life must also improve. Data from upper limb (UL) stroke research interventions suggest there is a discrepancy between improved capacity (measured via standardized clinical tests) and improved performance (measured via wearable motion sensors). This study addressed three questions about this discrepancy:
1. Is it common in routine, outpatient care, or just an artifact of intervention studies?
2. Is it unique to UL rehabilitation post stroke, or is it seen in walking rehabilitation too?
3. Is it only seen in persons with stroke, or is it a broader neurorehabilitation problem?
The study design was a longitudinal, prospective, observational cohort of 156 people participating in outpatient rehabilitation services at five clinics. Participants with stroke or Parkinson disease (PD) were enrolled adjacent to the time of their initial outpatient visit and assessed monthly during their episode of care. Question 1 was addressed by analyzing the full sample. Question 2 was addressed by comparing individuals with stroke undergoing occupational therapy with goals to improve UL function vs. individuals with stroke undergoing physical therapy with goals to improve walking. Question 3 was addressed by comparing the same stroke walking subgroup vs. individuals with PD also with walking goals. Individual, longitudinal capacity (Action Research Arm Test or walking speed) and performance (wearable sensor measurements of use ratio or steps/day) data were modeled to extract predicted change scores.
Bootstrapping determined whether an individual’s change was greater than 1 SE error of their own simulated distributions. Participants were classified into four outcomes: 1) improved capacity and improved performance (C+P+); 2) improved capacity and unimproved performance (C+P-); 3) unimproved capacity and improved performance (C-P+); and 4) unimproved capacity and unimproved performance (C-P-).
With respect to question 1, the majority of the full sample (69%) were classified as C+P-. A smaller portion of the sample (20%) were classified as C+P+, with the remaining participants (11%) classified as C-P-. With respect to questions 2 and 3, the discrepancy was not unique to the UL nor to stroke, as the proportions in the C+P- category were significantly worse in the stroke (X2=48.7, p<0.0001) and PD (X2=22.9, p< 0.0001) walking subgroups compared to the stroke UL subgroup.
These data indicate that the discrepancy between improvements in capacity vs. performance is a widespread problem in outpatient neurorehabilitation care. Our data do not indicate that performance cannot change, only that it often did not change in the current delivery model. These data open up the opportunity that if performance information were available, then patients and clinicians could act to address and improve it.
Categories
Motor Rehabilitation
Spinal Cord Injury: Do Residual Tissue Bridges Effect Neurophysiology and Functional Recovery?
Alyssa Canales, Marylu Cabello, Alondra Medina, Kelsey Baker
University of Texas Rio Grande Valley, Edinburg, USA
Abstract
Introduction: Spinal cord injuries affect about 300,000 individuals in the United States and cause varying degrees of motor and sensory impairment. The standard of care for those with a spinal cord injury typically involves traditional rehabilitation. Unfortunately, over a year of rehabilitation is often required to achieve meaningful improvements in motor or sensory function. In order to improve rehabilitative efforts for those with spinal cord injury, it is critical to identify biomarkers that could serve as indicators for those who would respond to standard rehabilitation efforts.
Objective: Following spinal cord injury, edema forms around the area of injury in the spinal cord. Recent work has suggested that the characteristics of the edema and neighboring tissue bridges in the spinal cord can influence both baseline function and functional recovery potential. The objective of our project is to assess magnetic resonance imaging (MRI) of the spinal cord and relate observed properties of the spinal edema and tissue bridges with functional recovery of the patient.
Methods: T2-weighted MRI images were collected in seven subjects with chronic cervical spinal cord injury (C2 to C7). Following MRI collection, all patients then participated in two weeks of rehabilitation. We assessed changes in motor function before and after rehabilitation using the nine-hole peg test and muscle grading. We defined the physical properties of the spinal scar using the open-source software, FSLview. We related the length, size and location of the spinal scar to functional recovery metrics using regression analysis. Statistical Package for the Social Sciences (SPSS) software was used for all statistical comparisons.
Results: We observed varying sizes of spinal edema in our patient population (0.1 mm3 to 3 mm3). Spinal edema varied substantially in length between patients, with some patients demonstrating spinal edema that spanned several cervical vertebrae levels. We observed that baseline function was related to dorsal and ventral tissue bridge sparing (r = .930 p = .007). We observed that more intensive stimulation was needed to produce a muscle twitch in patients with smaller tissue bridge (r= 0.55).
Conclusions: Identification of biomarkers could aid clinicians in determining what rehabilitation approaches to employ in a specific patient. Our work has found that properties of the spinal edema and neighboring tissue bridges are directly related to baseline function and recovery potential. Our findings encourage future research to evaluate mechanisms to improve the viability of tissue bridges in individuals with spinal cord injury, such as through use of stem cells, in order to improve functional recovery.
Categories
Spinal Cord Injury (SCI)
Impairment estimation from high dimensional motion data during functional task performance
Avinash Parnandi1, Anita Venkatesan1, Natasha Pandit1, Audre Wirtanen1, Emily Fokas1, Grace Kim2, Dawn Nilsen3, Heidi Schambra1
1NYU Langone, New York, USA. 2NYU Steinhardt, New York, USA. 3Columbia University Medical Center, New York, USA
Abstract
Introduction: Current approaches to quantify upper extremity (UE) motor impairment after stroke are impractical, limiting our ability to track recovery during rehabilitation. The UE-Fugl-Meyer Assessment (UE-FMA) is considered the current gold standard to measure UE impairment, but it is hampered by excessive time to administer and the requirement of a trained assessor. To address these issues, we present an approach that uses wearable sensors and deep learning (DL) to automatically quantify UE impairment during real-world functional tasks.
Methods: We studied 41 chronic stroke patients (22F:19M; 58.6 (21.3-84.3) years old; 22L:19R paretic; UE-FMA 48.9 (8-65)). We recorded upper body motion with 6 inertial measurement units (IMUs) while patients performed the FMA and a battery of functional tasks. We trained a long short-term memory (LSTM) DL model to estimate FMA scores from the recorded motion (training set n=33; testing set n=8). To ascertain the accuracy of the approach, we calculated the mean absolute error between the LSTM- generated scores and the FMA scores from a trained assessor. To identify a streamlined setup, we selectively reduced sensors and data dimensions and measured the error. Finally, to identify the contributions of UE and trunk sensors on the error, we performed occlusion analysis on the data.
Results: Using motions from FMA, the LSTM estimated FMA scores within 3.2 points of a trained assessor. Using motions from functional tasks, the LSTM estimated FMA scores within 4.1-10.4 points, with higher error for unstructured tasks. Using motions from tabletop task, we found that all 6 sensors and 58 data dimensions resulted in the lowest error (4.1 points). By reducing the sensor count, we saw an increase in error (5.8-7.9 points) with lower error for 3 sensors on the UE, and higher error for single sensor on arm and forearm. By reducing the data dimensions, we again saw an increase in error (7.0-13.0 points) with lower error for IMUs and higher error for accelerations. Finally, using motions from tabletop task, we found that for all patients, UE sensors were more informative (error 7.5 points) than trunk sensors (20.3 points). Interestingly, we found that trunk sensors became more relevant in patients with worse impairment (error 19.6 points) than those with low impairment (21.0 points).
Discussion: We present an approach that uses wearable sensors and DL to automatically estimate UE motor impairment. This approach is accurate on motion data from FMA and unrelated functional motions. However, it is possible that the LSTM may overfit the data given its high dimensionality, which may result in poor generalization despite good performance in our test set. Future work involves trialing other DL models that first learn the universal motion characteristics in healthy controls, and then use these characteristics to detect abnormalities in stroke patients.
Categories
Stroke
Detection of Stroke-Induced Spatial Neglect and Prediction of Neglected Visual Targets with an Augmented Reality (AR)-Encephalography (EEG) System
Jennifer Mak1, Deniz Kocanaogullari1, Xiaofei Huang2, Minmei Shih1, Emily Grattan1, Sarah Ostadabbas2, George F. Wittenberg1, Elizabeth Skidmore1, Murat Akcakaya1
1University of Pittsburgh, Pittsburgh, USA. 2Northeastern University, Boston, USA
Abstract
Spatial neglect is a common consequence after stroke. It is characterized by the inability to attend to or perceive contralesional stimuli. Current pen-and-paper clinical assessments such as the Behavioral Inattention Test (BIT) are not able to adequately quantify the extent of neglect, account for compensatory trunk and head movements, or yield consistent results. Our overarching goal is to develop a multipurpose system to progressively detect, assess, and rehabilitate neglect. We have developed a system that integrates augmented reality (AR) with electroencephalography (EEG) for neglect detection and estimation of the neglected field. This setup is easily adaptable for rehabilitation, is portable, and controls for the user’s head movements. We developed two versions of the Starry Night Test for the Microsoft Hololens AR headset: 1) a Clicker-Based Assessment to measure the ground-truth field-of-view (FOV) via reaction times to visual targets appearing among distractors, and 2) an EEG-Based Assessment to evaluate the neural response to those targets using 16 electrodes. Five patients with spatial neglect (SN) (two male, mean age = 61±19 years old) and five without spatial neglect (WSN) (four male, mean age = 54±23 years old) were recruited and categorized using BIT scores. All participants performed the Clicker-Based Assessment followed by the EEG-Based Assessment. In this preliminary exploratory analysis, we first aimed to identify spectral features for optimal neglect detection. The median reaction times from the Clicker-Based Assessment were thresholded into fast and slow responses. The power of five frequency bands (delta (0-4 Hz), theta (5-8 Hz), alpha (9-13 Hz), beta (14-30 Hz), and gamma (31-45 Hz)) was calculated for each baseline-corrected trial. A power ratio was calculated for each electrode and frequency band such that the log of the power spectral density in each fast response trial was divided by the average power in the slow response trials. Wilcoxon rank sum tests of the power ratios found statistically significant differences between SN and WSN groups in almost whole brain delta and beta and bilateral central-parietal theta, frontal-parietal alpha, and frontal-central gamma. A logistic regression analysis of the power ratios was able to separate SN from WSN with an area under the curve (AUC) of 0.886. We next aimed to use EEG signals to predict whether a target would be fast or slow using common spatial patterns as the feature extraction algorithm and regularized discriminant analysis combined with kernel density estimation (RDA+KDE) for classification. With 10-fold cross validation, the RDA+KDE yielded an average training AUC of 0.788 and average testing AUC of 0.760. These early results suggest that our system may be able to accurately detect neglect and use EEG to predict potentially neglected locations in the FOV as labeled by neglected targets.
Categories
Stroke
Cortical P300 amplitude reduction post concussion helps track neurophysiological change following ABI
Frank Palermo
University Colorado, Boulder, USA
Abstract
Pre season auditory cortical P300 baseline testing followed by post concussive comparison demonstrated greater than 50% reduction in amplitude. Evaluations were carried out over 5 years including 461 baseline tests and 47 post concussive tests. Pre-season testing compared with repeat preseason tests produced a variance (SD) of 12% within subjects. Post-concussive testing at 24 hrs showed a reduction of 4 SD progressively returning to within one SD at clearance to return to play in 70% of the athletes based on NCAA Guidelines. At the start of the following season, all returned to within one SD of baseline.
Additionally, we use the same cortical auditory P300 testing to establish baselines prior to rehab interventions following ABI.
Categories
Traumatic Brain Injury (TBI)
Decoding speech from human motor cortex using an intracortical brain computer interface
Daniel Rubin1,2, Tommy Hosman3, Anastasia Kapitonava1, John Simeral4,3, Sydney Cash1,2, Leigh Hochberg4,3,1,2
1Massachusetts General Hospital, Boston, USA. 2Harvard Medical School, Boston, USA. 3Brown University, Providence, USA. 4VA Providence Health Care System, Providence, USA
Abstract
Background: For people with brainstem stroke, locked-in syndrome, ALS, muscular dystrophy, and other neurologic conditions causing quadriparesis and severe dysarthria/anarthria, the loss of fluent communication is highly disabling. Commercially available assistive technologies harness residual volitional muscle control (e.g., eye movement), but may be cumbersome, slow, and ultimately may fail as disease progresses. Intracortical brain computer interfaces (iBCIs), including the BrainGate Neural Interface System, bypass lesional neural tissue to transmit information directly from cortex to an assistive device. Previously, iBCIs have allowed people with paralysis to type and write individual letters and decode phonemes. Here we describe efforts to decode intended speech using an iBCI. Understanding how the semantic information of language is translated into motor signals within cortex will enable the development of more accurate algorithms for decoding speech directly from neural activity.
Methods: Research is conducted with permission under an IDE from US FDA and the MGH IRB. A 37- year-old with quadriplegia from a spinal cord injury enrolled in the BrainGate clinical trial had two 96- channel microelectrode arrays placed chronically in dominant precentral gyrus. During recording sessions in his home, the participant read a series of pseudo-randomly presented words while we recorded neural activity. To vary context, word cues were presented as either text, pictures, or both, and another cue type indicated whether the word was to be pronounced with a normal or prolonged enunciation.
Results: Principal component analysis of the 192-channel neural activity identified the 35-dimensional feature space capturing the greatest variance across the task. We trained a support vector machine (SVM)-based decoder to distinguish between each spoken word and epochs of silence. In cross-validated analyses, the SVM correctly distinguished speech from silence in 96.6% of 2440ms epochs spanning 1176 spoken words over 3220 seconds of word reading. Individual articulatory elements of speech were well discriminated, with clustering of words observed most strongly on initial consonant sound. Decoding performance was enhanced in the elongated compared to normal enunciation context and in the combined picture and text compared to either picture or text alone contexts, suggesting that task context influences the discriminability of the neural representation of speech in motor cortex.
Conclusions: In this preliminary study, iBCIs can be used to detect intended speech directly from motor cortex. Ongoing work is focused on refining decoding algorithms to restore fluent speech-based communication to people with severe dysarthria and anarthria.
Categories
Cognitive/Language Rehabilitation
A Systematic Decomposition of Upper-Body Dressing
Emily Fokas1, Avinash Parnandi1, Zuha Ahmed1, Anita Venkatesan1, Natasha Pandit1, Audre Wirtanen1, Dawn Nilsen2, Heidi Schambra1
1NYU Langone, New York, USA. 2Columbia University, New York, USA
Abstract
Introduction: Upper-body dressing (UBD) is a particularly intimate activity of daily living (ADL) and is trained extensively in stroke rehabilitation. The amount of UBD training needed to achieve independence is currently unknown. Furthermore, aside from scoring the level of assistance needed, there exist no metrics to measure UBD abnormality. An approach to decompose UBD into recognizable units of motion would serve the quantitation of UBD training and the assessment of UBD motion abnormality.
We previously created a taxonomy to characterize functional upper extremity (UE) motions, identifying five classes of functional primitives (reaches, repositions, transports, stabilizations, idles). Functional primitives are units of motion that are strung together to complete most rehabilitation activities, but their use in UBD has not been assessed. Dressing engages clothes, fluid objects that envelop the UE, unlike the rigid, hand-sized objects targeted in most functional activities (e.g. utensils, handles). The resulting UBD movement strategies thus do not guarantee a mapping of the taxonomy. Here, we applied the taxonomy to UBD tasks to identify and operationalize phenotypic differences. Using the modified taxonomy, we identified the interrater reliability (IRR) of this UBD classification system.
Methods: We videotaped seven healthy adults (3M:4F, 59.1 ± 12.0 years) as they donned and doffed a t-shirt, button-up shirt, and zippered jacket. Two cameras were placed ~1 m orthogonal to the subject. Experts in the taxonomy (EF, AP, HS) examined videos from two subjects, and operationalized primitive characteristics that were unique to dressing. Two trained coders (EF, ZA) applied the taxonomy to the remaining subjects, and IRR (Cohen’s kappa) was assessed. Primitive counts were totaled across the three tasks for each subject, and then averaged across subjects.
Results: The taxonomy could account for all motions in the sampled UBD tasks. Dressing-specific primitives differed from standard primitives in that they often engage the full UE range of motion, are performed outside of the subject’s field of vision (i.e. behind the back), and involve sinuous or rotational motions. The three UBD tasks combined generated, on average, 110.8 ± 12.7 primitives, composed of 39.8 ± 3.9 reaches, 10.0 ± 1.2 repositions, 43.6 ± 4.4 transports, 8.2 ± 6.9 stabilizations, and 7.8 ± 2.2 idles.
IRR was moderate-to-high for most primitives: reach, k = 0.83; reposition, 0.96; transport, 0.71; stabilization, 0.08; idle, 0.90. Disagreements in primitive boundaries were, on average, 0.13 ± 0.22 seconds different.
Discussion: The proposed dressing taxonomy was effective at decomposing UBD into measurable primitive motions. It supported moderately high IRR except for clothing stabilization, where minimal motion was commonly confused for purposeful clothing transport. This dressing taxonomy could support the quantification of UBD training to identify training intensities needed for regaining independence.
Dressing primitives could also serve as the basis for kinematic abnormality assessment.
Categories
Motor Rehabilitation
Abnormal motor control in the arm and not in the finger is linked to increased CReST activity during an arc pointing task in chronic stroke patients
Myriam Taga1, Yoon N. G. Hong2, Charalambos C. Charalambous3, Sharmila Raju1, Jing Lin1, Pietro Mazzoni4, Jinsook Roh2, Heidi M. Schambra1
1Department of Neurology, NYU Langone, School of Medicine, New York, USA. 2Department of Biomedical Engineering, University of Houston, Houston, USA. 3Department of Basic and Clinical Sciences, Medical School, University of Nicosia, Nicosia, Cyprus. 4Department of Neurology, Washington University, School of Medicine in St. Louis, St. Louis, USA
Abstract
Introduction: In humans, the two major descending motor pathways are the corticospinal tract (CST) and corticoreticulospinal tract (CReST). Stroke damage to the CST results in a loss of motor control, which may result in compensatory upregulation of CReST. Due to its multi-spinal innervation, CReST upregulation may cause abnormal muscle co-activation in the paretic upper extremity (UE). Here, we studied how motor control relates to muscle co-activation during purposeful UE movement using either the arm or hand, as the balance of CST and CReST innervation may differ for these two UE effectors.
Methods: We studied eight chronic stroke patients (6 female, mean (range) age 64 (52 – 84) years, time post-stroke 4.1 (0.5 – 10) years) and eleven healthy controls (5 female, mean (range), age 53 (36 – 75) years). Subjects used their paretic arm or index finger (assigned in controls) to move an onscreen cursor through an arc-shaped channel. The UE was otherwise immobilized, preventing muscle co-activation from intruding on motor control. We recorded effector performance with an infrared camera. We recorded muscle activations of six bystander muscles with electromyography (EMG). For arm and finger movements, we quantified loss of CST function using a motor control abnormality score (error rate*path irregularity) and quantified CReST upregulation using muscle co-activation (average peak EMG amplitudes of bystander muscles). Higher motor control abnormality scores indicate worse motor control and higher muscle co-activations indicate more abnormal (i.e. widespread) muscle activation. For the arm and the finger, we examined differences between subject groups with Wilcoxon tests and relationships between motor control and muscle co-activation measures with Spearman rank correlations.
Results: Motor control abnormality scores were higher for patients than controls in the arm (p=0.005) and finger (p=0.001). Muscle co-activation was higher for patients than controls during arm (p=0.017) and finger (p=0.006) movement. In patients, greater motor control abnormality strongly correlated with greater muscle co-activation for the arm (rho=0.857, p=0.006) but not the finger (rho=0.190, p=0.651). In controls, muscle co-activation was not correlated with motor control for either effector (all p>0.832).
Discussion: We confirmed that patients had worse motor control and higher muscle co-activations than healthy controls during performance of controlled movements. Interestingly, poorer motor control scaled with increased muscle co-activation only in the arm of stroke patients. It is possible that CST and CReST innervation is more closely integrated in the arm than hand, and that loss of CST function for the arm may prompt a more calibrated upregulation of CReST.
Categories
Stroke
A preliminary investigation of the neural correlates of balance performance in healthy adults
Vyoma Parikh, Ann Medley, Hui-Ting Goh
Texas Woman’s University, Dallas, USA
Abstract
Background: Balance performance is crucial for independent living after neurologic injuries. Postural stability involves multiple neural mechanisms along the nervous system with the cerebellum and primary motor cortex (M1) being significant contributors. Therefore, interventions, such as non-invasive brain stimulation (NIBS), targeting these two neural loci are theorized to enhance balance performance. The effect of M1 NIBS on balance has been studied extensively but little is known about the effect of cerebellar NIBS. It is unclear which neural locus is a better target to enhance balance performance. As the first step in a series of experiments, we determined the role of two neural loci, M1 and cerebellum, in modulating balance performance in young healthy adults using repetitive transcranial magnetic stimulation (rTMS).
Methods: Fourteen healthy young adults (mean age= 27.8 years) received 5Hz M1 rTMS and 5Hz cerebellar rTMS in a cross-over randomized manner with a washout period of approximately 7days. Balance performance was assessed before and after rTMS using three tests on the Biodex Balance system SD: the limits of stability (LOS) test, the modified clinical test of sensory interaction on balance (mCTSIB) in which proprioception (firm vs. foam surface) and vision (eyes-closed vs. eyes-open) were manipulated, and the balance error scoring system (BESS) test in which the posture (feet together, tandem, single leg) and proprioception (firm vs. foam surface) were varied. Time and overall sway angle were used to quantify performance on the LOS test. mCTSIB and BESS performance were quantified by sway index with lower values indicating better balance performance.
Results: Participants improved significantly on the LOS time (p =.03) but not on the LOS overall sway angle after both M1 and cerebellar rTMS. The effects of rTMS on the mCTSIB tended to be mediated by stimulation target, proprioception, and vision (p = .06, ŋ2= 0.25). When proprioception was compromised (foam surface), cerebellar rTMS improved sway index under eyes-closed but not eyes-open condition whereas M1 rTMS improved sway index under eyes-open but not eyes-closed (p = .06). The effects of rTMS on the BESS test also tended to be mediated by stimulation targeting, posture and proprioception (p = .08, ŋ2= 0.19). M1, but not cerebellar, rTMS improved sway index across all three postures when standing on the firm surface (p = .04, ŋ2= 0.25).
Conclusion: Both M1 and cerebellar rTMS led to a comparable improvement in LOS time, suggesting a benefit for proactive postural control. However, a practice effect could not be ruled out. Cerebellar rTMS appeared to enhance balance performance when sensory feedback was compromised. In contrast, M1 rTMS tended to be favorable when the motor system was challenged. Our future directions include recruiting more healthy participants and extending the investigation to individuals with stroke.
Categories
Motor Rehabilitation
Expectation- and suggestibility-related placebo effects of tDCS on cognitive and motor training
Nicole Haikalis, Andrew Hooyman, Peiyuan Wang, Sydney Schaefer
Arizona State University, Tempe, USA
Abstract
Transcranial direct current stimulation (tDCS) may be a complementary therapy for augmenting cognitive or motor training in neurorehabilitation. However, studies using tDCS to enhance such training often yield null results. While differences in stimulation parameters (e.g., stimulation site, intensity) may contribute to mixed findings, it is also plausible that mixed findings are due to uncontrolled placebo effects that mask or inflate improvements in cognitive or motor performance post-stimulation. Intrinsic factors, like one’s expectations about a treatment and one’s overall suggestibility, are often sources of placebo effects in general. We have recently shown that the general public has higher-than-neutral expectations about tDCS; thus, the purpose of this study was to test for a placebo effect of tDCS on both cognitive and motor training. Eighty-seven participants (age: 22.20±4.16 years; 42 females) were randomly assigned to one of three groups: active anodal tDCS (n=29), sham tDCS (n=32), or no-tDCS control (n = 26). All participants completed a pre-test of the Shepard-Metzler Mental Rotation Task, followed by 20 minutes of Corsi Block Tapping Task (CBTT) training. tDCS stimulation (20-min, 2mA) was applied to the right parietal area (P4) during CBTT training (sham/anodal tDCS groups only). Participants were re-tested on mental rotation to determine changes in cognitive (visuospatial) performance, then completed 30 trials of motor training on a functional reaching task. After motor training, participants’ overall suggestibility, expectations about tDCS, and any previous tDCS experience were surveyed. Linear mixed-effects models tested for effects of tDCS and placebo on the three tasks. There was no effect of active or sham tDCS on mental rotation (p=.25). For the CBTT, there was a significant placebo effect, with both sham and active tDCS groups performing better after training than the control group (p<.05), with no difference in performance between sham and tDCS groups (p=.22). Participants with previous knowledge of tDCS showed more improvement on the CBTT during training compared to those who did not (p=.03), and participants who believed they received active tDCS improved more than those who believed they received sham tDCS (p=.017). While there were no significant tDCS or placebo effects on motor training (p=.49), higher suggestibility was associated with better performance on the motor task after training (p=.007). Thus, participants with high suggestibility may be more susceptible to placebo effects based on outside influences. This exploratory study is among the first to quantify a placebo effect of tDCS on both cognitive and motor training, which was stronger than the ‘treatment’ effect of tDCS and likely driven by participants’ suggestibility, expectations about tDCS, and prior tDCS experience. Future studies should not only measure and control for these factors but also identify ways to leverage placebo effects of tDCS to maximize therapeutic outcomes.
Categories
Other
Interaction of transcranial direct current stimulation (tDCS) and visual feedback in an ankle motor control task
Mark Cummings, Aditi Doshi, Farid Ihmoud, Lubna Shah, Sangeetha Madhavan
University of Illinois at Chicago, Chicago, USA
Abstract
Transcranial direct current stimulation (tDCS), a non-invasive neuromodulatory tool, has been demonstrated to facilitate motor learning and performance in healthy individuals; however, a consensus is yet to be achieved. The effects of tDCS on skill acquisition maybe task specific with visuomotor tasks more reliant on extrinsic feedback, thereby minimizing the effects of neuromodulation. Our objective was to explore if tDCS over the primary lower limb motor cortex differentially facilitates motor performance based on the availability of visual feedback. We hypothesized that ankle motor performance during a skilled visuomotor task will decrease with fading visual feedback, and that effects of tDCS will be greater as visual feedback declines.
22 neurotypical adults (mean (s.d) age = 24 (4) years, 11 males:11 females) performed ankle plantarflexion and dorsiflexion movements in a custom-built ankle tracking device to match a computer- generated target sine wave. Motor performance was measured using root-mean-squared error (RMSE), the difference between ankle position and target, normalized to each participant’s range of motion. Each participant attended two sessions, conducted a week apart, one without tDCS (visual feedback (VF) only) and one with tDCS (VF-tDCS). Each session was divided into two blocks containing 3 randomized VF settings: 1) 100%VF, full VF of the ankle position and target, 2) 0%VF, VF of ankle position but not the target, and 3) NoVF, participants were blindfolded during movement. During the VF-tDCS sessions, the first block included the application of 1 mA of tDCS to the lower limb M1 for 20 mins.
A one-way repeated measures ANOVA revealed that there was a main effect of VF on RMSE (p < .001). We found that the RMSE increased as the VF faded. A two-way repeated measures ANOVA showed a significant interaction of tDCS and VF (p = .027). Post-hoc analyses revealed significant differences in performance scores (change in RMSE scores between two blocks of each session) in the NoVF condition between different stimulation settings (VF-tDCS and VF) (p = .006). We found a 3% decrease in RMSE during tDCS compared to no tDCS application over the lower limb motor cortex for the NoVF condition.
Our results support that tDCS enhances ankle motor performance only when visual feedback is not available. These results may have clinical implications in neurological studies that rely on neuromodulation via tDCS to enhance lower limb motor control.
Categories
Motor Rehabilitation
Phonological Component Analysis augmented by anodal HD-tDCS: A case study examining behavioral and fMRI data in a patient with aphasia
Sara Pillay1, Cindy Li1, Priyanka Shah-Basak1, Joe Heffernan1, Lisa Conant1, Anna Frieberg1, Shelley Laitinen1, Samantha Hudson1, Jed Mathis1, Sabine Heuer2, Roy Hamilton3, Jeffrey Binder1
1Medical College of Wisconsin, Milwaukee, USA. 2University Wisconsin-Milwaukee, Milwaukee, USA. 3University of Pennsylvania, Philadelphia, USA
Abstract
People with aphasia often have phonological deficits as a result of damage to cortical and white matter pathways in the posterior perisylvian region thought critical for phonological retrieval processes. To understand what brain mechanisms support recovery in people with damage to this system, we measured brain activity in patients with phonological retrieval deficits as they read words aloud during fMRI1. Activation was greater during correct trials in the left angular gyrus (AG), a major node in the semantic network that lies immediately posterior to the phonological retrieval system. While this activation suggests that the perilesional semantic system promotes phonological retrieval, it is possible that patients are reorganizing damaged phonological processes within pre-existing language areas. To examine this hypothesis further, we studied the effect of a phonology-focused speech therapy intervention (Phonological Component Analysis, PCA) augmented with high-definition transcranial direct current stimulation (HD-tDCS) to the left AG. TDCS has been shown to enhance performance on some tasks, and we were interested to test whether it can optimize process-specific training methods and if this enhances neural reorganization.
We present a case study of a patient with phonological retrieval deficits who received anodal and sham HD-tDCS during two separate 2-week PCA therapy cycles, complete with task and resting-state fMRI sessions, and detailed language assessment battery at 3 times (pre-therapy, post-cycle 1, post-cycle 2). The patient was a 66-year-old woman 10-months post left MCA stroke. She received anodal stimulation to the left AG during the first cycle (4x1 montage, 2mA, 20 min, 10 days) and sham stimulation during the second cycle, approximately 3-months apart. Word and Picture items chosen for therapy were baseline-tested twice using an optimally efficient, adaptive procedure to identify items she reliably could not name, and to ensure that the items chosen represented an optimal level of difficulty. Matched lists were created to assess generalization. She improved on all lists after both cycles, with no clear benefit of anodal > sham. Therapy list item gains were maintained until the 10-week post-assessment time point. On the larger battery, she appeared to benefit from a-tDCS, and improved on tasks requiring phonological retrieval.
The patient was strongly left-hemisphere dominant for language pre- and post-therapy cycles. Rs-fMRI focused on characterizing connectivity changes between a priori defined semantic and phonological regions. Analysis revealed significant connectivity increases within left semantic, left phonological, and right phonological networks, and decreased connectivity between semantic and phonological networks after a-tDCS. After s-tDCS, gains in left phonological network connectivity were reduced, and connectivity between semantic and phonological networks moved toward baseline. The results demonstrate feasibility of pairing process-based speech therapy with tDCS and show that behavioral improvements may be facilitated by connectivity reorganization between phonological and semantic nodes, particularly in the left hemisphere.
Categories
Cognitive/Language Rehabilitation
Understanding the effects of cross-priming using non-paretic leg movement in severe stroke
Hyosok Lim1,2, Sangeetha Madhavan1
1Brain Plasticity Laboratory, Department of Physical Therapy, University of Illinois at Chicago, Chicago, USA. 2Graduate program in Rehabilitation Sciences, College of Applied Health Sciences, University of Illinois at Chicago, Chicago, USA
Abstract
Background & purpose: Movement-based priming is the use of specific movement paradigms to facilitate neural activity and augment the effects of motor therapy. Unilateral motor tasks are associated with bilateral activation of the motor cortex (M1). The recent focus on the adaptive role of the non-lesioned M1 in stroke survivors with severe deficits suggests the potential use of non-paretic limb movements as a novel cortical priming approach (cross-priming) to facilitate paretic limb motor recovery. This study aimed to determine if non-paretic lower limb movements have a priming effect on the paretic lower limb in people with severe stroke.
Methods: Eleven individuals with severe lower limb impairment post stroke (paretic FMLE score < 21 & absence of contralateral motor evoked potentials (MEP) in the paretic tibialis anterior (TA)) participated in the study (10.5±7.6 years post stroke; paretic FMLE score 17.0±2.7). Participants performed three 20- min priming conditions using their non-paretic ankle in a randomized order separated by 7 days: dynamic movement (DP), isometric resistance (IP), and sham priming (SP). DP and IP involved tracking visual targets that corresponds to 80% of participant’s maximum ankle range and strength respectively, while SP included subthreshold electrical stimulation of the posterior tibial nerve on the medial ankle while seated. Contralateral MEPs from the non-paretic TA and ipsilateral MEPs from the paretic TA, and transcallosal inhibition (TCI) from the non-lesioned to lesioned M1 were measured using transcranial magnetic stimulation. Motor control of the paretic ankle was measured using a reaction time task. All outcomes were measured before, immediately post, 30-min post, and 60-min post priming.
Results: Contralateral MEPs from the non-paretic TA increased by 24% and 21% immediately after DP and IP respectively (p<0.05). Ipsilateral MEPs from the paretic TA increased by 26% after DP (p<0.01), but not IP. These changes in MEPs were significantly greater compared to SP (all p<0.05) and retained 30-min and 60-min after priming. Although not significant, TCI increased by 27% after DP while IP decreased by 6% and reaction time of the paretic ankle improved after DP (11%) and IP (12%). No changes in contralateral and ipsilateral MEPs, TCI, and reaction time were observed during SP.
Conclusion: Our results indicate that the activation of the non-lesioned M1 after dynamic movement priming using the non-paretic ankle facilitated the ipsilateral descending pathways with mild improvement of paretic limb motor control. Isometric resistance priming decreased inhibition to the lesioned M1, however, changes in contralateral MEPs from the paretic TA could not be determined. These data provide evidence for the potential benefits of a novel priming paradigm using non-paretic limb movements to deliver alternative treatments for severely impaired stroke survivors.
Categories
Stroke
Curious case of sudden onset bilateral lower extremity weakness with T2 cord signal hyperintensity: What we know and what we don’t
Viswanath Aluru, Nidhi Purohit, Gollamudi Reddy
Ochsner Clinic Foundation, New Orleans, USA
Abstract
Introduction: 57-year-old man with a past medical history significant of extensive spine degenerative disease presented to ED with sudden onset bilateral lower extremity weakness, sensory loss, loss of bladder control. Patient was given steroids and closely monitored in neuro intensive unit. MRI brain was evident for linear T2 cord signal hyperintensity in the thoracic region. Lumbar puncture was evident for inflammation. A neuro surgical intervention was not needed at that time. Hospital course was complicated by a lower extremity deep venous thrombosis. Patient was medically stable but left with paraparesis, urinary incontinence, sensory and gait impairment.
Results: Upon admission to the neurorehabilitation unit, he was able to participate in intensive therapy. He made progress in terms of gait, balance, ADLs. Steroids were tapered. He was continued on anticoagulation for lower extremity thrombus. He continued to present with bilateral lower extremity weakness and urinary retention.
Discussion: We present a challenging case of an acute onset spinal cord syndrome associated with T2 cord linear hyperintensity in the thoracic region. Patient’s history of extensive spine degenerative disease might have resulted in a spinal ischemic attack due to foraminal stenosis related to postural changes.
However, we cannot rule out the possibility of an acute transverse myelitis syndrome or an acute inflammatory demyelinating process given clinical presentation and response to steroids. On the other hand, thoracic spinal cord with fewer anastomoses in the anterior region is at greater risk of ischemia/infarction which could have resulted in acute onset of symptoms.
Conclusion: Spinal cord syndrome along with cord signal changes need thorough evaluation to rule out underlying causes. MRI findings need to be corroborated along with clinical presentation and lab studies in final diagnosis and management. Follow-up imaging should be considered in all patients. Treatment should be directed at underlying condition to avoid unnecessary adverse treatment effects.
Categories
Neural Repair Mechanisms
Magnetic and electrical stimulation of the corticospinal pathway to assess residual connectivity in individuals with severe hemiparesis post-stroke: Preliminary results of a feasibility study
Mary Ellen Stoykov1,2, Carley Butler2, George F Wittenberg3,4, Carolee J Winstein5, Monica Perez1,2
1Shirley Ryan Abilitylab, Chicago, USA. 2Northwestern University, Chicago, USA. 3University of Pittsburgh, Pittsburgh, USA. 4VA Pittsburgh HS, Pittsburgh, USA. 5University of Southern California, Los Angeles, USA
Abstract
Upper limb hemiparesis is a common post-stroke disability. The neural mechanisms that enable individuals with stroke to regain control of their paretic limb remain poorly understood. Although there is broad consensus that presence or absence of a motor evoked potential (MEP) elicited by transcranial magnetic stimulation (TMS) over the motor cortex is a useful biomarker to predict functional recovery in individuals post-stroke,1-4 some studies have reported contrasting results. For example, MEP(-) and MEP(+) individuals post-stroke experienced clinically meaningful changes after an intervention.5 Also, Nazarova et al showed that when more than one muscle is tested in the same individual, some are classified as MEP(-) while others are MEP(+).6 Thus, there is a need to assess, in more detail, the link between MEP(-) status, descending connectivity, and functional recovery.
Recent evidence suggests that, compared to cortical MEPs, electrical stimulation directly on the corticospinal axons may be more sensitive to detection of residual connectivity in people with severe spinal cord injury.7 MEPs elicited by TMS over the primary motor cortex are inherently affected by the temporal dispersion of descending volleys including direct (D) and indirect (I) waves. Here, we propose to compare residual connectivity by using magnetic and electrical stimulation of the corticospinal pathway. Specifically, we will assess MEPs elicited at the cervicomedullary junction (known as CMEPs) as a potential biomarker for post-stroke recovery in individuals with severe UE hemiparesis.8 We hypothesize that individuals previously classified as MEP(-) will have a greater number of positive responses to cervicomedullary stimulation as compared to TMS.
Methods: We will test 10 chronic stroke survivors with severe UE hemiparesis (defined as an UE Fugl- Meyer9 score of ≤30) and categorized previously as MEP(-). Exclusion criteria include contraindications to TMS.10
MEP status of the affected biceps brachii (BB), extensor carpi radialis (ECR), and first dorsal interosseous (FDI) muscles is confirmed using single pulse TMS (Magstim, 200, UK). Although participants were previously categorized as MEP(-) in a specific muscle, we are confirming MEP status in 3 muscles during voluntary contraction. We are also recording maximal voluntary contractions.
To collect CMEPs, high-voltage electrical stimulation is passed between two small gold-cup electrodes placed behind the mastoid process at the cervicomedullary junction during voluntary contraction of affected upper limb muscles.
Results: All three subjects tested, categorized as MEP(-) by TMS in the BB, ECR, and FDI, showed the presence of CMEPs; an indication of residual descending connectivity in the corticospinal tract. Two subjects had CMEPs in all three muscles.
Discussion: To our knowledge, this is the first study that examines the presence of CMEPs in post-stroke individuals. Our preliminary data suggest that a CMEP can commonly detect residual descending connectivity, as compared with TMS. These findings have implications for post-stroke rehabilitation.
Categories
Neural Repair Mechanisms
Effect of Gamification with Social Incentives on Increasing Daily Steps after Stroke: A Randomized Clinical Trial
Kimberly Waddell1,2, Mitesh Patel1,3, Kayla Clark1, Tory Harrington1,4, S. Ryan Greysen1,2
1University of Pennsylvania, Philadelphia, USA. 2Crescenz VA Medical Center, Philadelphia, USA. 3Ascension Health, St. Louis, USA. 4Continuum Clinical, Philadelphia, USA
Abstract
Background: Regular physical activity post-stroke can help reduce morbidity, improve quality of life, and help prevent recurrent strokes but individuals are largely inactive after stroke. Most interventions to increase physical activity are delivered within a rehabilitation clinic and may luck sufficient support to help translate skills to the home environment. Interventions that leverage insights from behavioral economics (incentives, feedback) can help facilitate behavior change but have not been evaluated in adults with stroke.
Purpose: To evaluate the impact of a remote, gamification with social incentives intervention for increasing daily steps among community dwelling adults with stroke.
Methods: This was a two-arm, randomized clinical trial. Participants were randomized to a control or gamification with social incentives arm. Participants were issued a wrist-worn activity tracker (Fitbit Inspire 2), selected a step goal that was a 33%, 40%, or 50% increase above their baseline, and synced their daily step count to Way to Health, a remote monitoring platform. The control arm received device feedback only. The gamification arm participated in an 8-week game with loss-framed points and levels and received daily and weekly performance feedback. Participants in the gamification arm also selected a support partner from their social network who was informed of the participant’s game progress via a weekly email. The primary outcome was the difference in the change in mean daily steps from baseline between arms and was evaluated using a linear mixed effects regression model, adjusted for participant random effects, baseline steps, and calendar month. The secondary outcome was the difference in proportion of days participants achieved their step goal during the 8-week intervention, examined using a generalized linear mixed effects model, adjusted for participant random effect and calendar month, with 1000 bootstrap samples.
Results: Thirty-four individuals were randomized to the control (n=17) or gamification (n=17) arm. Overall, participants had a mean (SD) age 59 (15.3) and were 27.7 (58.7) months post-stroke.
Participants in the gamification with social incentives arm had a significantly greater change in mean daily steps from baseline (adjusted difference, 981 steps; [95% CI 201, 1762], P =0.01), compared to the control group. Participants in the gamification with social incentives arm also had a significantly higher adjusted proportion of days achieving their step goal (adjusted difference 0.41; [95% CI 0.38, 0.43], P < 0.001), compared to the control group.
Conclusions: A gamification with social incentives intervention was effective for increasing daily steps and step goal achievement among community-dwelling adults with stroke. This remote, individualized intervention incentivized individuals to increase daily physical activity in their home environment and required no in-person visits. Interventions that leverage behavioral economic principles can help individuals with stroke increase daily physical activity, which may mitigate post-stroke disability and improve longer-term outcomes.
Categories
Stroke
Wearable Myoelectric Interface for Neurorehabilitation (MINT) of Arm Function in Chronic Stroke
Abed Khorasani, Vivek Paul, Nathan Hung, Prashanth Prakash, Torin Kovach, Joel Hulsizer, Marc Slutzky
Northwestern University, Chicago, USA
Abstract
Abnormal muscle co-activation (also called abnormal synergies) is a significant cause of impaired arm movement after stroke. We have shown that a myoelectric computer interface (MyoCI) that provides feedback to stroke survivors about abnormal co-activation can enable reduction of abnormal co- activation and associated arm impairment. Here, we demonstrate a novel, wearable version of the MyoCI called the myoelectric interface for neurorehabilitation (MINT). MINT conditioning requires people with chronic stroke to control customized computer games with their muscles. The MINT device records surface EMGs from 2-3 arm muscles and wirelessly controls games on a laptop or tablet. By mapping the muscle activities to orthogonal directions of cursor movements, the games operantly condition the participants to learn to reduce abnormal co-activation between the muscles. MINT thus enables high- intensity, high-dose training using an inexpensive device at home. Here we report interim results of a randomized controlled trial of home-based MINT conditioning in severely impaired stroke survivors (Fugl- Meyer score <30). Participants were asked to train 90 min per day, 6 days a week, for a total of 6 weeks, with different variants of the paradigm (2 or 3 muscles at a time) vs. a sham control group that trained on one muscle at a time. We surveyed participants about their engagement with the game using a modified Intrinsic Motivation Inventory. To date, thirty-five participants have completed the 6-week training. They averaged 85±18 min of daily training and about 300 daily repetitions. More than 92% of participants surveyed found the training enjoyable and motivating. A few participants had significant difficulty with the laptops due to inexperience with computers. The primary functional outcome was the Wolf Motor Function Test (WMFT, timed portion). At 6 weeks, WMFT showed a trend of greater improvement relative to baseline in experimental groups than in sham controls (combined mean -3.6 vs -2.0 s, p=0.3, t-test). The group training 3 muscles at a time had a strong trend of greater improvement more than sham (-7.3 s, p=0.10) at 6 weeks. Improvement was even greater at 4 weeks after the training ended (-16.9 s in 3-muscle group, p=0.03; -9.5 s overall groups compared to -2.1 s for sham, p=0.03). MINT is a motivating way to provide home-based, high-intensity therapy to stroke survivors with a new mechanism of action. Early results suggest that it improves arm function and is not just overcoming learned non-use.
Categories
Stroke
Sing for your Saunter: Musical Cues to Improve Gait in People With Parkinson Disease With and Without Dementia
Lauren Tueth, Gammon Earhart, Elinor Harrison
Washington University School of Medicine, St Louis, USA
Abstract
Background: Parkinson disease (PD) is a neurodegenerative disorder with both motor and cognitive symptoms. Gait is impaired in most individuals with PD, leading to many investigations regarding methods for improving gait quality in PD. Our team previously demonstrated that external cues, such as listening to music, as well as internal cues, such as singing aloud or in one’s head, can be used to improve gait in PD. Both external and internal cues are associated with longer strides, and internal cues are associated with stable or reduced step-to-step variability as compared to external cues which are associated with increased step-to-step variability. Although both external and internal cueing methods can be effective for individuals with PD, less is known about how these cues impact gait in individuals with both Parkinson disease + dementia (PDD).
Methods: Our target sample is 23 individuals with a diagnosis of PDD, of which we have data from four individuals to report in this Abstract. Participants with PDD were matched based on age and UPDRS score to individuals with PD without dementia for comparison. APDM sensors were worn on the wrists, feet, sternum, and lumbar spine while performing all tasks. The gait tasks included: uncued walking, walking while listening to music (music) and walking while mentally singing (mental). Participants performed each of these for three trials of 30 seconds each. Tasks were performed at cue rates of 90%, 100%, 110%, and 120% of uncued walking cadence.
Results: Percent change between uncued gait and each cueing condition was calculated to determine ability to modulate gait. While individuals with PD (N=4) demonstrated increased velocity, stride length, and cadence at cue tempos of 110% and 120%, individuals with PDD (N=4) demonstrated much smaller increases in velocity and cadence and had decreased stride lengths for all cueing tempo and conditions except for mental 120%, which had only a 0.17% increase in stride length. For velocity, individuals with PD averaged a percent change of -3.46%, 4.49%, 12.48%, and 13.54% for mental cueing at tempos 90%, 100%, 110%, and 120% respectively. In contrast, for the same cues, individuals with PDD had changes of -6.59%, -0.74%, 0.54%, and 1.67%.
Conclusion: These preliminary data suggest that external cueing and internal cueing may have less impact on gait in PDD as compared to PD. The change to velocity appears to be smaller in individuals with PDD compared to those with PD, and individuals with PDD appear to take shorter strides even with higher tempo cues. These results should be interpreted with much caution, as a larger sample is needed.
Categories
Motor Rehabilitation
Virtual Reality Assessment of Arm Choice Under Cognitive Load
Cory Potts1, Shailesh Kantak1,2, Laurel Buxbaum1
1Moss Rehabilitation Research Institute, Jefferson University, Elkins Park, USA. 2Department of Physical Therapy, Arcadia University, Elkins Park, USA
Abstract
Background: A perplexing feature of motor disability after stroke is that 40-80% of individuals fail to use their paretic arm effectively despite having the capacity to do so1. The disparity between use and capacity is referred to as non-use and remains a central challenge in stroke rehabilitation. In this experiment we examined the contribution of attentional capacity to arm non-use. The capacity theory of attention posits that attention is a resource that is effortfully allocated to objects and actions to meet task demands2. Prior work has shown that under higher-demand conditions, neurotypical controls show more frequent use of the dominant arm even when to do so is inefficient (cross-midline reaches)3. Using a novel virtual reality (VR) task that mimics the demands of reaching to real-world objects in a cluttered array (e.g., a kitchen cabinet), we extended this work to individuals with stroke. We tested the hypothesis that despite having the motor capacity to reach with the paretic arm, stroke survivors would show reduced use of that arm under higher-load task conditions.
Method: 14 individuals with mild-to-moderate chronic stroke and 10 age-matched neurotypical controls reached for target objects in a VR array. In the Basic condition, the target belonged to a different semantic category than distractors (e.g., household target, animal distractors). In the Hard condition, the target and distractors were from the same category (e.g., household target and distractors). In the Dual- Task condition, the target and distractors were from the same category, and participants concurrently performed a secondary task. Accuracy, initiation time, trajectory, movement time, and arm choice were tracked by the VR system. In secondary analyses we examined relations between arm use and measures of sensorimotor and clinical function.
Results: Arm choice was analyzed with generalized linear mixed effects modeling. For individuals with stroke, an interaction emerged between Condition (Basic, Hard, Dual) and Target Side (Contralesional, Ipsilesional), χ2 (2) = 16.74, p< .001. Proportions of paretic arm use within the contralesional workspace decreased for higher load conditions (Basic = .68, Hard = .58, Dual = .60), indicating more frequent cross-midline reaches with the non-paretic arm. Results for controls replicated previous work on arm choice and cognitive load.
Discussion: With increased task demand, there was decreased use of the paretic arm for individuals with stroke and the non-dominant arm for controls as well as evidence of decreased efficiency of arm choice. These results support prior evidence that arm non-use is not merely a function of motor capacity, and the specific hypothesis that non-use is significantly influenced by attentional factors. These data thus highlight an important interaction of cognitive and motor factors in arm choice after stroke.
Categories
Stroke
Development of a rehabilitation data repository: the first step to creating a learning health system focused on precision rehabilitation
Margaret French1, Kelly Daley2, Preeti Raghavan1, Stephen Wegener3, Pablo Celnik1
1Johns Hopkins University, Baltimore, USA. 2Johns Hopkins Hospital, Baltimore, USA. 3Johns Hopkins Universirty, Baltimore, USA
Abstract
Purpose: Learning health systems are organizations that constantly adapt and modify clinical practice based on knowledge that is generated from clinical care. One key component of a learning health system is having easy access to systematically collected and documented data from the electronic health records (EHR) that can then be used to generate new knowledge that guides clinical care. Here we will describe the development of a rehabilitation data repository that contains EHR data from Johns Hopkins Medicine (JHM) for individuals with a history of stroke. This resource will be used to answer future research questions, thus, serving as the first step in establishing a learning health system that is focused on function at JHM.
Description: After IRB approval for the creation of the rehabilitation data repository, we defined the patient cohort for inclusion in the repository. Patients included in the cohort had to meet at least one of the following criteria: 1) admission to JHM by the Neurology Stroke Service, 2) treatment by the stroke specialty service line within outpatient rehabilitation services, or 3) documentation of an ICD-10 code associated with a stroke diagnosis during an inpatient or outpatient encounter with a Physical Medicine and Rehabilitation provider or therapist. In an iterative process with software engineers from the Core for Clinical Research Data Acquisition at JHM, we refined the list of ICD-10 codes and documentation location of these ICD-10 codes to improve the accuracy of the patient cohort. The final cohort includes 13,600 individuals with a history of stroke. Based on manual chart reviews, 82% of those included had a history of stroke, while 17% did not. We then identified the variables to be extracted from the EHR. We identified general information, such as demographics, medical history, billing information, and inpatient, outpatient, and emergency department encounters, to be included. Additionally, we partnered with rehabilitation informaticians to identify the location of rehabilitation specific data, including functional outcome measures, to be included in the rehabilitation repository.
Future Work: We are currently validating the data that was extracted to ensure the plausibility and accuracy of the extracted data. We will then create a data dictionary that contains metadata about the cohort and data included in the repository.
Summary of Use: Once completed, researchers and clinicians will to have access the data dictionary to determine what research questions can be answered with the repository. Once IRB approval is obtained for specific research questions, the data repository will be available to answer those questions and generate new knowledge to drive the function-focused learning health system at JHM.
Categories
Other
Feasibility and compliance of remote monitoring of physical, cognitive, and emotional function in individuals after stroke
Margaret French1, Junyao Li2, Ryan Roemmich1, Meghan Beier1, Peter Searson1, Stephen Wegener1, Pablo Celnik1, Preeti Raghavan1
1Johns Hopkins University, Baltimore USA. 2Johns Hopkins University, Baltimore, USA
Abstract
Purpose: Current assessments of physical, cognitive, and psychosocial function are made at clinical visits with measures that provide only a discrete snapshot of an individual’s function and low information content data. Remote monitoring may overcome these limitations by providing more frequent and more detailed information; however, feasibility of and compliance to remote monitoring is a potential barrier, particularly in individuals with neurologic conditions. Thus, the purpose of this pilot project is to examine the feasibility and compliance of individuals after stroke to a two month remote monitoring program for physical, cognitive, and psychosocial function.
Methods: Sixteen individuals (60.3 ± 11.3 yo; 4F) with history of stroke were instructed to wear a Fitbit daily and to sync at least every five days. If the device had not been synced in five days, they received an automatic reminder to do so. To assess feasibility and compliance to remote monitoring of physical function, we calculated the average number of reminders provided and the average percentage of the day the device was worn. Additionally, participants completed cognitive and psychosocial assessments at 0, 4, and 8 weeks. Participants had 7 days to complete the assessments, but were instructed to complete them as soon as possible. Additional reminders were provided automatically one and three days after the testing start date if the assessments were not completed. To assess feasibility and compliance to remotely monitoring cognitive and psychosocial function, we calculated the percentage of assessments completed and the average number of reminders required for each completed assessment.
Results: One participant withdrew from the study due to the time commitment of study and another was lost to follow up; thus, 14 participants were included in the analysis. Twenty-four reminders to sync the Fitbit were provided to eight different participants, resulting in an average of 1.7 ± 1.5 reminders per person during enrollment. On average, participants wore the Fitbit 92.1 ± 7.3% of the day during the study. Of the 42 assessments of cognitive and psychosocial function that were initiated, 33 (78%) were completed. For each completed assessment, an average of 1.9 ± .9 reminders were provided.
Conclusion: Individuals after stroke wore the Fitbit for a large portion of the day and a majority of the cognitive and psychosocial assessments were completed. However, a number of reminders were required to ensure the device was synced and the assessments were completed. The frequency of reminders required to reach this level of compliance would hinder the feasibility of remote monitoring if the system had not been automated due to the human resources required. Thus, automatic systems would facilitate the feasibility of remote monitoring in larger groups of individuals.
Categories
Stroke
Higher amyloid correlates to greater loneliness during the COVID-19 pandemic
Abigail Kehrer-Dunlap, Rebecca Bollinger, Beau Ances, Susan Stark
Washington University in St. Louis, St. Louis, MO, USA
Abstract
Background: Loneliness is a growing epidemic for millions of older adults and is associated with poor physical health, depression, anxiety, mortality, and an increased risk of Alzheimer disease (AD).
Psychosocial characteristics such as loneliness, anxiety, and depression are associated with symptomatic AD, but it is unclear whether these emerge during the preclinical phase of AD. The COVID-19 pandemic imposed sudden constraints on social interactions for older adults, with many older adults reporting loneliness, anxiety, and depression. It is imperative to understand how at-risk groups, including individuals with preclinical AD, may be impacted by prolonged feelings of loneliness throughout the COVID-19 pandemic. The purpose of this study was to examine the relationship between amyloid accumulation and psychosocial characteristics of loneliness, anxiety, and depression in cognitively normal older adults with and without preclinical AD during the COVID-19 pandemic.
Methods: Participants in this cross-sectional study were community-dwelling older adults recruited from an ongoing longitudinal cohort study at the Knight Alzheimer Disease Research Center in St. Louis, MO. A global Clinical Dementia Rating® Scale score of 0 (i.e., cognitively normal) was required for enrollment in this study. Cortical amyloid burden was measured through using PiB and AV45 PET tracers, and centiloids were used to synchronize measures. Demographic characteristics and six additional questions about personal and social behaviors were collected via self-report. Loneliness and feelings of depression were measured using the 20-item UCLA Loneliness Scale and Patient Health Questionnaire 9, respectively.
Anxiety was measured using the Patient-Reported Outcomes Measurement Information System Emotional Distress–Anxiety short form 4a and the Hospital Anxiety and Depression–Anxiety subscale. Spearman correlation was used to examine the relationships between amyloid and psychosocial characteristics.
Results: The 108 participants had a mean age of 75.0 ± 5.5 and average amyloid burden, measured by centiloid, of 22.2. Consistent with previous findings, higher amyloid accumulation was significantly associated with greater loneliness. Depression and anxiety were not significantly associated with higher amyloid accumulation.
Conclusions: Understanding the relationships of psychosocial characteristics, such as loneliness, anxiety, and depression, to amyloid accumulation in cognitively normal older adults can aid in better understanding the preclinical phase of AD. As the physical and psychosocial effects of the COVID-19 pandemic will persist far beyond the conclusion of the pandemic, it is imperative to continue examining psychosocial symptoms and the impact of prolonged loneliness on older adults with and without preclinical AD. Additional research is needed with a larger, more diverse sample to examine these psychosocial symptoms in preclinical AD.
Categories
Other
Motor control and cognitive deficits impact gait coordination in individuals with stroke
Prakruti Patel, Neha Lodha
Colorado State University, Fort Collins, USA
Abstract
Background: Safe mobility relies on rhythmic, anti-phase coordination between lower limbs during walking. Despite regained ability to walk independently, deficits in gait coordination often persist after stroke, and increase the risk for falls. Identifying stroke-specific impairments that affect gait coordination is essential for developing targeted interventions for safe mobility. Recent evidence highlights that poor mobility outcomes after stroke are related to impaired force modulation and attention; however, whether these impairments impact gait coordination is not known. Thus, the aim of our study was to determine the contribution of bilateral force coordination and selective attention to gait coordination in stroke survivors.
Methods: Chronic stroke survivors (N = 14; 66.31 ± 13.04 years) and controls (N=11; 66.87 ± 11.13 years) performed - 1) overground walking at preferred speed, 2) bilateral ankle force tracking task that required anti-phase coordination between isometric ankle dorsiflexion forces generated with each leg, and 3) visual search task involving a rapid response to a target stimulus embedded within distractors. We measured gait coordination during overground walking with phase coordination index (PCI) such that higher PCI values indicated poor gait coordination. We measured cross-correlation coefficient and time lag between two legs in bilateral force tracking task. Positive values for cross- correlation coefficient and higher values for time lag indicated poor bilateral force coordination. We measured response time on the visual search task. Longer response times indicated poor selective attention.
Results: The stroke group showed increased PCI (p = 0.01) compared with the control group. The stroke group had less negative cross-correlation coefficient (p = 0.001) and increased time lag (p =0.05) between bilateral forces. The stroke group demonstrated longer response times than controls (p = 0.01) during visual search. In the stroke group, PCI was positively correlated with cross-correlation coefficient (r = 0.57, p = 0.02) and the response time (r = 0.53, p = 0.03). For all participants, hierarchical multiple regression was performed with PCI as dependent variable. Cross-correlation coefficient entered at stage 1 accounted for 40.9% variance in PCI (p < 0.01). Adding response time to the model explained an additional 10.5% of variance in PCI (p < 0.05). In the final model, we included group (stroke or control) as an independent variable and found that group was not a significant predictor of PCI (p > 0.05). Thus, cross-correlation coefficient and response time together predicted PCI in the stroke and control groups.
Conclusion: Poor gait coordination in stroke survivors is associated with decline in both motor and cognitive abilities. Specifically, impairments in bilateral force coordination and selective attention affect gait coordination. Improving the ability to coordinate forces and attention may improve gait coordination and thereby, enhance safe mobility after stroke.
Categories
Stroke
Identifying racial and ethnic outcome disparities after discharge from acute inpatient rehabilitation
Amanda Herrmann1,2, Ella Chrenka1,2, Marny Farrell1,3, Leah Hanson1,2, Steven Jackson1,2,3
1HealthPartners Neuroscience Center, St. Paul, USA. 2HealthPartners Institute, Bloomingon, USA. 3Regions Hospital, St. Paul, USA
Abstract
Background: Over the last couple decades, racial and linguistic disparities have been reported for acute rehabilitation outcomes. The goal of this retrospective study was to determine whether disparities exist in our inpatient rehabilitation facility (IRF) outcomes (function independence measure (FIM) scores and discharge destination) for white vs. non-white patients and English speaking vs. Limited English Proficiency (LEP) patients.
Methods: Data were exported from the Uniform Data System of Medical Rehabilitation, a national rehabilitation outcomes measurement program in which our Commission on Accreditation of Rehabilitation Facilities (CARF) accredited IRF participates. We exported data for all patients ≥18 years who discharged from our IRF from 2013-2019. Patients were categorized by race (white or non-white) and language (English vs. LEP). Change in FIM score and discharge destination were compared between groups using appropriate statistical tests (t-test, chi-square test, ANCOVA). All analyses were performed in SAS 9.4 using two-sided p-values of 0.05.
Results: We identified 2,518 patients ≥18 years who discharged from our IRF from 2013-2019. Of these, 528 (21.0%) identified as non-white and 179 (7.1%) reported a primary language other than English.
Patient race was associated with change in FIM, with non-white patients estimated to have a 1.9 smaller change in FIM score (0.3 to 3.5, p = 0.02) compared to white patients after adjustment for possible confounders. Shorter length of stay, older age, a non-stroke diagnosis, and experiencing a complication while in IRF were also associated with a smaller improvement in FIM score. Similarly, LEP patients were expected to have a 3-point smaller change in FIM score compared to English speaking patients (3.1 (0.7, 5.5), p = 0.01).
Discharge destination significantly differed between white and non-white patients (p<.001). White patients were more often discharged to a skilled nursing facility (white: 24%, non-white:18%) and non- white patients were more often discharged with home health services (white: 8%, non-white: 14%).
When stratified by health insurance, this racial disparity persisted in patients who received Medicare or Medical Assistance (p <.001). For patients with commercial insurance there was no difference in discharge destination associated with race (p = 0.86). A similar pattern was identified when comparing English and LEP patients, with LEP patients more likely to be discharged with home health services (English: 9%, LEP: 20%, p<.001).
Conclusion: Overall, we found that racial and linguistic disparities in rehabilitation outcomes (FIM and discharge destination) exist within our CARF certified IRF; however, the results should be cautiously interpreted due to the relatively low representation of minority patients and the complexities of rehabilitation. Regardless, our organization is committed to reducing health care disparities by raising awareness, identifying gaps in care, and creating innovative solutions. The next step will be to explore opportunities for prospective interventions aimed at reducing health disparities.
Categories
Other
Comparing the accuracy of open-source pose estimation methods for measuring gait kinematics
Edward Washabaugh1,2, Thanikai Adhithiyan Shanmugam2, Rajiv Ranganathan3, Chandramouli Krishnan2,4
1Wayne State University, Detroit, USA. 2Michigan Medicine, Ann Arbor, USA. 3Michigan State University, East Lansing, USA. 4Michigan Robotics Institute, Ann Arbor, USA
Abstract
Background: Open-source pose estimation is rapidly reducing the costs associated with motion capture. Where specialized cameras, markers, and software packages were once needed to perform accurate motion capture, machine learning now permits this analysis to be performed with a single camera. This technology could be particularly valuable for clinical gait analysis, which is often performed qualitatively due to the prohibitive cost and set-up required for conventional motion capture.
Research Question: How do open-source pose estimation software packages compare in their ability to measure kinematics during gait analysis?
Methods: An existing dataset that contained video and synchronous motion capture data from 32 able- bodied participants while walking was used in this analysis. Sagittal plane videos were analyzed with four open-source pose estimation methods—OpenPose, Tensorflow MoveNet Lightning, Tensorflow MoveNet Thunder, and DeepLabCut—to extract keypoints (i.e., landmarks) and calculate hip and knee kinematics. The absolute error when using each markerless pose estimation method was computed against conventional marker-based optical motion capture. Errors were compared between pose estimation methods over the gait cycle using statistical parametric mapping.
Results: Pose estimation methods differed in their ability to measure kinematics. OpenPose and Tensorflow MoveNet Thunder methods were most accurate for measuring hip kinematics, averaging 3.7± 1.3 deg and 4.6 ± 1.8 deg (mean ± std) over the entire gait cycle, respectively. When measuring knee kinematics, OpenPose was most accurate, averaging 5.1 ± 2.5 deg of error over the gait cycle.
Significance: Pose estimation technology could fundamentally change how gait analysis is performed in the clinic. This is the first study to systematically compare open-source pose estimation software packages on their ability to measure gait kinematics. While this technology will continue to improve, OpenPose currently an appealing and accurate option for clinical gait analysis.
Categories
Motor Rehabilitation
Towards individualized Transcranial Magnetic Stimulation for motor recovery from hemiparesis: study of Corticomuscular Network
Gansheng Tan1,2, Jixian Wang1, Jinbiao Liu1, Yixuan Sheng1, Qing Xie1, Peter Brunner2, Honghai Liu3
1Shanghai Jiao Tong University, Shanghai, China. 2Washington University in St. Louis, St. Louis, USA. 3Harbin Institute of Technology (Shenzhen), Shenzhen, China
Abstract
Background and aims: Hemiparesis affects more than 60% of chronic stroke patients. It limits patients’ daily activities and reduces quality of life. Transcranial Magnetic Stimulation (TMS) opens new and exciting avenues to improving motor function in patients affected by stroke. However, effectiveness of TMS in post-stroke motor recovery varies substantially across patients. Individualization of stimulation parameters and using repetitive TMS (rTMS) are thought to overcome this limitation and improve the therapeutic potential. The first step towards this promise is to develop a mechanistic understanding that links post- stroke reorganization, motor recovery, and effects of rTMS. Working towards this goal, we propose the concept of a Corticomuscular Network, based on which we investigated the neural plasticity after stroke, neural correlates of motor recovery, and the effect of rTMS.
Methods: We acquired electroencephalography (EEG) and electromyography (EMG) from 21 patients with hemiparesis while they performed a robotic-assisted isokinetic push-pull task before and after rTMS treatment and from 7 healthy controls doing the same task. The patients were randomly allocated to one of three groups: HF-rTMS, LF-rTMS, and Sham. HF-rTMS group received 15min of 10Hz rTMS targeting the ipsilesional primary motor cortex with 10 pulses per train and inter-train intervals of 1s. LF-rTMS group received 15min of 1Hz rTMS intervention totaling 900 pulses that targeted contralesional primary motor cortex. LF-rTMS protocol was applied to Sham group except the coil was tilted to be perpendicular to the skull. Standardized permutation mutual information (SPMI) was used to quantify Functional brain connectivity (FBC), corticomuscular coherence (CMC), intermuscular coherence. CMN was constructed based on coherence with electrodes as nodes and coherence as the weight of an edge connecting two nodes. Graph theory-based features of CMN including small-worldness, assortativity, and centrality were analyzed.
Results: Our analyses revealed that patients manifested higher CMC (p<0.01) between the sensorimotor cortex and upper limb muscles compared to healthy controls. We found coherence and centrality CMN to increase between pull and push phase in patients, but not in healthy controls. Further, we found FBC between ipsilesional motor cortex and other brain areas after stroke (p<0.01) to be increased and negatively correlated (p<0.01, ) with Upper Limb Fugl-Meyer Assessment. 1Hz rTMS decreased the centrality of the stimulated site and inhibited CMC (p<0.01) while 10Hz rTMS induced mixed changes across the CMN. Interestingly, only a small proportion of effects of both rTMS protocols related to motor recovery or post-stroke plasticity.
Conclusion: This study sheds light on relationship between post-stroke nervous system reorganizations, motor recovery, and effects of rTMS. Although CMC serves as a biomarker of stroke, only FBC between the ipsilesional frontal cortex and other brain areas were indicative of motor recovery. This suggests that restitution of FBC should be the target of rTMS treatment.
Categories
Stroke
Polarity dependent effects of bi-hemispheric tDCS when paired with contralaterally controlled functional electrical stimulation (CCFES) on chronic post stroke corticospinal output: A TMS study
David A. Cunningham1,2, Kevin H. Cheng1,2, Amy Friedl2, Ela B. Plow3, Kenneth B. Baker3, Richard D. Wilson1,2, Jayme S. Knutson1,2
1Case Western Reserve University, Cleveland, USA. 2MetroHealth Center for Rehabilitation Research, Cleveland, USA. 3Cleveland Clinic, Cleveland, USA
Abstract
Introduction: Post-stroke loss of upper-limb function is typically characterized by difficulty opening the paretic hand due to paresis of the finger extensors. CCFES is an electrical stimulation therapy that gives the patient control of both timing and intensity of stimulation to their finger and thumb extensors and thereby enables intention-driven hand opening. We propose that non-invasive brain stimulation using tDCS will work in synergy with CCFES to improve dexterity and reduce upper extremity impairment compared to CCFES alone. A conventional bihemispheric-tDCS montage involves exciting the ipsilesional motor network (i.e. anodal stimulation) while inhibiting the contralesional motor network (i.e. cathodal stimulation). Because the CCFES control strategy requires volitional motor activity, combining conventional-tDCS with CCFES may evoke greater motor rehabilitation gains than CCFES alone.
Alternatively, there may be greater therapeutic benefit if tDCS is applied in an unconventional montage, i.e., inhibiting the ipsilesional while exciting the contralesional motor networks. This approach is supported by homeostatic metaplasticity models, which emphasize that brain networks are self- protective and will prevent over-excitability and, thus, predict that the combination of two excitatory neuromodulatory interventions will limit the capacity for improvement.
Objective: The purpose of this preliminary study is to compare the effects of conventional tDCS and unconventional tDCS when paired with CCFES-assisted finger tracking on ipsilesional corticospinal output.
Methods: Seven mild-to-moderately impaired chronic stroke survivors (>6 months) participated in a randomized crossover design consisting of tDCS (conventional and unconventional) delivered during and before CCFES-assisted finger tracking. All sessions were separated by at least one week. Within each session, TMS assessments of corticospinal output (I/O curves) were made before and 45 minutes after 40 minutes of 1) conventional bihemispheric tDCS + CCFES 2) unconventional bihemispheric tDCS + CCFES, and 3) sham tDCS + CCFES. All active sessions were randomized to receive either active or sham stimulation before and during CCFES-assisted finger tracking (e.g. active-tDCS (20 min) followed by sham-tDCS during CCFES (20 min)). Sham + CCFES consisted of sham-tDCS delivered before and during CCFES.
Results: We found an overall benefit of delivering unconventional-tDCS during CCFES (F = 9.02, p =.004). Twenty minutes of unconventional-tDCS + CCFES produced a greater increase in ipsilesional corticospinal output to the paretic hand than conventional-tDCS + CCFES (t = 2.82, p=.03) or sham tDCS+ CCFES (t= -4.28, p=.005). Relative to sham stimulation, there was a 29±6.7% increase in ipsilesional output following unconventional-tDCS + CCFES, with no change following conventional-tDCS + CCFES. We found no added benefit with either the conventional or the unconventional-tDCS montage when delivered before CCFES (F=.31, p=.74).
Conclusion: These preliminary findings suggest that unconventional-tDCS during functional electrical stimulation may promote motor training-induced plasticity. Future work will investigate neurophysiologic mechanisms and extend these findings to a multi-session randomized clinical trial.
Categories
Stroke
Transcallosal Inhibition in hand and arm muscles of chronic stroke and healthy controls
Leticia Hayes1, Myriam Taga1, Charalambos Charalambous2,3, Sharmila Raju1, Jing Lin1, Elisa Stern1, Heidi Schambra1
1Department of Neurology, NYU Langone, School of Medicine, New York, USA. 2Department of Basic and Clinical Sciences, Medical School, University of Nicosia, Nicosia, Cyprus. 3Center for Neuroscience and Integrative Brain Research (CENIBRE), Medical School, University of Nicosia, Nicosia, Cyprus
Abstract
Background: Interhemispheric inhibition is a normal neurophysiological process, potentially important for precision unilateral movement and suppression of bilateral movements. Transcranial Magnetic Stimulation (TMS) can be used to examine this inhibition, generating an ipsilateral silent period (iSP) in a contracting muscle. iSP reflects inhibition between active motor areas and their motor homologues in the opposite hemisphere. Previous studies have found stronger inhibition in distal than proximal upper extremity muscles, but it is unknown if this differential inhibition is altered after stroke. Here, we examined this differential inhibition in both subject groups using TMS to probe an arm muscle (biceps; BIC) and intrinsic hand muscle (first dorsal interosseous; FDI).
Methods: In this cross-sectional observational study, we examined 15 healthy subjects (7 females; mean age: 54 (44-81) years; mean Fugl-Meyer Assessment (FMA) score: 65 (63-66)) and 16 chronic stroke subjects (10 females; mean age 62 (44-85) years; mean UE FMA score: 49 (23-64); mean time since stroke: 5 (0.5- 14.4) years). Using each subject’s structural MRI, we placed a 5 × 5 virtual grid (25 sites) over primary and secondary motor areas of their contralesional hemisphere (assigned in healthy controls). We applied single-pulse TMS at 100% maximum stimulator output on each grid site during paretic BIC or FDI activation. At each grid site, we measured iSP strength (mV*ms) as the area of EMG suppression normalized to baseline EMG. We first identified the iSP hotspot for each muscle, operationalized as the grid location with the strongest iSP. We used linear regression and post hoc student’s t-tests to examine differences in iSP strength between muscles, subject groups, and their interaction.
Results: iSP was present in nearly all healthy subjects (BIC: 14/15, FDI: 15/15) and in most stroke subjects (BIC: 11/16, FDI: 15/16). iSP hotspots were predominantly located in the premotor cortex for both healthy and stroke subjects and rarely overlapped for both muscles. In healthy and stroke subjects, inhibition was stronger in FDI than BIC (healthy p < 0.0001; stroke p < 0.0001). In FDIs, inhibition was comparable for both subject groups, but in BICs inhibition showed a trend being stronger in patients (p = 0.065). The difference in inhibition between muscles was different for the subjects groups, being less distinct in patients than controls (muscle*group; p = 0.024).
Conclusion: In both healthy and stroke subjects, we confirmed that transcallosal inhibition from the contralesional hemisphere is stronger for a distal than proximal muscle. We newly show that this differential inhibition was less striking in stroke subjects, whose BICs showed strong inhibition alongside their FDIs. These findings suggest that following stroke, transcallosal inhibition loses some muscle-specificity. The behavioral consequence of this change is not clear.
Categories
Stroke
The Impact of SSRIs on Motor and Visual Recovery in Stroke Patients Undergoing BCI Intervention
Anthony Bui1, Alexander Remsik2, Vivek Prabhakaran2
1University of Wisconsin School of Medicine and Public Health, Madison, USA. 2University of Wisconsin School of Medicine and Public Health - Department of Radiology, Madison, USA
Abstract
Introduction: Evidence from past research suggests that pharmaceutically increased serotonin levels in the brain during BCI interventions may promote recovery in stroke survivors with acquired motor disability.
Methods: SIS subdomains (Mobility, ADL, Hand Function), NIHSS subdomains (Best Gaze, Visual, Facial Palsy, Motor Arm, Motor Leg, Limb Ataxia), and 9-Hole Peg Tests (affected and unaffected arm) were used to analyze whether there was greater motor or visual recovery in BCI intervention participants who had taken SSRIs compared to those who had not. Data was acquired from 58 stroke survivors. The average age of participants who had taken SSRIs within 90 days post-stroke or during BCI treatment was 57.36 (n = 18). The average age of participants who did not take SSRIs was 58.12 (n = 40). Effectiveness of recovery was calculated using (Completion - Baseline)/Baseline for each subdomain and compared between the SSRI versus non-SSRI cohorts.
Results: Although there were some mean differences in functional recovery between the SSRI and non-SSRI cohorts after BCI intervention, these differences were not statistically significant. Average recovery for SIS Mobility with SSRI was .41; without SSRI was .14 (p = .52). Average for SIS ADL with SSRI was .11; without SSRI was .096 (p = .91). Average for SIS Hand Function with SSRI was 8.3; without SSRI was 2.75 (p = .25). Average for NIHSS Best Gaze with SSRI was .056; without SSRI was -.025 (p = .13). Average for NIHSS Visual with SSRI was -.056; without SSRI was .026 (p = .51). Average for NIHSS Facial Palsy with SSRI was -.056; without SSRI was -.13 (p = .50). Average for NIHSS Motor Arm with SSRI was -.22; without SSRI was .025 (p = .17). Average for NIH Motor Leg with SSRI was 0; without SSRI was .025 (p = .77). Average for NIHSS Limb Ataxia with SSRI was -.11; without SSRI was -.025 (p = .56). Average for NIHSS Sensory with SSRI was .11; without SSRI was -.025 (p = .17). Average for 9HPT affected arm with SSRI was -.073; without SSRI was -.12 (p = .65). Average for 9HPT unaffected arm with SSRI was -.12; without SSRI was -.11 (p = .82).
Discussion: For stroke survivors who undergo BCI intervention, there seems to be no significant difference in restoration of motor or visual function with SSRIs versus without SSRIs. As such, this data cannot entirely support the findings of the FLAME trial that SSRI administration and subsequent increased serotonin during motor intervention will result in greater functional recovery. Some group mean differences suggest that administration of SSRIs during BCI intervention may increase or decrease recovery potential for stroke survivors; however, more research is needed to understand this relationship before becoming clinically relevant.
Categories
Motor Rehabilitation
Role of the Vestibular System in the control of locomotion
Carl Tchoumi1, Mindy Levin1, Anatol Feldman2
1McGill University, Montreal, Canada. 2Université de Montreal, Montreal, Canada
Abstract
Background: Fundamental questions in the field of motor control concern the understanding of how human movements are produced and controlled, and determining what systems in the brain are involved in this control. In our research project, we will investigate the role of the vestibular system (VS) on the control of lower limb movements at different phases of gait. We will compare the traditional view of direct programming of motor output by the brain to an alternate view based on the physical principle that the brain controls movements by changing physical systems called the referent configuration of the body. hypothesis. According to this e referent control view, discrepancies between the actual body configuration and referent body configuration lead to muscle activation that enables the actual body configuration to move towards the referent body configuration and hence produce movement from one location to another. Also, that the actual and referent leg postures can match each other at specific locomotor phases and that suggest that referent (R) postures are shifted forwards in space to produce locomotion. Phases at which both actual and R postures match are marked by minimized EMG activity in multiple muscles, which are called, EMG minima.
Methods: Healthy subjects and stroke subjects will walk form a starting point to a designated position at a cadence of 110 steps/min. The participants will be required to take 8-10 steps, with GVS stimulation applied from the 6th to the 10th step. Five optotrak markers will be placed on five bony landmarks, 10 EMG electrodes will be placed on calf muscles of both legs, as well as 2 accelerometers in order to observe leg movements. There will be three conditions involving different head positions, one with head facing forward, and the other two with the head either facing left or right to produce a slight deviation with GVS stimulation. Each condition will have 10 trials.
Expected Outcomes: Changes in motor evoked potentials (MEPs) evoked by GVS stimulation will be used to evaluate the influence of the VS at different phases of the gait cycle. We will also compare the phase shifts in motion between instances with and without GVS.
Implications: Our findings will help to choose between the two alternative frameworks in the analysis of human gait provide support for the referent configuration hypothesis, which will helping to improve our understanding about how human locomotion occurs and the roles played by the VS in the control of locomotion and understanding how the vestibular system might play a role in certain disorders of locomotion. In the future, we might explore the clinical implications of these findings in identifying or treating disorders in locomotion.
Categories
Other
Sensitivity to change and responsiveness of the upper-extremity Fugl-Meyer in individuals with acute stroke
Baothy Huynh1, David Lin2, Julie DiCarlo2, Teresa Kimberley1, Perman Gochyyev1, Jessica Ranford2
1MGH Institute of Health Professions, Boston, USA. 2Massachusetts General Hospital, Boston, USA
Abstract
The upper extremity (UE) section of the Fugl-Meyer Assessment (FMA-UE) is one of the most widely used outcome measures in stroke rehabilitation and is often used to stratify research participants into stroke severity groups, predict functional outcomes, and assess motor impairment and function. Sensitivity is the ability of a measure to detect change that exceeds what can be attributed to error, regardless of if the change is considered relevant or clinically meaningful. Responsiveness is the detection of change that is clinically meaningful from a patient, provider, or researcher’s perspective. Responsiveness is commonly reported through the minimally clinically important difference (MCID). The MCID is the smallest amount of change that is considered clinically meaningful and reveals important insight into the clinical utility of interventions. Current estimates of MCID for the FMA-UE have varying levels of rigor, do not stratify by baseline severity, or are generalizable only to the chronic stroke population. The purpose of this research was to assess sensitivity and responsiveness of the FMA-UE using distribution-and anchor-based methods in acute post-stroke individuals and to examine if these estimates vary by baseline stroke severity. This present study was a secondary analysis of data from the Stroke Motor Rehabilitation and Recovery Study (SMaHRT), a natural history longitudinal study of UE motor recovery after acute stroke. Patients were recruited from the inpatient stroke service at Massachusetts General Hospital. Eighty-two participants with mild, moderate, or severe UE hemiparesis were administered the FMA-UE during their acute stroke hospitalization and again at 6-weeks follow-up. Participants rated their perceived amount of UE recovery on a global rating of change scale (GROC) and clinicians rated participant’s status on the modified Rankin Scale (mRS), a measure of global disability. Individuals were stratified by stroke severity using baseline FMA-UE scores. Sensitivity to change was assessed using Cohen’s effect size (ES), standardized response means, standard error of measure, and minimal detectable change scores. MCID of the FMA-UE were estimated using receiver operating characteristic curve analysis with the GROC and the mRS as anchors. Total sample analysis revealed moderate sensitivity to change (ES, d = 0.5). However, stratified subgroup analysis revealed high sensitivity to change for mild (ES, d = 1.14), moderate (ES, d = 3.10), and severe (ES, d = 1.11) groups. Total sample and all subgroups demonstrated statistically significant changes from baseline to 6-weeks. The estimated MCID of the FMA-UE for acute stroke patients using the total sample were 13 and 9 anchored to the GROC and mRS, respectively. This suggests that the MCID in the acute stroke population is higher than previously reported values for chronic stroke. The reported sensitivity and MCID estimates provide guidance for both clinicians and researchers to assess and interpret clinically relevant changes following treatment.
Categories
Stroke
Efficacy of Corsi Block Tapping Task as a viable visuospatial training approach: A proof-of-concept
Sydney Schaefer, Andrew Hooyman, Nicole Haikalis, Randy Essikpe, Peiyuan Wang
Arizona State University, Tempe, USA
Abstract
Although there is a growing industry for cognitive training (e.g., Lumosity, CogniFit), there is substantial debate about the efficacy of such training. Current scientific evidence suggests that global cognition, along with certain aspects of cognition like working memory, reasoning, and verbal abilities, do not necessarily improve following various cognitive training programs. Very little research has been done, however, on whether and how visuospatial abilities can be improved, above and beyond severe hemineglect post-brain injury. Furthermore, recent work posits that enhancing visuospatial abilities, if possible, could have positive downstream effects on motor rehabilitation. Thus, the purpose of this proof- of-concept study was to test whether a single session of training on a computerized version of the Corsi Block Tapping Task (CBTT, a spatial working memory task) was efficacious in improving visuospatial ability. Fifty-three young adults were assigned to one of two groups: a control group (mean age = 21.4; 10 females) or a training group (mean age = 21.5; 17 females). Visuospatial ability was measured using a computerized version of the Shepard-Metzler Mental Rotation test (reaction time and accuracy) and was assessed before and after a 20-minute training session of the CBTT for the training group. The control group was simply assessed on the Mental Rotation test two times, separated by 20 minutes of rest. Both tasks were assessed using open-source programs within the Psychology Experiment Building Language (PEBL). Results showed that CBTT performance improved with practice, indicating that the training task is learnable. Results also demonstrated a significant interaction between time (pre vs. post) and group (control vs. training) on mental rotation performance (p = .04), with the training group performing on average 123 ms faster on accurate trials than the control group at post-test. This proof-of- concept study suggests that improving visuospatial ability may be feasible through targeted (rather than global) cognitive training. We acknowledge that this study was not randomized, as the groups were collected at two different sites. Future studies will investigate various doses of CBTT training, as well as longer-term retention in cognitively-intact and -impaired individuals.
Categories
Cognitive/Language Rehabilitation
A Novel Trunk-based Index of Performance as a Biomarker of Upper Limb Motor Impairment in Stroke
Daniele Piscitelli1,2, Melanie C. Baniña1,2, Timothy K. Lam3, Kay-Ann Allen3, Joyce L. Chen3,4, Mindy F. Levin1,2
1School of Physical and Occupational Therapy, McGill University, Montreal, Canada. 2Feil/Oberfeld Jewish Rehabilitation Hospital/CRIR Research Centre, Laval, Canada. 3Hurvitz Brain Sciences Research Program, Sunnybrook Research Institute, Toronto, Canada. 4Faculty of Kinesiology and Physical Education, University of Toronto, Toronto, Canada
Abstract
Introduction: Several upper limb (UL) motor tasks have been developed to identify prognostic biomarkers to predict recovery and to detect intervention effects after rehabilitation treatment. Recovery has been investigated using separate kinematic variables describing features of the endpoint performance (i.e., accuracy and speed) or the quality of joint and body segment (i.e., trunk displacement) movements. However, UL performance and movement quality measures in isolation may not provide a true profile of functional recovery. Indeed, endpoint performance is influenced by compensatory use of the trunk to assist with endpoint displacement and accuracy. This current study investigates the validity of characterizing motor performance of a functional UL task using a combination of endpoint and trunk kinematic data reflecting movement quality.
Methods: This is a nested-cohort study within a multi-center double-blind randomized clinical trial. Twenty-five chronic stroke individuals were recruited. At three baseline evaluations, separated by at least 2 weeks, participants performed a reaching task to a target placed at arm’s length with their more affected UL. Trunk and UL kinematic data were recorded with a 3D Optotrak system. A novel index of performance (IPt) was computed based on Fitts’ Law that incorporated accuracy and speed of endpoint movements (i.e., motor performance) and corrected by the amount of sagittal trunk displacement (i.e., movement quality). UL motor impairment was evaluated with the Fugl-Meyer Assessment (FMA), while UL activity was assessed with the Action Research Arm Test (ARAT). Receiver operating characteristic curve (ROC) and the area under the ROC curve (AUC) were used to determine the IPt predictive ability for distinguishing between levels of UL motor impairment severity (mild to severe, cut-off based on FMA < 50/66.). Regression analysis was used to investigate the relationship between IPt and the UL clinical scales (FMA and ARAT).
Results: The ROCs and AUC demonstrated the ability of the IPt to distinguish between UL levels of impairment. For each of the three baseline evaluations, the AUC was 0.792 (p=0.002), 0.848 (p<0.001), 0.800 (p=0.002). Preliminary results showed that IPt explained 42% (R2 = 0.43, p=0.001) and 31% (R2=0.31, p=0.005) of the variance in the FMA and ARAT, respectively.
Conclusion: The IPt based on a reaching motor task incorporating movement quality and performance is a promising tool that could be applied in research and clinical settings. These preliminary results highlight how kinematic variables may be used in conjunction to develop a potential biomarker of sensorimotor UL impairments. Our findings encourage further studies of the reliability and predictive power of IPt in stroke survivors.
Categories
Motor Rehabilitation
Motor Cortical Map Reorganization in Persons with Cervical Spinal Cord Injury (SCI) is Related to Upper Limb Prehension Capability
Jia Liu1, Tarun Arora2, Kyle O’Laughlin1, Gail Forrest3, Svetlana Pundik4, Kevin Kilgore5, Anne Bryden5, Steven Kirshblum3, Ela Plow1
1Cleveland Clinic, Cleveland, USA. 2University Health Network, Toronto, Canada. 3Kessler Foundation, West Orange, USA. 4Louis Stokes Cleveland VA Medical Center, Cleveland, USA. 5MetroHealth System, Cleveland, USA
Abstract
Purpose: In survivors with chronic cervical SCI, we investigated (1) the characteristics of motor cortical maps devoted to less affected and more affected upper limb muscles and (2) the relationship between motor cortical map characteristics and upper limb motor impairment.
Materials and Methods: Fifteen individuals with C1-C8 SCI (AIS A-D) and 6 age-matched able-bodied controls underwent motor cortical mapping for biceps (C5) and triceps (C7) using navigated transcranial magnetic stimulation (TMS). Scalp sites on a 7-by-7 grid (1cm2 resolution) centered over the biceps hotspot were targeted at 100% maximal stimulation output intensity with 5 pulses per site, while surface EMG was acquired from biceps and triceps of the contralateral arm (i.e., the arm with weaker triceps in participants with SCI, and the arm alternated by dominancy in consecutively enrolled controls). Motor evoked potentials (MEPs) generated in each muscle were normalized to the maximal MEP (max-MEP) amplitude of biceps. The following motor map characteristics were determined from all active sites (i.e., with MEPs > 1/8 max-MEP) for each muscle: area (count of active sites), volume (total MEP amplitude), center of gravity (COG), mean MEP amplitude, and the mean MEP amplitude in the top 50% and the bottom 50% of the map (expressed relative to max-MEP). Graded Redefined Assessment of Strength Sensibility and Prehension (GRASSP) was collected in participants with SCI as a measure of functional impairment. A Group X Muscle mixed analysis of variance was performed with Bonferroni correction to compare map characteristics. Pearson’s correlation was used to determine the relationship between map characteristics and GRASSP scores in participants with SCI.
Results: While mapping characteristics for biceps and triceps did not differ in able-bodied controls, participants with SCI demonstrated significantly larger map area (40.4±12.1 vs 14.5±16.3 cm2, p=0.002), map volume (2395.5±1017.0 vs 385.4±532.5 %max-MEP *cm2, p=0.001), and MEP amplitudes (whole map: 54.5±19.8 vs 13.3±13.3 %max-MEP, p=0.001; top50%: 67.6±19.8 vs 3.9±15.1 %max-MEP, p=0.001; bottom50%: 28.0±9.5 vs 12.8±12.5 %max-MEP, p=0.003) for biceps than triceps. No differences in antero-posterior or medio-lateral coordinates of map COG were noted for biceps and triceps in either group of participants. In addition, mapping variables for neither muscle were found different between two groups of participants. Mean MEP amplitude difference in the bottom 50% map between biceps and triceps was positively correlated with Prehension (Qualitative Prehension: r=0.652, p=0.012; Quantitative Prehension: r=0.543, p=0.045) but not Strength and Sensation GRASSP scores.
Conclusion: In a typical cervical SCI, motor cortical map devoted to the less-affected biceps is over- represented than that devoted to the more-affected triceps, with relevance for upper limb motor function. Mapping characteristics may serve as potential biomarkers of response to treatments geared at augmenting control of more-affected musculature in SCI.
Categories
Spinal Cord Injury (SCI)
Wearable activity monitors as part of physical activity intervention for people with neurodegenerative diseases: opportunities and considerations
Hai-Jung Steffi Shih1, Philippa Morgan-Jones2, Katrina Long3, Abigail Schreier1, Lori Quinn1,4, Ciaran Friel5
1Teachers College, Columbia University, New York, USA. 2Cardiff University, Cardiff, United Kingdom. 3San Jose State University, San Jose, USA. 4Columbia University Irving Medical Center, New York, USA. 5Northwell Health, New York, USA
Abstract
Introduction: Wearable physical activity (PA) monitors such as Fitbits may facilitate PA uptake. Despite their popularity in cancer and obesity research, PA monitors have not been widely adopted in neurodegenerative disease research. While monitors show promise as a powerful tool for PA intervention, there are challenges unique to neurodegenerative diseases that warrant consideration. The purpose of this study was to evaluate the feasibility of using Fitbits in a PA intervention for people with Parkinson’s (PD) and Huntington’s disease (HD), and highlight considerations for future research.
Methods: Twenty-seven participants with PD (N=13, Hoehn & Yahr I or II) and HD (N=14, premanifest or early- stage) enrolled in Preactive, a 4-month PA coaching intervention grounded in self-determination theory. Participants received a Fitbit Charge 2, a disease-specific workbook, and 4-6 individualized coaching sessions. Fitbit-use allowed for: 1) participant self-monitoring, and 2) therapist review to inform coaching and goal-setting. Weartime was determined using minute-level heart rate data, with a valid wear day defined as 10 hours or more of wear. Steps and metabolic equivalent (MET) data were extracted for a week at baseline and follow-up. MET*minute per week was subsequently calculated with a minimal threshold of 3 METs.
Results: Retention rate of the intervention was 85%. Reasons for dropouts (n=4) were time constraints (n=2), mood disorders (depression and apathy), and cognitive impairments. Participants had a mean(SD) of 92.3(9.2)% valid wear days over the intervention period. Average daily weartime was 18.4(4.5) hours, with an initial 2-week period of high weartime variability. Regardless of diagnosis, participants emerged as either day wearers (average daily weartime 13.0(1.2) hours) or day and night wearers (average daily weartime 21.9 (1.0) hours). Steps and MET*minutes data showed a different distribution between diagnoses – those with PD had mean(SD) 7564(2804) steps/day and 9039(3987) MET*minutes/week at baseline and 8313(2783) steps/day and 10319(4934) MET*minutes/week at follow-up; those with HD had 11193(3342) steps/day and 12358(5048) MET*minutes/week at baseline and 10165(3484) steps/day and 13192(6224) MET*minutes/week at follow-up.
Discussion: Implementation of Fitbits in a PA intervention for people with early-stage PD and HD was feasible. The spontaneous divide between day wearers and day and night wearers may be an important consideration for researchers and clinicians who wish to gain insight on sleep. An early trend of lower steps and MET*minutes for individuals with PD compared to HD emerged, possibly reflecting hypokinesia in PD and hyperkinesia in HD. Future research should consider cognitive impairments, disease-specific symptoms, and medication as they may affect Fitbit use and data interpretation.
Categories
Other
Worse Performance of Instrumental Activities of Daily Living Associates with Markers of Neurodegeneration
Audrey Keleman, Rebecca Bollinger, Julie Wisch, Beau Ances, Susan Stark
Washington University School of Medicine, St. Louis, MO, USA
Abstract
Objective: In preclinical Alzheimer disease (AD), biomarkers such as amyloid plaque accumulation and evidence of neurodegeneration (e.g., reduced hippocampal volume and weaker intra-network connections) are apparent in the brain, but memory symptoms may not have occurred. Performance- based assessments of Instrumental Activities of Daily Living (IADL) have been able to detect mild cognitive impairments but have not yet been examined in preclinical AD. A performance-based IADL assessment may be sensitive and detect early functional impairment in the preclinical stage of AD.
Methods: In this cross-sectional analysis of an ongoing longitudinal cohort study, cognitively normal (CN) older adults completed 3 standardized IADL tasks (shopping, checkbook balancing, and medication management) from the Performance Assessment of Self-Care Skills (PASS) in their home. Biomarkers of AD were assessed through magnetic resonance imaging (MRI) within 2-3 years of an individual’s completion of the PASS and include: amyloid accumulation measured by positron emission tomography (PET) using one of two tracers [Pittsburgh Compound B (PIB) and 18F-AV-45 (Florbetapir)], and markers of neurodegeneration, including hippocampal volume and resting-state functional connectivity (rs-fc) signature. Rs-fc signature is a summary value representing intra-network connection strength.
Associations between performance on PASS tasks and the biomarkers of AD were quantified by Pearson point-biserial partial correlations (controlling for age and sex), as biomarker data was normally distributed and PASS scores fell between two categories (acceptable or optimal quality of performance). Correlations with a 95% confidence interval (CI) that did not include 0 were considered significant.
Results: 161 CN participants (mean age 74.3 years, 55% female) were included. After controlling for age and sex, worse performance of IADL tasks was associated with smaller hippocampal volumes (Pearson’s r: 0.302, p=.02, 95% CI: .179 to .416) and weaker within-network rs-fc network connections (Pearson’s r: 0.276, p=.03, 95% CI: .078 to .453), but not amyloid accumulation.
Discussion: This study suggests that worse performance of IADL tasks may be associated with markers of neurodegeneration among CN older adults. These findings could lead to a better understanding of functional changes that may associate with neurodegeneration prior to the onset of noticeable memory symptoms in AD or related dementias.
Categories
Cognitive/Language Rehabilitation
Healthcare Resource Utilization and Costs in Adult Patients With Spasticity – A Matched Cohort Analysis
Michael Hull1, Vamshi Ruthwik Anupindi1, Jing He1, Natalya Danchenko2, Mitchell DeKoven1, Jonathan Bouchard3
1IQVIA, Falls Church, USA. 2Ipsen, Boulogne-Billancourt, France. 3Ipsen, Cambridge, USA
Abstract
Background: Spasticity is a velocity-dependent increase in muscle tone that is associated with upper motor neuron lesions and is seen in individuals with neurological etiologies such as cerebral palsy (CP), multiple sclerosis (MS), stroke, or injuries to the central nervous system (CNS). Spasticity affects more than a half million individuals in the United States (U.S.) alone, and the loss of function, increased pain, and/or abnormal limb posture that results from spasticity can significantly impact patient quality of life. This analysis evaluated incremental healthcare resource utilization and healthcare costs in adult patients with spasticity and one of the following etiologies: CP, MS, stroke, or CNS injury (cases) versus a random sample of patients without spasticity (controls) over a 1-year period from a U.S. payer perspective.
Methods: Retrospective, observational analyses were conducted using administrative healthcare claims from the IQVIA PharMetrics® Plus database of commercially or self-insured members from October 1, 2015, to December 31, 2019. Cases were identified based on International Classification of Diseases, 10th Revision (ICD-10) diagnosis codes for evidence of spasticity (index date) and a background etiology (6 months before or up to 7 days post-index) with 6 months of continuous enrollment before the index date (pre-index) and 12 months following the index date (post-index). Index date was defined as the date of the first evidence of spasticity and a background neurological etiology before or up to 7 days post-index. The control group was identified based upon no evidence of spasticity, along with continuous health plan enrollment for 6 months before the index date (pre-index) and 12 months following the index date (post-index). Cases were direct matched 1:1 to the control group. Pharmacy, outpatient, and inpatient costs were included in all-cause cost analyses (USD).
Results: Overall, 215,739 cases were matched to controls. Cases had statistically higher mean±SD ($29,912±70,494 vs $7,464±24,003) and median ($9,113 vs $1,979) total post-index healthcare costs (both p<0.0001) compared to controls. Cases also had significantly more mean prescription fills (31.6±36.1 vs 15.6±21.4), physician office (19.4±21.4 vs 6.0±9.5), and physical therapy visits (12.9±24.5 vs 1.5±7.4) compared to controls. Higher mean pharmacy ($7,056±24,926 vs $2,220±13,341), physician office ($2,979±13,235 vs $834±5,914), and inpatient costs ($11,245±50,998 vs $1,486±11,183) were also observed for cases (all p<0.0001) compared to controls.
Conclusions: In this retrospective real-world study, adult patients with spasticity had higher mean and median healthcare costs and healthcare resource utilization than matched controls did, resulting in a higher economic burden. More effective treatment of spasticity therefore represents a considerable potential for cost-savings within managed healthcare systems.
Categories
Other
Absence of perilesional neuroplastic recruitment in chronic post-stroke aphasia
Andrew DeMarco, Candace van der Stelt, Sachi Paul, Elizabeth Dvorak, Elizabeth Lacey, Sarah Snider, Peter Turkeltaub
Georgetown University, Washington, DC, USA
Abstract
A prominent theory proposes that neuroplastic recruitment of perilesional tissue supports aphasia recovery. This idea emphasizes the role of tissue immediately surrounding the lesion, where animal studies have both observed dysfunction and suggested that collateral axonal sprouting and synaptogenesis may support functional recovery. Motor stroke recovery, in particular, appears to rely on functional take-over by perilesional sensorimotor or primary motor cortices. These findings have informed models of aphasia recovery, which stipulate that when language tissue is damaged, alternative perilesional processors may become recruited to support outcomes, especially around small lesions.
Here, we test the perilesional plasticity hypothesis using two fMRI tasks in two groups of patients with prior aphasia diagnosis.
Two cohorts totaling 81 chronic (>6 months) left-hemisphere stroke patients with prior aphasia diagnosis, and 80 control participants underwent BOLD fMRI using either a naming task or a reliable and valid semantic decision task. Individualized perilesional tissue was defined by various dilation operations applied to anatomical lesion tracings. The brain was separated into language tissue, language-capable tissue, and non-language tissue based on independent task-specific meta-analytic maps. For each tissue type, linear mixed-effects modeling examined differences in activity (patient vs control) in concentric perilesional shells. The analysis was performed brainwide in 4mm perilesional shells (0-4mm, 4-8mm, 8- 12, 12-16, and >16mm) and also in atlas regions with adequate lesion coverage either near (4-16mm) or far from (>16mm) the lesion boundary. For each task, the relationship between perilesional language activation and a) lesion size and b) aphasia severity were also examined.
In both tasks, patients exhibited reduced activity, relative to controls, in perilesional language and language-capable tissue up to 8mm from the lesion boundary, and up to 12mm from the lesion boundary in the naming task. Although a few cortical regions exhibited greater activity irrespective of distance from the lesion, or only when distant from the lesion, no regions exhibited increased activity only when near the lesion. Larger lesions (>100cc) but not smaller lesions (<50cc) were associated with reduced language activity, relative to controls, irrespective of distance from the lesion. When using the reliable fMRI task, a reduction in language activity was significantly related to aphasia severity, independent of lesion size.
We find no evidence for neuroplastic recruitment of perilesional tissue in aphasia beyond its typical role in language. Rather, our results are consistent with alternative hypotheses that changes in left- hemisphere activation during recovery are related to normalization of language network dysfunction and possibly recruitment of alternate cortical processors. These findings clarify left-hemisphere neuroplastic mechanisms supporting language recovery after stroke.
Categories
Neural Repair Mechanisms
Exoskeletons increase paretic limb use in stroke survivors during a bimanual virtual reality reaching task
Alexander Brunfeldt1, Barbara Bregman1, Peter Lum2
1Georgetown University, Washington, USA. 2Catholic University of America, Washington, USA
Abstract
We used exoskeleton and virtual reality devices to increase the use of the paretic limb in hemiplegic stroke survivors. This feasibility study in 3 chronic stroke survivors shows that patients alter upper extremity coordination, when gravity assistance is applied to the impaired limb, in a bimanual reaching task. We discuss these results in the context of graded-constraint, allowing for bimanual control during forced-use therapies.
Nearly 8 million Americans live with disability cause by stroke. Device-based rehabilitative therapies focus on aiding the impaired limb when patients engage in goal-directed, unimanual tasks. Recently, bilateral therapies have been developed; however, they fail to show sustained benefits, likely due to a lack of resemblance to activities of daily living. The goal of this study was to develop a rehabilitation platform, comprising exoskeleton and virtual reality technologies, to increase the use of the paretic limb in a bimanual reaching task.
Three hemiplegic stroke participants reached for targets in a virtual reality environment using both hands. They completed 162 reaches divided into 3 blocks. Following baseline, we used an exoskeleton to provide 50% arm weight compensation to the paretic limb and used wrist weights to provide 50% arm weight resistance to the non-paretic limb. We removed the exoskeleton and wrist weights during the retention block. We used electromyography to approximate muscle activity in the biceps brachii. Relative contribution (RC) was calculated as the displacement of the paretic arm divided by the sum of displacements for both arms. Muscle contribution (MC) was calculated as the root mean square of paretic arm muscle activity divided by the sum of activity for both arms.
During baseline, RC of the impaired limb ranged from 43 to 49%, and MC of the impaired bicep ranged from 35 to 43% in 3 mild to moderately impaired patients (Fugl-Meyer Upper Extremity scores of 43, 37, and 44). During loading, RC increased by 5.6%, 1.1%, and 2.4% respectively; MC decreased by 8.3%, 11.8% and 21% respectively. These data suggest our device increases limb displacement while simultaneously reducing the muscular effort in the impaired limb. Importantly, these results closely match data from our previous work in 12 healthy controls, where we found a 1.3% increase in RC coincides with a 12% decrease in MC. By collecting more data on stroke patients, we will quantify this tradeoff between coordination and muscle activity modulation, allowing us to optimize the exoskeleton mechanics to maximize paretic limb use.
We demonstrate our platform is well tolerated by mild to moderately impaired stroke patients. With further development, researchers and clinicians can use our platform to fine-tune the level of limb constraint based on the individual needs of the patient and allows for bilateral constraint-based therapies.
Categories
Motor Rehabilitation
Effects of repetitive transcranial magnetic stimulation of contralesional dorsal premotor cortex on interhemispheric functional connectivity in severe chronic stroke
Xin Li1, David Cunningham1,2,3,4, Ken Sakaie5, Mark Lowe5, Yin-Liang Lin1,6, Steven Wolf7, Adriana Conforto8, Andre Machado9, Akhil Mohan1, Kyle O’Laughlin1, Xiaofeng Wang10, Morgan Widina1, Ela Plow1,9
1Cleveland Clinic Lerner Research Institute, Cleveland, USA. 2Case Western Reserve University, Cleveland, USA. 3MetroHealth Medical Center, Cleveland, USA. 4Cleveland Functional Electrical Stimulation Center, Cleveland, USA. 5Cleveland Clinic Imaging Institute, Cleveland, USA. 6National Yang Ming Chiao Tung University, Taipei, Taiwan. 7Emory University School of Medicine, Atlanta, USA. 8Hospital Das Clínicas/São Paulo University, São Paulo, Brazil. 9Cleveland Clinic Neurological Institute, Cleveland, USA. 10Cleveland Clinic Quantitative Health Sciences, Cleveland, USA
Abstract
Introduction: Up to 50% of stroke survivors experience persistent, severe upper limb paresis despite intensive rehabilitation.1 Treatment options for severely impaired survivors are often limited and ineffective. Repetitive transcranial magnetic stimulation (rTMS) is a well-known technique typically given to facilitate excitability of the ipsilesional primary motor cortex (iM1) for chronic post-stroke motor improvement.2,3 Despite promising early evidence, this approach has failed to produce motor gains in severely impaired survivors. Recent evidence indicates that the dorsal premotor cortex in the (undamaged) contralesional hemisphere (cPMd) may be a more suitable target based on alternate and available, bi-hemispheric and ipsilateral connections.4–6 Our recent study has shown that a single session of rTMS delivered to facilitate cPMd enhanced paretic limb reaching speed in association with strengthened interhemispheric connectivity in severely impaired survivors who otherwise showed little motor gains with conventional rTMS delivered to facilitate iM1.7
Purpose: In an ongoing study, we are evaluating whether multiple sessions of cPMd rTMS given in conjunction with rehabilitation can produce clinically meaningful and longer-lasting motor functional gains in severely impaired stroke survivors. We hypothesize that cPMd rTMS combined with rehabilitation will produce greater motor functional gains than conventional iM1 rTMS combined with rehabilitation in severely impaired stroke survivors based on increase in interhemispheric functional connectivity.
Methods: In a randomized, assessor-blinded, pilot clinical trial, 14 participants with chronic severe stroke-related motor impairment, defined as lack of 10° extension in paretic wrist/thumb/fingers or absence of TMS-based motor evoked potentials in paretic finger extensor muscles, were randomized to receive cPMd rTMS or iM1 rTMS both in conjunction with upper limb therapy, delivered twice a week for 6 weeks. Assessments of motor impairment and function were made at baseline, 6wk end-of-treatment and 3-month follow-up. Assessment of interhemispheric connectivity was made with neurophysiology tested using TMS and resting state functional magnetic resonance imaging (rs-fMRI),8 both collected at baseline and 6wk end-of-treatment.
Results: Preliminary findings indicate participants who receive cPMd rTMS in conjunction with rehabilitation experience greater reduction in motor impairment (Upper Extremity Fugl-Meyer [UEFM] 6wk-gain, mean±SEM, 4.57±1.39) and greater gains in motor function (Wolf Motor Function Test [WMFT] normalized rate 6wk- gain, 4.97±2.05) compared to those that receive iM1 rTMS in conjunction with rehabilitation (UEFM gain 3.5±1.26, WMFT gain 1.31±3.36). In a sub-sample of participants that completed rs-fMRI assessments (n=5, due to COVID pause or claustrophobia), we find those that receive cPMd rTMS show greater increases in interhemispheric functional connectivity between cPMd and ipsilesional motor areas in association with higher gains in motor function.
Discussion: Addition of anticipated 10 more participants is expected to give insight into effects and mechanisms of a novel brain stimulation approach designed to target available substrates in severely impaired stroke survivors.
Categories
Stroke
Early diagnosis of spasticity in acute post-stroke patients
Mindy F. Levin1,2, Alice Misana1,2, Marie-Hélène Boudrias1,2, Alexander Thiel3,4, Theodore Wein3,5,6
1School of Physical and Occupational Therapy, McGill University, Montreal, Canada. 2Centre for Interdisciplinary Research in Rehabilitation, Montreal, Canada. 3Department of Neurology and Neurosurgery, McGill University, Montreal, Canada. 4Jewish General Hospital, Montreal, Canada. 5Montreal Neurological Hospital, Montreal, Canada. 6McGill University Health Center, Montreal, Canada
Abstract
Spasticity is a major disabling condition following stroke. It is clinically defined as a pathological velocity- dependent increase in muscle reflex activity that affects both resting muscle tone and voluntary movement production.1 Recovery of lost motor function in patients with stroke may be affected by spasticity, which most commonly develops in elbow flexors and ankle plantarflexors.2,3 However, despite its clinical relevance, the evolution of spasticity over the first 3 months after stroke is not clearly understood. Indeed, common clinical measures of spasticity, such as the Modified Ashworth Scale (MAS),4 lack reliability and objectivity5 and do not take into account the neurophysiological origin of spasticity. The objectives of this study are: 1) to determine the feasibility of examining the evolution of spasticity development over the first 3 months post-stroke; 2) to examine the extent to which changes in a new physiological measure of spasticity (tonic stretch reflex threshold, TSRT, and its velocity sensitivity, μ) can detect the presence/absence of spasticity; and 3) to determine the sensitivity of TSRT/ μ to detect spasticity compared to MAS and other common clinical measures. We hypothesize that 1) TSRT/μ will detect the presence of spasticity earlier than MAS/clinical measures, and 2) TSRT/μ will have a greater correlation with motor impairments and activity limitations than MAS.
Beginning from admission, outcomes are measured in patients hospitalized for first-ever stroke (≥18 y/ o) with resulting hemiparesis. Measurements are done once weekly for 12 weeks with a follow-up assessment at week 16. Primary outcome measures are TSRT/μ and MAS/clinical measures. TSRT/μ is measured with a portable clinical device called Montreal Spasticity Measure (MSM), which has been shown to have good inter-rater and test-retest reliability.6,7
The research protocol is feasible for the patient and clinical research teams in acute-care, rehabilitation and follow-up care centres. Of the 7 patients recruited to date, 3 completed all 16 weeks of data collection, 3 recovered before the trial end and 1 withdrew . In 5 patients examined for the occurrence of elbow flexor spasticity, TSRT/μ in elbow flexors was detected earlier than the MAS. MAS and increased reflex activity was only evident in 1 patient. Increases in TSRT (towards normal values) were related to higher values of μ (towards normal values). Neither TSRT/μ, MAS or any of the clinical measures (strength, ROM, sensation) was related to UL impairment scores (Fugl-Meyer Assessment-UL) in this small cohort. FMA-UL scores were highly correlated with isometric strength of elbow flexors and extensors.
The results of this study will improve our ability to detect and diagnose spasticity earlier and/or predict which patients will develop spasticity in order to deliver more effective treatments, improve prognosis and improve patient quality of life.8.9
Categories
Stroke
Assessing sensorimotor function after stroke. A survey of neurorehabilitation clinicians
Joanna Eskander1,2, Michael Borich3, Trisha Kesar3, Darcy Reisman1,4, Jennifer Semrau1,2
1Department of Biomechanics and Movement Science Program, University of Delaware, Newark, DE, USA. 2Department of Kinesiology and Applied Physiology, University of Delaware, Newark, DE, USA. 3Division of Physical Therapy, Department of Rehabilitation Medicine, Emory University, Atlanta, GA, USA. 4Department of Physical Therapy, University of Delaware, Newark, DE, USA
Abstract
Nearly 75% of people living with stroke have somatosensory impairments, which include tactile sensation and proprioception. Deficits in somatosensation can impair sensorimotor integration, causing difficulty with activities of daily living, balance, and mobility. Previous survey studies revealed that clinicians report assessing sensation often in individuals post-stroke, frequently relying on non- standardized assessments. We developed a survey of neurorehabilitation clinicians to identify which sensory assessments were considered clinically important, how often sensory testing was completed, and the priority level for capturing sensory function during a typical evaluation post-stroke. We hypothesized that somatosensation was evaluated less frequently than motor function, and that clinicians demonstrate low utilization of standardized somatosensory assessments and rely more on non- standardized somatosensory assessments.
We distributed an electronic survey to clinicians (e.g. MD, RN, PT, OT) to examine the clinical utility and importance of somatosensory assessment. The survey was structured into several categories: demographics, typical evaluation, knowledge of sensory assessments, and barriers in practice. Inclusion criteria included clinicians who had treated at least one individual post-stroke in the past 6 months. A total of 185 participants responded to the survey with an average post-stroke caseload of 39% ± 25.
Most respondents were US-based occupational [34%] and physical therapists [41%], primarily in acute care [45%].
Preliminary results revealed that prioritization of assessing sensory status during an evaluation was high [90%]. However, the most used standardized (Semmes-Weinstein Monofilaments [40%]) and non- standardized (light touch screening [81%]) assessments are both in the tactile domain, with neither addressing any aspect of proprioceptive status. While evaluation durations varied among respondents (median=66 ± 95 minutes) prioritization of sensory assessment viewed as a percentage of total evaluation time was relatively low (median=6.25% ± 3.8). Additionally, 60% of respondents devoted most assessment time to motor function (e.g., ROM, strength) during clinical evaluation, but length of sensory assessment was ranked 6th of 11 domains included in the survey. A paired Wilcoxon test showed significantly greater familiarity (p<0.01) and clinical use (p<0.01) for non-standardized versus standardized sensory assessments. Sixty percent of clinicians reported barriers to using sensory assessments, with the top barriers including lack of time, access to standardized assessments, and priority within workplace.
Our preliminary results suggested that while there was limited utilization and prioritization of sensory assessments after stroke, those that were completed were typically related to the sensation of touch, despite other sensory aspects (e.g., proprioception) being integral to sensorimotor control and stroke recovery. Further, our results suggested that important factors such as time, access, and the lack of clinical priority placed on somatosensation may drive these differences. Overall, this survey suggested that there is a need for greater clinical priority to be placed on sensory assessment post-stroke to maximize the functional status and recovery of stroke survivors.
Categories
Sensory Rehabilitation
Motor Imagery has Variable Effects on Peripheral Nerve Recovery
Taewon Kim1, Susan Mackinnon1, Jana Dengler2,3, Benjamin Philip1
1Washington University School of Medicine, Saint Louis, USA. 2Sunnybrook Health Sciences Centre, Toronto, Canada. 3University of Toronto, Toronto, Canada
Abstract
Recovery of function after upper extremity peripheral nerve injury may depend, in part, on sensorimotor cortex remapping to reflect changes in the periphery. Motor imagery (imagined movement) affects cortical sensorimotor representations, but it remains unknown whether these two concepts involve the same mechanism; i.e., whether imagery-driven cortical plasticity is a bottleneck for functional recovery after peripheral nerve injury. Here, we performed a small proof-of-concept pilot study to determine whether motor imagery can improve hand movement performance in patients with unilateral upper extremity peripheral nerve injury. All participants (currently n=4, plus one omitted due to orthopedic injury in their other hand) practiced by imagining the movement and sensation of rapid thumb-to-finger tapping with their injured hand, alternating between 2 unique sequences (thumb to 1-4-2-3 and 4-1-3-2). Real movements were evaluated in each hand, before and after training, to determine percentage change in finger-tapping rate from pre-training to post-training. A short-term group (n=2) practiced for a 20 minute session, which comprised: thirty blocks of imagined movement practice, each block 20 sec, alternating between sequences, with 15 seconds between blocks. A long-term group (n=2) of chronic patients (> 1 year since injury, plateaued performance with injured hand) performed an identical 20 minute practice session, but performed it every day for 14 days.
Both groups improved their finger-tapping rate with both hands after motor imagery practice, but the pattern depended on the training duration. The short-term participants improved more for the injured hand than the uninjured hand (one participant 36.6%/11.9%, other participant 6.1%/1.5%). However, the long-term participants improved more for the uninjured hand than the injured – i.e. more for the untrained hand than the trained hand (one participant 1.4%/17.9%, other participant 10.0%/21.2%). It remains unclear whether these differences arise from the training duration or the participants, but the group criteria offers one possible interpretation: the requirements for the long-term group to be >1 year post-surgery and have reached a stable plateau of hand function. Therefore, our preliminary results suggest that motor imagery can improve hand function after upper extremity peripheral nerve injury, but may not provide further benefit to chronic patients whose recovery has stalled.
Categories
Motor Rehabilitation
Willed movements versus passive observation during Mirror Therapy and Video Therapy in hemiplegic patients: a behavioral and EEG maps comparison
Pascal Giraux1,2, Ahmed Adham2, Clara Pfenninger3, Diana Rimaud1, Julia Touly1
1Adult PRM department, University Hospital of Saint-Etienne, Saint-Etienne, France. 2Lyon Neuroscience Research Center (CRNL), Trajectoires team, INSERM UMR-S U1028, CNRS UMS 5292, Lyon, France. 3Laboratoire Inter-Universitaire de Biologie de la Motricité, EA 7424, Univ Lyon, UJM-Saint-Etienne, Saint- Etienne, France
Abstract
Background and aims: Mirror Therapy (MT) and Video Therapy (VT) are two rehabilitation techniques based on visuomotor feedback that are routinely used motor training of the upper-limb in post-stroke hemiplegia. The motor task performed by the impaired limb can be passive observation, motor imagery or willed movements. This motor intention difference could influence the efficacy of these techniques. In these experiments, we studied how this motor intention gradation modulate the perceived intensity of the kinesthetic illusion and its related EEG maps during MT and VT.
Methods: 10 chronic hemiplegic patients performed a VT session involving wrist movements of their impaired limb either in a passive observation (PO) condition or in a willed movement (WM) condition (VT device: IVS™, Dessintey, France). The intensity of the illusion was evaluated by a visual analog scale (VAS from 0 to 10) as well as the angle of the perceived movement. In a subsequent study, EEG data (32 channels) from 10 chronic stroke patients and 15 healthy control subjects were collected during MT and VT with three distinct motor tasks conditions: passive observation, motor imagery and willed movements. Time-frequency maps were analyzed.
Results: The illusion intensity was significantly greater during WM (4.1 ± 2.9) than PO (2.0 ± 3.0) (p < 0.005). The perceived angle was also greater during WM (13.5 ± 21.6 °) than PO (28.7 ± 27.1 °) (p = 0.005). EEG frequency maps demonstrated a modulation of ERD and β-rebound according to the motor task, both in controls and stroke patients, with greater variability in patients according to the brain lesion and the level of motor recovery.
Discussion: Willed movement coupled with positive visual feedback enhances the perceived illusion of movement as compare to a passive observation task in post-stroke hemiplegic patients. This intentionality effect goes with a neurophysiological modulation of ERD and β-rebound and could influence the efficacy of visuomotor feedback therapies (mirror or video therapy).
Categories
Motor Rehabilitation
Mapping the Human Cervical Spinal Cord with Electrical Stimulation for Neurorehabilitation
James R. McIntosh1,2, Evan F. Joiner1, Jacob L. Goldberg2, Lynda M. Murray3,4, Bushra Yasin1,2, Anil Mendiratta1, Steven C. Karceski2, Earl Thuet5, Oleg Modik2, Evgeny Shelkov2, Christopher Mandigo1,5, K. Daniel Riew1,2,5, Noam Y. Harel3,4, Michael S. Virk2,5, Jason B. Carmel1,2
1Columbia University, New York, USA. 2Weill Cornell Medicine, New York, USA. 3Icahn Sch. of Med. at Mount Sinai, New York, USA. 4James J. Peters VA Med. Ctr., Bronx, USA. 5New York Presbyterian, New York, USA
Abstract
While epidural spinal cord stimulation has emerged as a powerful modality for recovery of leg movement, how it should be targeted to the cervical spinal cord to activate arm and hand muscles is not well-understood, particularly in humans.
We sought to map muscle responses to posterior epidural cervical spinal cord stimulation in humans. We hypothesized that lateral stimulation over the dorsal root entry zone (DREZ) would be most effective, and responses would be strongest in the muscles innervated by the stimulated segment.
Eighteen people undergoing clinically indicated cervical spine surgery were enrolled. During surgery, bipolar stimulation was performed in midline and lateral positions at multiple exposed segments; eight arm and three leg muscles were recorded on each side of the body.
Across all segments and muscles tested, lateral stimulation over the DREZ produced stronger muscle responses than midline (in 91.5% of conditions). These stronger responses were due to a combination of lower stimulus intensity needed to produce a response (lower threshold) and increased recruitment at greater stimulus intensities. As hypothesized, muscles innervated at a cervical segment had the largest responses from stimulation at that segment, with similar responses in adjacent segments. However, responses were also observed in muscles innervated at more distant levels of the cervical enlargement as well as leg muscles. We also observed that two distinct clusters emerged at C4-C6 vs C8-T1 that produced strongly similar muscle activation internally, while producing different muscle activation from each other.
In the human cervical spinal cord, strong responses to lateral stimulation are likely due to the proximity of stimulation to afferent axons which are likely recruited at relatively low thresholds, similar to the putative mechanism of lumbar epidural stimulation. More similar than expected muscle responses with stimulation of adjacent cervical segments argues for local circuit integration, and the distant muscle responses suggest activation of long propriospinal connections. This map may guide cervical stimulation to improve arm and hand function.
Categories
Spinal Cord Injury (SCI)
Feasibility and Impact of transcranial photobiomodulation on fine hand motor skill learning in non-disabled young adults
Alexandra Messur, Jocelyn Penteck, Bokkyu Kim
SUNY Upstate Medical University, Syracuse, USA
Abstract
Introduction: Transcranial Photobiomodulation (tPBM) using near-infrared LED lights is a promising non- invasive brain stimulation to enhance experience-dependent neuroplasticity for motor skill learning. This pilot clinical study examined the feasibility and impact of the tPBM to the primary motor cortex (M1) on fine hand motor skill learning in non-disabled young adults.
Methods: Thirteen non-disabled young adults were recruited. All participants performed the baseline fine hand motor behavioral assessments, including Box and Block Test (BBT) and Modified Box and Block using chopsticks (mBBT) on each hand, and chopstick object pick-up test (COPT) on the non-dominant hand. Then, they performed transcranial magnetic stimulation (TMS) testing for corticospinal excitability of the first dorsal interosseus (FDI) on a non-dominant hand. Participants were then randomly allocated into three groups: 1) sham, 2) pre-practice tPBM experimental group, and 3) post-practice tPBM experimental group. Participants in the experimental groups received either pre- or post-practice tPBM stimulation for 15 minutes on the contralateral primary motor cortex to the non-dominant hand. The motor skill practice consisted of 50 trials of object pick up using a pair of chopsticks, as they performed in the COPT. Sham group received sham light therapy either before or after the motor practice. All groups were subjected to a 10-minute post-practice COPT and TMS testing. Participants returned 24 hours following their first visit to perform a COPT as well as a TMS testing. During the COPT and motor skill practice sessions, we recorded the kinematics of fine hand motor performance using stereo cameras and computer vision object tracking. Kruskal Wallis tests were used to compare kinematic variables among different groups at each time point.
Results: All participants completed the study protocol without any adverse effects. Object movement duration and peak velocity were statistically different between groups at the immediate post-practice timepoint. The pre-practice tPBM group showed the greatest improvements in the movement duration and peak velocity at the immediate post-practice timepoint. There was no statistically significant group difference in kinematic variables at the 24-h post-practice time point. There were no statistically significant group differences in BBT, mBBT, and TMS variables at 24-hour post-practice time point.
Discussion: We confirmed that the tPBM in conjunction with motor skill practice is safe and feasible. Further, COPT performance data support that the Pre-practice tPBM to M1 can enhance fine hand motor performance in non-disabled adults. This result may indicate that the tPBM to M1 can prime the M1 to enhance motor skill performance. However, there was no impact of tPBM on fine hand motor skill learning. Further studies with a large sample size are needed to determine the effectiveness of tPBM on motor performance and motor skill learning.
Categories
Motor Rehabilitation
Contributions of the more affected arm and hand for bimanual tasks: insights about action selection and performance in chronic stroke survivors
Marika Demers, Lauri Bishop, Amelia Cain, Nicholas Schweighofer, Carolee Winstein
University of Southern California, Los Angeles, USA
Abstract
Background: Many stroke survivors experience decreased use of the more affected arm and hand in everyday activities (1–3), with a frequent disparity between motor capacity (what one can do) and performance (what one chooses to do). While the disparity between capacity and performance is complex and multifactorial, it may lead to maladaptive spasticity, functional decline, higher disability levels, and limited community reintegration (4,5). Expert consensus recommends the use of objective and longitudinal standardized assessments to diagnose the disparity between capacity and performance (6). Yet, standardized assessments may not reflect performance of everyday activities or the real time action selection to use the more affected arm and hand particularly during bimanual tasks. This study aims to quantify the action selection strategies and movement count ratio in chronic stroke survivors when performing a commonly used clinical assessment of bimanual arm and hand activity limitations, the Chedoke Arm and Hand Activity Inventory-7 (CAHAI).
Methods: The sample comprises 33 chronic stroke survivors with a wide range of arm and hand motor impairments (Fugl-Meyer scores range 4-66) and five young neurotypical participants. Each participant was video recorded while they completed the seven bimanual, functional tasks included in the CAHAI. From video recordings, movement counts were annotated by an experienced occupational therapist. The arm used to accomplish each task component was also identified. A ratio between the movement counts of the paretic arm over the movement counts of both arms was computed from the video annotation metrics for each task.
Results: Neurotypical adults performed all tasks bimanually and used each arm relatively equally (average dominant arm use ratio: 51.0 ± 1.5%). Stroke survivors had an average paretic arm use ratio of 26.2 ± 15.6% (range: 5.4-51.7%). The task demands influenced whether stroke survivors chose to complete the task with or without the more-affected arm. For example, 87.7% of stroke survivors used both hands to ‘trace a line with a ruler’ (average ratio = 27.4%). In contrast, ‘Pouring a glass of water’ and ‘Buttoning 5 buttons’ were performed bimanually by 33.3% (average ratio: 16.7%) and 36.7% (average ratio: 17.9%) of stroke survivors, respectively. For those who did not use the more-affected arm for bimanual tasks, they either accomplished them unimanually with the less-affected arm or used alternative compensatory strategies. More generally, the relationship between the number of tasks accomplished bimanually and FMA scores was strong (r2= 0.883).
Conclusion: Understanding task-level strategies (e.g., use ratio, bimanual vs unimanual) to accomplish bimanual tasks across a range of motor impairment levels provides valuable insight about action selection that goes beyond global motor capacity. This insight can begin to inform the development of personalized and effective restorative treatment interventions to increase paretic arm and hand use for everyday activities.
Categories
Stroke
The Use of Transcranial Magnetic Stimulation for Upper Extremity Motor Assessment at the Bedside During Acute Stroke Hospitalization: A Feasibility Study
Isha Vora1, David Lin2,3,4, Yi-Ling Kuo5, Russell Banks6, Julie DiCarlo2,3,4, Leigh Hochberg2,3,4, Teresa Kimberley1
1MGH Institute of Health Professions, Boston, USA. 2MGH Center for Neurotechnology and Neurorecovery, Massachusetts General Hospital, Boston, USA. 3Department of Neurology, Massachusetts General Hospital, Boston, USA. 4RR&D Center for Neurorestoration and Neurotechnology, Providence VA Medical Center, Providence, USA. 5SUNY Upstate Medical University, Syracuse, USA. 6Linus Health, Boston, USA
Abstract
Transcranial Magnetic Stimulation (TMS) is a non-invasive brain stimulation technique that can be used to investigate corticospinal tract integrity and cortical excitability. TMS is a common investigative tool in the stroke population, though is most often used during the chronic phase of stroke. Assessing the potential of gathering TMS measures and examining their relationship to behavioral outcomes during acute stroke hospitalization are important first steps in developing optimally effective intervention strategies for people after stroke. This study aims to assess the feasibility of collecting TMS measures at the bedside during acute stroke hospitalization in the United States and to associate TMS neurophysiology with upper extremity (UE) behavioral measures. The Stroke Motor reHabilitation and Recovery sTudy (SMaHRT) is a prospective, observational cohort study taking place on the inpatient neurology service at Massachusetts General Hospital. Forty-two adults (65.44 ± 12.52 years) with stroke resulting in UE weakness who could follow simple commands in English and had no contraindications to receiving single-pulse TMS were eligible for and included in this study. Fifteen eligible participants did not receive TMS testing due to medical acuity, short hospital length of stay, or COVID precautions. TMS was performed within 5.2 ± 2.2 days of stroke. TMS measures collected at the bedside included motor evoked potentials (MEPs), resting motor threshold (RMT), cortical silent period (cSP) and ipsilateral silent period (iSP). Behavioral measures collected included UE Fugl-Meyer Assessment (FMA-UE), grip strength, shoulder abduction finger extension (SAFE) score and 9-hole peg test. MEPs, RMT, cSP or iSP were collected in at least one hemisphere for 27/42 participants. A mobile TMS cart, coordination with nursing staff and use of position adjusting hospital chairs/beds maximized successful testing. On average, TMS testing took 30-45 minutes to perform. Patients’ whose lesioned hemisphere was MEP+ compared to MEP- had better FMA-UE, SAFE, 9HPT and grip strength scores (p<0.01). Lower RMT values were also associated with better FMA-UE (rho=-0.6, p<0.05), SAFE (rho=-0.5, p=0.07), and Grip Strength (rho=-0.4, p=0.23) scores. There were no clear linear trends between behavioral outcomes and cSP or iSP. We found that it is feasible to gather several TMS measures at the bedside during acute stroke hospitalization. Interdisciplinary coordination, a mobile TMS cart, and participant positioning are essential to successful TMS testing. In addition to MEP, RMT has added value for relating to UE motor behavior during acute stroke hospitalization. Future research should focus on long-term follow up to address the potential predictive value of acute TMS measures for recovery.
Categories
Stroke
Validating the measurement of upper extremity sensorimotor behavior utilizing a tablet device in neurologically intact and stroke populations
Devin S Austin, Makenna Dixon, Joshua GA Cashaback, Jennifer A Semrau
University of Delaware, Newark, DE, USA
Abstract
Intact upper extremity sensorimotor function is required for the successful performance of daily activities. As a result of stroke, this function is often compromised and requires accurate and objective measurement of sensorimotor impairment to develop effective rehabilitation interventions. Current clinical measures often lack precision, and devices (e.g., robotics) that are objective and sensitive to changes in neurologic function are often impractical for performing in-home assessments and are absent from many clinical environments. We aim to develop a portable, tablet-based, application (App) for quantifying motor control in stroke survivors. Our goal was to validate the developed application and accompanying data analysis against previously validated robotic measures (Kinarm) of upper limb motor function in stroke.
Eleven young controls (YC), 16 older controls (OC), and 7 stroke survivors (SS) performed a Visually Guided Reaching task (VGR) using the Kinarm Robotic Exoskeleton and a Samsung Galaxy tablet. In the robot and tablet versions of VGR, subjects viewed their fingertip as a cursor and were instructed to hold their finger (digitizing pen with tablet) at a central target until one of eight peripheral targets appeared. In the robot task, participants made 10cm reaching movements as quickly and accurately as possible. In the tablet task, they made 6cm movements due to screen size constraints. In both tasks, participants completed 80 trials pseudorandomized to each of the eight targets. Our analyses aimed to determine similarities between motor behavior measured by the robot vs. tablet. We used the previously validated Kinarm Standard Analysis (KSA) and a custom analysis to evaluate Reaction Time (RT) and Initial Direction Error (IDE) as measurements of sensorimotor function. Here, we compared the KSA to our custom analysis to quantify any differences in VGR performance between devices.
Our results demonstrated that sensorimotor behavior measured with the tablet vs. the robot was nearly identical within each of our three subject groups. No significant differences were found for between- device comparisons for RT (p-values, YC=0.58, OC=0.60, SS=0.75) or IDE (p-values, YC=0.29, OC=0.99, SS=0.81), demonstrating that the tablet is suitable for data collection when a robot is unavailable.
Furthermore, no significant differences were found in a parameter-to-parameter comparison between the KSA and our VGR analysis for RT (p-values, YC=0.44, OC=0.06, SS=0.09) or IDE (p-values, YC=0.73, OC=0.06, SS=0.06), suggesting that our analysis is an accurate replication of the validated version of the KSA. Overall, our results suggest that a standard tablet (e.g., iPad, Galaxy Tablet) is a suitable device for measuring upper limb motor function in individuals with stroke. The use of such a device has considerable advantages as it is inexpensive, convenient for both patients and clinicians, and has the potential for collecting large longitudinal datasets in the home or clinic that capture day-to-day sensorimotor recovery after stroke.
Categories
Stroke
Are there wrist-worn sensor metrics that are a better proxy for functional arm/hand behaviors than “activity counts” for chronic stroke survivors?
Marika Demers1, Lauri Bishop1, Justin Rowe2, Daniel Zondervan2, Carolee Winstein1
1University of Southern California, Los Angeles, USA. 2Flint Rehabilitation Devices, Irvine, USA
Abstract
Background: Activity monitors provide a unique opportunity to monitor motor performance of stroke survivors outside clinical and laboratory settings. To date, researchers have mostly been using “activity counts” based on accelerometry thresholds to quantify arm movements. However, activity counts are arbitrary units and may not be intuitive to interpret for someone who has sustained a stroke. Having a simple and intuitive functional measure of arm movements is especially important, if feedback from activity monitors is to be used to motivate behavior change in stroke survivors. This project aims to determine if metrics derived from wrist-worn activity monitors (i.e video annotated movement counts, active movement time) are a better proxy for functional arm/hand behaviors than accelerometry-derived activity counts of chronic stroke survivors.
Methods: Thirty chronic stroke survivors (10 women, age: 57.0 ± 9.6) with mild to severe arm motor impairments (Fugl-Meyer Assessment (FMA): 37.3 ± 17.6, range: 4-66) took part in this study. The Chedoke Arm and Hand Activity Inventory-7 (CAHAI), the FMA and the Reaching Performance Scale (RPS) were administered. For the CAHAI, participants were outfitted with two wrist-worn activity monitors (MiGo, Flint Rehabilitation Devices) and video recorded (V). Each video was annotated to identify arm and hand movements (i.e, video-derived movement countV) and time actively movingV each arm/hand. For each arm, time in active movementM and movement countM were computed from the activity monitors (M), alone and compared to the video recording metrics.
Results: Movement metrics derived from the activity monitors were strongly related to metrics from the video annotations (Active Movement time: r2=0.955; Movement count: r2=0.889). Asymmetries in arm use for bimanual tasks were quantified as a ratio (paretic/total movementsV) and expressed as a percentage of paretic arm use: 26.2 ± 15.6%; and time in active movementV for the paretic arm: 30.6 ± 15.1% (neurotypical symmetry ratio: ~50%1). The relationship between the FMA and movement time and count were strong (Active Movement timeM: r2=0.831, Movement countM: r2=0.859). Similarly, movement metrics derived exclusively from the activity monitors were strongly related to the RPS (Movement timeM: r2=0.779, Movement countM: r2=0.819) and the CAHAI (Movement timeM: r2=0.867, Movement countM: r2=0.886).
Conclusions: Monitor-derived Time in active movement is a valid proxy for functional arm/hand behaviors. We speculate that this metric be more intuitive to interpret for clinicians and stroke survivors than movement or activity count metrics. As a next step, we will explore stroke survivors’ feedback preferences to identify which metrics are the most meaningful and intuitive to motivate behavior change. Insights gained from stroke survivors will help guide future efforts to effect durable gains in paretic arm/hand use in the natural environment.
Categories
Stroke
Contralesional M1 reorganization depends on stroke lesion volume and functional output of M1 of the lesioned hemisphere
Cathrin Buetefisch1, Marc Haut2, Kate Revill3, Scott Shaeffer4, Lauren Edwards4, Deborah Barany4,5, Samir Belagaje4,6, Fadi Nahab4, Neeta Shenvi7, Kirk Easley7
1Departments of Neurology, Rehabilitation Medicine, Radiology, Emory University, Atlanta, USA. 2Departments of Behavioral Medicine and Psychiatry, Neurology, Radiology West Virginia University, Morgantown, USA. 3Department of Psychology, Emory University, Atlanta, USA. 4Department of Neurology, Emory University, Atlanta, USA. 5Department of Kinesiology, University of Georgia, Athens, USA. 6Marcus Stroke and Neuroscience Center, Grady Memorial Hospital, Atlanta, USA. 7Rollins School of Public Health, Emory University, Atlanta, USA
Abstract
After stroke, increases in task-related contralesional primary motor cortex (M1) activation and excitability have been reported1. While this may indicate cortical reorganization, its extent and outcome are incompletely understood. In non-human primates2, 3 and rats4-6, contralesional motor area reorganization is related to the extent of M1 injury (ipsilesional M1, il_M1). Here we test the hypothesis that the extent of il_M1 and/or ipsilesional corticospinal tract (il_CST) damage, as measured by structural MRI (lesion volume, CST lesion load) or the electromyographic response to TMS, determines the extent of contralesional M1 (cl_M1) reorganization and its relationship to affected hand function.
Patients (N= 35, 16M/19F, mean age= 59.0 years) with a single ischemic stroke < 1 month affecting M1 or its output system and corresponding hand paresis underwent MRI of the brain, TMS measures of ipsi- and contralesional M1 excitability, and assessment of hand function with the Jebsen Taylor Hand Function Test (JTT). High-resolution structural scans were collected according to the HCP Lifespan Project protocol7. TMS was applied to contra- and ipsilesional M18. Single TMS pulses were applied to the hotspot of the extensor carpi ulnaris muscle at increasing intensities up to 80% of maximum stimulator output (MSO) to generate a stimulus response curve (SRC) for all patients8. Patients with no measurable response to TMS of il_M1 at 100% MSO were designated MEP-. Linear regression analyses were performed to analyze how a continuous predictor (log-transformed normalized lesion volume, il_CST and il_CSTM1 lesion load), MEP presence, and the interaction between the predictor and MEP presence affected the two primary outcomes: cl_M1 excitability, as measured by the area under the SRC and hand function (JTT). The mean cl_M1 excitability differed between MEP+ and MEP- patients. The relationship between cl_M1 excitability and lesion volume depended on MEP presence: (p=0.055). Cl_M1 excitability increased as lesion volume increased in the MEP- group (p =0.015), but not in the MEP+ group. There were positive but nonsignificant relationships between cl_M1 excitability and both CST and CSTM1 lesion load irrespective of MEP response. Significantly less impaired hand function was seen in MEP+ than MEP- patients. Increasing lesion volume (p=0.026), CST (p=0.015) and CSTM1 (p=0.006) lesion load were significantly related to greater impairment in hand function. Greater Cl_M1 excitability was related to more impaired hand function in MEP- (p=0.005) but not in MEP+ patients. We conclude that during the subacute post-stroke recovery period, the presence or absence of a MEP response seems to indicate critical M1 output with respect to stroke injury related contralesional M1 reorganization and hand function. Consistent with non-human primate results, affected hand function is dependent on CST and CSTM1 lesion load, regardless of MEP presence.
Categories
Stroke
Evaluating the Microbiome to Boost Recovery from Stroke: The EMBRS Study
Tyler Hammond1, Arnold Stromberg1, Lumy Sawaki1, Ai-Ling Lin2
1University of Kentucky, Lexington, KY, USA. 2University of Missouri, Columbio, MO, USA
Abstract
Advances in stroke therapies have lowered stroke mortality, but survivors are often left with severe cognitive and emotional impairments. Accumulating evidence from animal studies suggests that gut microbes modulate brain plasticity via the bidirectional gut-brain axis and may play a role in stroke rehabilitation. An imbalanced microbial community, or dysbiosis, has been shown to occur following stroke, causing a systemic flood of neuro- and immunomodulatory substances due to increased gut permeability and decreased gut motility. Here we measure post-stroke increased gut dysbiosis and how it correlates with gut permeability and subsequent cognitive impairment.
We recruited 12 participants with acute stroke, 12 healthy control participants, and 18 participants who had risk factors for stroke, but had not had a stroke. We measured the gut microbiome with whole shotgun sequencing on stool samples. We measured cognitive and emotional health with the NIH toolbox. We normalized all variables and used linear regression methods to identify gut microbial levels associations with cognitive and emotional assessments.
There was no difference in alpha diversity between the groups. Beta diversity analysis using the Bray- Curtis Index revealed that the bacteria populations of the stroke group were statistically dissimilar from the risk factors (p=0.001) and healthy control (p=0.039) groups on ANOSIM test. Those in the acute stroke group had a significantly lower abundance of proteobacteria (mean 1.11% lower, p=0.009) and verrucomicrobia (mean 1.24% lower, p=0.024). The acute stroke group had higher levels of the leaky gut marker alpha-1-antitrypsin in the stool than the control groups (stroke=1365±774, risk factors=677±476, healthy=500±294, p=0.018). Stroke participants scored lower on the picture vocabulary and list sorting tests than those in the control groups. Stroke participants who had higher levels of roseburia (b= 0.501, p=0.010) and Odoribacter splanchnicus (b=0.243, p=0.039) performed better on the list sorting task. Stroke participants who had lower levels of Clostridium bolteae (b=-0.552, p=0.010), Clostridium symbiosum (b=-0.557, p=0.008) and higher levels of negaviticutes (b=0.407, p=0.030) performed better on the picture vocabulary task. Stroke participants who had higher levels of Lachnospiraceae (b=0.448, p=0.002), Flavonifractor plautii (b=0.945, p=0.001), and Coriobacteriia (b=0.467, p=0.009) reported higher levels of sadness. Stroke participants who had higher levels of alistipes (b=0.281,p=0.040) reported higher levels of pain.
We found that microbial communities are disrupted in a stroke population, showing dissimilarity from the control groups on beta diversity. Some bacteria were associated with markers of recovery. Many of these bacteria have previously been reported to have correlates to health and disease. For example, roseburia is a marker of good gut health, Odoribacter splanchnicus has anti-inflammatory effects, Clostridium bolteae is associated with autism, and alistipes has been associated with depression. This preparatory study will lay the foundation for the development of therapeutics targeting the gut following stroke.
Categories
Stroke
The impact of the COVID-19 pandemic on rehabilitation outcomes and care post-stroke in Quebec
Palak Vakil1,2,3, Perrine Ferré4, Johanne Higgins2,5,6, Louis-David Beaulieu7, Claude Vincent8,9, Kimberley Singerman3, Diana Zidarov2,5,6, Marie-Hélène Milot10, Marie-Hélène Boudrias1,2,3
1McGill University, Montreal, Canada. 2Centre for Interdisciplinary Research in Rehabilitation of Greater Montreal, Montreal, Canada. 3Jewish Rehabilitation Hospital, CISSS-Laval, Laval, Canada. 4Villa Medica Rehabilitation Hospital, Montreal, Canada. 5University of Montreal, Montreal, Canada. 6Institut de réadaptation Gingras-Lindsay-de-Montréal, CIUSSS-CSMTL, Montreal, Canada. 7University of Quebec at Chicoutimi, Saguenay, Canada. 8Center for Interdisciplinary Research in Rehabilitation and Social Integration, Quebec, Canada. 9Laval University, Quebec, Canada. 10Centre de recherche sur le vieillissement, University of Sherbrooke, Sherbrooke, Canada
Abstract
Background: Rehabilitation for persons with disabilities (cognitive/sensory/ motor) due to stroke follows standards of high-quality care. Quebec’s healthcare system has been severely disrupted due to the COVID-19 pandemic, adversely impacting rehabilitation services. Additionally, being infected with COVID-19 can aggravate motor and cognitive deficits impairing recovery.
Purpose: To assess how individuals infected with COVID-19 (COVID+) before or while being admitted to a COVID-19 designated rehabilitation centre affected their care continuum and potential of recovery compared to those that were not infected with COVID-19 (COVID-).
Methods: A retrospective analysis of medical charts is currently being performed in COVID-19 designated rehabilitation centres in Quebec. Preliminary data from 59 COVID+ stroke patients admitted and discharged from Villa Medica Rehabilitation Hospital between February 2020 and May 2021 are presented along with data from an equal number of COVID- patients who were admitted and discharged during the same period. The main outcome measure was the Functional Independence Measure (FIM: total score= 126). Demographic information and descriptive markers of rehabilitation care such as the average length of hospital stay, and the number of treatments received from Physical and Occupational therapists (PT/OT) were also collected.
Results: The number of men and women in each group (Male: Female=28:31) was similar. The two groups were also similar in terms of age (76±12 years for COVID+; 76±10 years for COVID-; p>0.56). The COVID+ group had significantly lower FIM scores on admission compared to the COVID- one (68.3±24.5 vs 79.3±21.4; p<0.03) and at discharge (96.8±25.9 vs 105.8±19.9; p<0.03). The length of stay was significantly longer for the COVID+ group (70.6±38.3 vs 43.4±24.6 days; p<0.001). The absolute number of PT and OT sessions was higher for the COVID+ (PT 39.2±24.1 vs 27.4±17.9; p<0.01; OT 37.9±23.6 vs 25.8±14.7; p<0.002). We found that 56% of COVID+ stroke patients reached the minimal clinically important difference (MCID=22) in their FIM scores at discharge compared to 66% in the COVID- group.
Conclusions: Preliminary results show that being infected with COVID had a significant impact on the rehabilitation outcomes and care post-stroke. COVID+ individuals were more impaired at admission and discharge, and they had a longer length of stay. Also, although they received more PT and OT sessions, the level of independence of COVID+ individuals remained lower at discharge compared to the COVID- cohort.
Implications: Data collection is ongoing to obtain a total of around 100 post-stroke COVID+ patients admitted to 10 rehabilitation centres within Quebec during the pandemic. These preliminary results provide a better understanding of the impact of the pandemic on post-stroke rehabilitation care in terms of the provision of rehabilitation services to maintain/improve the physical condition/function of this population.
Categories
Stroke
NeuroCuresNY: A Novel Clinical Trial Platform to Find Treatments for Chronic Neurologic Disability
M. Cristina Falo1, Marissa Wuennemann1, Amy Bialek1, Susan Wortman-Jutt1, Jeremy Hill2,3, Russell Hardesty2, Timothy Fake2, Jonathan Wolpaw2,3, Bradford Berk4,5, Rajiv Ratan1,6, Tomoko Kitago1,7
1Burke Neurological Institute, White Plains, USA. 2National Center for Adaptive Neurotechnologies, Stratton VA Medical Center, Albany, USA. 3State University of New York at Albany, Albany, USA. 4Aab Cardiovascular Research Institute, Department of Medicine, University of Rochester School of Medicine and Dentistry, Rochester, USA. 5University of Rochester Neurorestoration Institute, University of Rochester School of Medicine and Dentistry, Rochester, USA. 6Feil Family Brain and Mind Research Institute, Weill Cornell Medicine, New York, USA. 7Department of Neurology, Weill Cornell Medicine, New York, USA
Abstract
Background: Chronic neurological disability due to stroke is an area of great unmet need, with more than 7 million stroke survivors in the United States. NeuroCuresNY (NCNY), a multi-center clinical trial platform, was formed to address this need by finding treatments that significantly reduce impairment for patients who are living with the long-term effects of stroke. NCNY leverages the resources and expertise of academia, government, and industry to establish a platform trial to efficiently evaluate the safety and efficacy of multiple experimental treatments for chronic stroke and other neurological disorders.
Pre-clinical and clinical studies suggest that the nervous system is more receptive to training for a short period of time after injury (i.e. the sensitive period). In chronic stroke, improvements with training alone are generally limited; however, therapeutics such as drugs or devices that are designed to protect or repair a circuit, reopen circuit plasticity, enhance axonal conduction, or alter the immune system may reopen the sensitive period to make training more effective. Combinatorial treatments coupling intensive training with plasticity-promoting interventions are therefore a promising approach to enhancing neurological recovery in the chronic stage.
NCNY will use robotic training devices because they deliver standardized, quantifiable, intensive upper- extremity training for stroke survivors focused on impairment reduction, as well as objective measurement of motor outcomes. We hypothesize that combinatorial interventions will lead to greater reductions in upper extremity motor impairment than training alone for chronic stroke survivors.
Participants: Eligible participants will be community-dwelling chronic stroke survivors (at least six months post- ischemic stroke) with residual upper limb weakness.
Study design: All participants will undergo six weeks of robot-assisted upper limb therapy (RT), three times per week, for a total of eighteen sessions.
Participants will be randomized to an intervention, which will consist of RT combined with one of the investigational products (IP), or RT combined with placebo/sham. There may be multiple IPs that are tested within the platform at a given time.
Outcome assessments will be conducted before training, mid-training, end of training, and at 6 and 12 weeks post-training. The primary outcome measure will be the change in upper limb motor impairment (Fugl-Meyer) while secondary outcomes will be change in Action Research Arm Test and Stroke Impact Scale. Exploratory outcomes will be the change in upper limb kinematics.
Conclusions: NeuroCuresNY will facilitate the discovery of novel combinatorial treatments to reduce chronic neurological disability.
Categories
Motor Rehabilitation
An objective method for analyzing ipsilateral motor evoked potentials (iMEPs) in stroke survivors with severe upper limb hemiplegia
Akhil Mohan1, Xin Li1, Jayme S Knutson2, Morgan Widina1, Bei Zhang3, Ela B Plow1, David A Cunningham2
1Cleveland Clinic Lerner Research Institute, Cleveland, USA. 2MetroHealth Rehabilitation Institute, Cleveland, USA. 3Case Western Reserve University, Cleveland, USA
Abstract
Introduction: Ipsilateral motor evoked potentials (iMEPs) elicited using transcranial magnetic stimulation (TMS) are often used to measure excitability of ipsilateral (uncrossed) pathways to paretic muscles1,2,3. Evidence indicates that ipsilateral (uncrossed) pathways may serve as alternate resources to support recovery in stroke survivors with insufficient (crossed) ipsilesional substrate4,5,6. However, there are no widely accepted methods to define the presence or absence of iMEPs. Both manual and semi-automatic (objective) methods have been used but there is no clear and consistent guideline on which method to choose and why or if multiple methods should be combined7. Here, we investigated the agreement between objective and manual methods with the goal of providing precision to subjective manual methodology and minimize the time and investigator effort required for the analysis.
Methods: Twelve chronic stroke survivors with severe upper limb hemiplegia (Upper Extremity Fugl- Meyer (UEFM) mean (SD) = 22.0(9.6)), characterized by lack of 10-degree extension in wrist or fingers, underwent iMEP testing using TMS. IMEPs were collected from the paretic biceps brachii (BB) by delivering suprathreshold TMS (30-40 pulses) to contralesional motor areas. Participants were asked to maximally contract the paretic BB and the non-paretic triceps brachii simultaneously while turning their neck to the paretic side, a method that elicits a higher number of and larger iMEPs3. Two experienced investigators subjectively measured the presence and absence of iMEPs (manual method), which involved identifying an MEP waveform in the ipsilateral biceps muscle 3-10ms after the appearance of an MEP waveform in the contralateral (non-paretic) biceps muscle. Four objective methods based on standard deviation (SD) were used to measure the presence or absence of iMEPs. Reliability (% agreement) of the objective methods was compared against the manual method.
Results: Percent agreement of detecting iMEP presence or absence across two raters was 81.9%, indicating strong inter-rater agreement for manual method. Of the objective methods, iMEP detection performed using criterion threshold of background mean electromyography (EMG) + 2SD and background mean EMG + 3SD methods yielded the best percent agreement with the manual method (72.5% and 76.1% respectively), indicating moderate agreement.
Discussion: There is room for improvement for current objective methods to meet the percent agreement based on the manual method. Future work to improve the objective method (e.g., patient-specific feature detection and optimization) remain to be explored.
Categories
Motor Rehabilitation
Effects of Rhythmic-Based and Tonal-Based Music Interventions on Upper Extremity Movements in Individuals with Parkinson’s Disease: A Scoping Review
Yi-An Chen, Emily Bell, Julia Baker, Meredith Parrott, Jessica Rosales
Georgia State University, Atlanta, USA
Abstract
Background: Emerging evidence of music-based interventions, including rhythmic cueing and tonal music listening/playing, has been shown to alleviate motor symptoms and improve functioning in individuals with Parkinson’s Disease (PD). Rhythmic-based interventions utilize external beats to influence motor execution; tonal-based interventions incorporate melody and harmony of the music to motivate and affect movement performance. However, the majority of the PD studies exclusively focused on rhythmic cueing for gait, which is a naturally paced movement. As the upper extremity (UE) movements are more varied and may be less rhythmed (e.g., dressing, handwriting), we wonder whether different types of intervention based on different mechanisms may provide distinct benefits in UE movements.
Objective: To examine the effects of rhythmic- and tonal-based music interventions on UE movements and functions in individuals with PD.
Methods: A systematic literature search was completed in August 2021 with five databases (Cochrane, PubMed, Medline, CINAHL, and Embase). We identified a total of 23 articles that include (1) music-based interventions with rhythmic and/or tonal components, and (2) UE outcome measures in individuals with PD. Included articles were divided into three categories: rhythmic-based interventions (n=12), tonal- based interventions (n=7), and combined interventions (tonal component with a superimposed rhythmic cue, n=4).
Results: Studies using rhythmic-based interventions (e.g., metronome, drumbeats) consistently showed positive results in UE movements, including finger tapping (n=9), finger-to-thumb opposition (n=2), grasping (n=1), supination/pronation (n=1), and thumb flexion/extension (n=1). Three of these studies also found that faster-paced rhythms are more effective at improving UE finger movements (e.g., tapping, opposition) than slower rhythms. Tonal-based interventions that used newly created novel music demonstrated similar positive UE improvements in finger tapping (n=3) and wrist movement (n=1). However, studies (n=3) that used songs that participants are familiar with demonstrated no significant improvements, possibly due to distraction. We also found that studies that assessed UE functions (n=3), such as spoon-use or self-feeding tasks, showed no improvements. Lastly, studies that used the combined interventions demonstrated significantly positive results in finger tapping (n=3) and handwriting function (n=1).
Conclusion: Overall, only a small number of music-based intervention studies (n=23) examined UE movements and functions in PD. Limited evidence showed that all types of music-based interventions demonstrated benefits in improving discrete and rhythmed UE movements (e.g., finger tapping).
However, only 4 studies measured functional outcomes, and simply 1 article showed positive results (in handwriting). More direct examination of the training effects on daily functions is needed to support the use of music-based interventions in PD. Additionally, faster rhythms and novel music may better improve PD patients’ UE movements. Yet, more studies and further investigation of the mechanisms are needed to identify essential components of music-based interventions to promote the efficacy of the treatments for PD.
Categories
Motor Rehabilitation
Association between online motor-cognitive game performance and APOE e4 carrier status among older adult Mindcrowd users
Andrew Hooyman1, Matt Huentelman2, Sydney Schaefer1
1Arizona State University, Tempe, USA. 2The Translational Genomics Research Institute, Pheonix, USA
Abstract
A major challenge in pre-clinical Alzheimer’s Disease (AD) clinical trials is identifying individuals who are most likely to develop AD or will experience significant cognitive decline during the trial period. Because people who are carriers of the APOE e4 gene are at higher risk of experiencing cognitive decline and dementia than non-carriers, APOE genotyping is commonly used as a screening or ‘enrichment’ tool for AD clinical trial enrollment. There is growing concern, however, regarding genetic testing among asymptomatic individuals. Emerging evidence from our laboratory suggests that motor behavior may provide unique insight into the progression of AD, which could serve as a low-cost biomarker. Thus, the purpose of this study was to examine if performance on a web-based motor-cognitive game (Super G) is associated with APOE e4 carrier status. The goal of the Super G game is to transport an astronaut avatar (in two dimensions) from a start planet to a goal planet using the arrow keys on a computer keyboard in 4.5 seconds or less. This study is in collaboration with MindCrowd, an online initiative aimed at recruiting one million people as part of an AD research repository. We emailed 662 older adults whose APOE e4 carrier status is banked within the MindCrowd repository a link to the Super G website requesting them to play a minimum of 75 trials (equaling ~5.6 minutes of play time). Overall, 58 participants (APOE e4 carrier = 23) completed the Super G task (mean age = 56.8 ± 7.7; 42 females). Data showed that Super G performance (measured as the time at which the astronaut leaves the start planet, i.e., response time) significantly predicted APOE e4 carrier status (p = .014) while controlling for age, cognition, sex, and education. Interestingly, APOE e4 carriers demonstrated better response time than non-carriers. This is consistent with recent research demonstrating an APOE-e4 advantage on a visual working memory task, relative to non-e4 carriers. In conclusion, results of this study suggest that performance on the web- based Super G task could be an alternative screening or enrichment tool for AD clinical trial recruitment.
Categories
Cognitive/Language Rehabilitation
Estimated functional connectivity derived from clinical MRI predicts performance on a cognitive-IADL measure after acute stroke
Abhishek Jaywant1, Joan Toglia2, Zijin Gu3, Keith Jamison1, Faith Gunning1, Amy Kuceyeski1
1Weill Cornell Medicine, New York, USA. 2Mercy College, Dobbs Ferry, USA. 3Cornell University, Ithaca, USA
Abstract
Introduction: Stroke is associated with changes in resting state functional connectivity, which predict cognitive function in the subacute and chronic stages of recovery. How functional connectivity is associated with acute cognitive changes after stroke remains relatively unknown, but has important implications for understanding the neural mechanisms underlying early cognitive dysfunction and developing early interventions when neuroplastic potential is high. Major barriers to resting state fMRI in acute stroke are cost and feasibility, particularly for patients who are too ill or cannot tolerate lengthy scans. In this preliminary analysis, we applied a novel technique to estimate functional connectivity (eFC) from lesion masks derived from clinically acquired structural MRI. We explored how eFC in regions of the executive control network (ECN) correlated with performance on a measure of cognitively-based instrumental activities of daily living (cognitive-IADL).
Methods: Participants were 11 individuals with stroke acutely hospitalized at NewYork-Presbyterian Hospital/Weill Cornell Medical Center. Participants underwent standard-of-care clinical (structural) MRI on average 3.1±2.6 days after onset of stroke symptoms. We created lesion masks and used the Network Modification (NeMo) Tool to estimate the impact of the lesion on each participant’s structural connectome. Each participant’s estimated structural connectome then served as input to a previously published deep neural network that estimates functional connectivity (Tozlu et al., 2021 NeuroImage: Clinical). We calculated regional eFC node strength by averaging eFC in each ROI of the ECN. Participants completed the Weekly Calendar Planning Activity (WCPA), a standardized, ecologically-valid, and performance-based measure of cognitive-IADL. The WCPA assesses planning, working memory, flexibility, and inhibition by requiring examinees to input appointments into a mock weekly calendar/schedule while following a set of specific rules and inhibiting distractions. Performance variables analyzed included the proportion of appointments entered correctly (accuracy) and the number of rules correctly followed. Spearman correlation coefficients were used to explore the association between regional eFC node strength and performance on the WCPA.
Results: Accuracy on the WCPA was negatively associated with eFC node strength in the left (r = -.67, p = .047) and right superior frontal cortex (r = -.59, p = .096), and with eFC node strength in the left rostral middle frontal cortex (r = -.65, p = .06). Number of rules followed on the WCPA correlated negatively with eFC node strength in the left superior frontal cortex (r = -.78, p = .01).
Conclusions: Functional connectivity in the ECN, estimated from standard-of-care clinical MRI, correlates with cognitive-IADL ability acutely after stroke. Negative correlation coefficients suggest that superior and middle frontal regions may exhibit maladaptive compensation in the setting of structural lesions (i.e., increased eFC associated with worse performance). Novel therapeutics targeted at the ECN may be relevant in promoting early cognitive recovery after stroke.
Categories
Stroke
Evidence of excessive hip extension during a step-up task as compensation for distal joint impairment in individuals with bilateral cerebral palsy
Vatsala Goyal, Theresa Sukal-Moulton
Northwestern University, Chicago, IL, USA
Abstract
Introduction: Individuals with bilateral cerebral palsy (CP) are often affected by motor impairments in the knee and ankle joints, consistent with damage to corticospinal tracts. Previous literature has shown that the hip may compensate for distal joint impairment during steady-state walking. Community ambulation requires more challenging gait tasks such as stepping up, but research investigating these tasks is limited. We hypothesized that compared to individuals with typical development, individuals with bilateral CP would generate less knee and ankle moments and more hip moments during a step-up task.
Methods: We recruited individuals aged 5-19 years with and without bilateral CP who were able to step up without assistive devices or orthotics. Participants were outfitted with reflective markers and were instructed to step up onto a 10.2-cm platform 5-15 times, leading with either the dominant foot (without bilateral CP) or the most-affected foot (bilateral CP). Data from a motion capture system and force plates were used to compute lower limb joint moments throughout the experiment. Hip, knee, and ankle joint moments in the sagittal plane were normalized by body weight and summed to calculate the support moment, defined as the extension moment required to keep the body upright. We identified the maximum support moment during two stance phases of the step-up task: the push-off phase, which typically requires large ankle plantarflexion moments, and the pull-up phase, which typically requires large knee extension moments. Hip, knee, and ankle contributions to support moment were averaged to descriptively compare the groups.
Results: During the push-off stance phase, participants without CP (n = 11, age = 12.7 years) generated -0.024 Nm/kg of hip extension, while participants with CP (n = 5, age = 10.6 years) generated -0.645 Nm/kg of hip extension. Participants without CP also generated +0.076 Nm/kg of knee flexion and -0.857 Nm/kg of ankle plantarflexion, while participants with CP generated +0.061 Nm/kg of knee flexion and -0.596 Nm/kg of ankle plantarflexion. During the pull-up stance phase, participants without CP generated -0.192 Nm/kg of hip extension, while participants with CP generated -0.830 Nm/kg of hip extension. Participants without CP also generated -0.828 Nm/kg of knee extension and -0.251 Nm/kg of ankle plantarflexion, while participants with CP generated -0.382 Nm/kg of knee extension and -0.356 Nm/kg of ankle plantarflexion.
Discussion: Compared to individuals with typical development, individuals with bilateral CP showed a marked decrease in ankle plantarflexion moments during the push-off phase and knee extension moments during the pull-up phase. There was also a corresponding increase in hip extension moments during both stance phases, as has been seen in steady-state walking. This is likely a neural and biomechanical compensation for distal joint impairment.
Categories
Other
Multimodal Longitudinal Assessment of Infant Brain Organization and Recovery in Perinatal Brain Injury
Ellen Sutter1,2, Catarina Saiote2, Ryan McAdams2, Douglas Dean III2, Raghavendra Rao1, Michael Georgieff1, Bernadette Gillick2,1
1University of Minnesota, Minneapolis, USA. 2University of Wisconsin-Madison, Madison, USA
Abstract
Background and Objectives: Perinatal brain injury, including stroke or brain bleed, is a common cause of cerebral palsy (CP). 50-75% of infants with perinatal stroke will develop life-long motor impairment.
Infancy is a key period of motor system development and organization, particularly of the corticospinal tracts (CSTs). The CST is the primary pathway by which the cerebral cortex controls movement, and its organization impacts long-term motor function. However, little is known about the interaction between recovery after perinatal brain injury, CST organization, and resultant motor outcome during infancy. This gap hinders the development of strategies for earlier diagnosis and interventions during a period of heightened neuroplastic potential, with impact on lifelong outcomes.
Transcranial magnetic stimulation (TMS) is a non-invasive brain stimulation technique that can safely assess the excitability of the motor pathway, as a biomarker of tract functional integrity. TMS also provides insight into corticospinal circuitry and neuroplastic changes in connectivity that may occur after early brain injury. In pioneering studies, our laboratory has established the safety and feasibility of TMS in children and infants. To further understand the mechanisms underlying development of brain circuitry, we can additionally measure CST structural integrity with diffusion Magnetic Resonance Imaging (MRI).
By examining these multimodal biomarkers in relation to motor development and CP diagnosis, we may improve understanding of neurodevelopmental trajectories after early brain injury, with implications for early diagnosis and intervention.
Description: We present a protocol for a multi-modal longitudinal assessment of brain development and motor function in infants, following perinatal stroke/brain bleed. The protocol incorporates assessments at five separate ages (corrected for prematurity) in early development: 0-2, 3-6, 12, 18, and 24 months. We will use TMS (after 3 months) and MRI to measure CST functional and structural integrity, and development of the motor pathway circuitry during this period. We will also assess the association between cortical excitability and integrity and motor function/outcomes, examined with the General Movements Assessment, the Hammersmith Infant Neurological Examination, and the Bayley Scales of Infant and Toddler Development. The study implementing this protocol is currently funded by the National Institute of Child Health and Human Development.
Significance: This multimodal early assessment protocol will lead to identification of unique bioindicators related to motor outcome and neuroplasticity for infants with perinatal brain injury. It will subsequently inform early detection, diagnosis, and intervention and improve understanding of brain development after perinatal brain injury.
Categories
Motor Rehabilitation
Structural Neural Correlates of Objective and Patient-Reported Measures of Function and Health Status After Stroke
Julie DiCarlo1,2, Kimberly Erler3, Abhishek Jaywant4, Perman Gochyyev3, Jessica Ranford1, Steven Cramer5,6, David Lin1,2
1Massachusetts General Hospital, Boston, USA. 2Department of VA Medical Center, Providence, USA. 3MGH Institute for Health Professions, Boston, USA. 4Weill Cornell Medicine, New York, USA. 5University of California, Los Angeles, USA. 6California Rehabilitation Hospital, Los Angeles, USA
Abstract
Introduction: Patient-reported outcome measures provide valuable insights into health status and function, but their relationships with objective outcome measures remain incompletely understood. Here we evaluate the relationship between these two classes of measures using dimensionality-reduction techniques in patients after acute stroke and examine their associated patterns of neuroanatomical injury.
Methods: Sixty-four adults with upper extremity motor deficits after stroke were serially assessed (acute, sub-acute, early-chronic, chronic) using objective motor-based measures (Upper Extremity Fugl- Meyer, Barthel Index, modified Rankin Scale, Box and Blocks, 9- Hole Peg, Grip Strength) and patient- reported measures (PROMIS-Global Physical, Mental, and Social Health, Patient Health Questionnaire-9) of function and health status. At each timepoint after stroke, exploratory factor analysis was performed to identify the underlying factorial structure of these 10 measures by evaluating the relationships between them. As a validation step, confirmatory factor analysis was then performed to further confirm and evaluate the reduction of the dimensionality of the response data into a smaller set of extracted factors. Voxel-Based Lesion Symptom Mapping was used to examine the relationship between factors and patterns of neuroanatomical injury.
Results: In the battery of stroke outcome measures, two uncorrelated factors were identified and retained, accounting for ≥78% of the overall variance across outcomes at every timepoint. Objective measures loaded onto Factor 1 separately from patient-reported measures which loaded onto Factor 2. Results were consistent at each serial timepoint after stroke. Objective motor-based measures were related to injury to subcortical brain regions, particularly the corticospinal tract, while patient-reported measures were related to injury to cortical brain regions including the insula and inferior parietal lobe.
Discussion: Two distinct factors representing objective and patient-reported measures of function and health status were extracted from a battery of stroke outcome measures scored across the first year post-stroke. Each factor was associated with injury to brain regions concordant with the content of the represented assessments. These findings emphasize the distinct behavioral elements and neuroanatomical underpinnings for objective and patient-reported outcome measures after stroke and have potential implications for targeted therapies.
Categories
Stroke
The Feasibility of a Remote Physical Activity Monitoring Program for Rural Veterans with Stroke or Parkinson’s Disease
Kimberly Waddell1,2, Mitesh Patel2,1,3, Jayne Wilkinson1,2, Robert Burke1,2, Sreelatha Koganti1, Stephanie Wood1, James Morley1,2
1Crescenz VA Medical Center, Philadelphia, USA. 2University of Pennsylvania, Philadelphia, USA. 3Ascension Health, St. Louis, USA
Abstract
Background: Rural Veterans represent nearly half of the Veterans Affairs patient population but experience limited access to high-technology care, report worse physical health and quality of life (QOL), and are underrepresented in research. Rural Veterans with chronic neurological conditions, such as stroke or Parkinson’s disease (PD), are particularly vulnerable to poor outcomes due to limited access to specialized care. Regular physical activity can improve overall health, QOL, and mitigate motor and non- motor symptoms of disease. Little is known about the daily physical activity levels, and predictors of activity, of rural Veterans with stroke or PD.
Purpose: To examine the feasibility of a remote physical activity monitoring program and predictors of daily physical activity among rural Veterans with stroke or PD.
Methods: This remote, observational study enrolled Veterans with a rural or highly-rural VA designation and a history of stroke or PD. Potential participants were identified using administrative data, mailed a recruitment letter, contacted the study team via phone, and provided oral consent. Potential predictors of physical activity (e.g. self-efficacy, apathy, social support) were quantified with surveys completed online via Way to Health, a remote monitoring platform. Participants were issued a wearable device (Fitbit Inspire HR) to wear for 30 days, and synced their daily step counts to the Way to Health platform. Program feasibility was assessed by recruitment response rate, study completion rate, and device adherence. Individual and group daily steps were evaluated descriptively. We examined predictors of physical activity using a linear mixed effects regression model that adjusted for participant random effect and diagnosis. Predictors of interest included sociodemographic, clinical, and survey data.
Results: The cohort was comprised of 40 rural-dwelling Veterans (20 stroke, 20 PD), who resided a mean (SD) distance of 252 (105) miles from the study location. Recruitment response rates were 6% for stroke and 11% for PD. Veterans who did not respond were demographically similar to those who enrolled.
Twenty-five individuals with stroke consented to participate (80% completion rate) and 28 consented with PD (71% completion). The 30-day device adherence rates were similar for both cohorts (95.2% stroke; 97.0% PD). Individuals with stroke had a mean (SD) 6447 (5532) steps per day and those with PD achieved 3476 (3269) steps. Results from the adjusted analysis indicated that, after adjusting for diagnosis, age was the only significant predictor of daily steps (-265 steps, 95% CI [-401, -123], P < 0.001).
Conclusions: A remote physical activity monitoring program for rural Veterans with stroke or PD was feasible. Veterans who were older took significantly fewer steps per day. The evolution of scalable, remote monitoring technology affords new opportunities to extend care to rural-dwelling individuals that may help improve self-management of chronic health conditions and attenuate rising healthcare costs.
Categories
Stroke
Voluntary Muscle Activation Increases the Threshold at which an Electrical Stimulus Is Detected Post-Hemiparetic Stroke: Preliminary Findings
Ninghe Cai, Alan Duong, Eileen Medina, Netta Gurari
Northwestern University, Chicago, USA
Abstract
Accurately perceiving tactile stimuli is necessary to seamlessly interact with and manipulate objects. In individuals who are neurologically intact, voluntarily activating one’s muscles increases the threshold at which a tactile stimulus is detected. This gating of tactile signals during voluntary muscle activation is postulated to occur in the cuneate nucleus based on animal studies. In individuals post-hemiparetic stroke, the integrity of white matter tracts at the cuneate fasciculi is decreased on the contralesional side. This decreased integrity could impact the transmission and gating of tactile signals and, subsequently, tactile perception at the paretic arm. Studies have reported deficits in detecting tactile stimuli at the paretic arm when an individual is at rest; however, it remains unaddressed whether tactile perception is further impacted during voluntary muscle activation. Consequently, our study investigates the effect of stroke on one’s perception of a tactile stimulus when they voluntarily activate their arm muscles.
Six individuals with hemiparetic stroke (µ± age: 62±4 years) and six similarly-aged individuals without neurological impairments (µ± age: 64±9 years), i.e., controls, participated. All controls and four participants with stroke were right arm dominant. The participants with stroke were assessed as having severe to mild levels of motor impairments according to the upper extremity Fügl-Meyer motor assessment (UE-FMA) (µ±/range: 40±18/16-57). Participants sat in a Biodex chair with their testing forearm fixed to a mechatronic device. Participants received a single-pulse, 100µs square wave electrical stimulus at the palmar aspect of the fingertip of their testing arm’s middle finger. Participants reported when they detected this electrical stimulus at rest (baseline) and while isometrically flexing about the elbow to 25%, and 50% of their maximum voluntary torque (MVT). The detection threshold was identified using an up-down staircase method.
The µ± detection threshold in controls for the baseline, 25%MVT, and 50%MVT was 4.52±1.29mA, 5.06±1.10mA, and 5.40±1.24mA at the dominant fingertip and 3.85±1.07mA, 3.91±0.93mA, and 4.52±1.22mA at the non-dominant fingertip. The detection threshold in participants with stroke for the baseline, 25%MVT, and 50%MVT was 4.29±1.35mA, 4.67±1.31mA, and 4.60±1.40mA at the non-paretic fingertip and 5.60±1.80mA, 6.27±2.36mA, and 6.75±2.46mA at the paretic fingertip.
These preliminary results show that detection of the electrical stimulus was poorer at the paretic fingertip than at the non-paretic fingertip and both fingertips of controls when at rest. Additionally, voluntary muscle activation made detection of the electrical stimulus poorer for both the controls and participants with stroke, with the greatest effect at the paretic fingertip of the participants with stroke. These initial findings on how stroke impacts tactile perception during voluntary muscle activation are relevant for considering how to assess deficits and provide effective rehabilitation treatments to survivors of a stroke.
Categories
Stroke
Paired stimulation targeting spinal cord is more effective than targeting sensorimotor cortex
Ahmet Asan, Ajay Pal, Jason Carmel
Columbia University, New York, USA
Abstract
Timed co-activation of sensory and motor systems is used to strengthen their interaction, but the best site of convergence is not known. The most common paired stimulation protocol used in humans, paired associative stimulation, targets the sensorimotor cortex. We have shown that pairing suprathreshold motor cortex stimulation with subthreshold spinal cord (SC) stimulation strongly augments the resulting muscle response. Here we compared convergence in the motor cortex with convergence in the SC by altering the latency between the two sites of electrical stimulation. We hypothesized that converging these two stimuli in the SC would generate a stronger increase in evoked muscle responses.
To test our hypothesis, we measured the latencies between the SC and motor cortex and used these to time paired stimulation to converge at each site. In rats, we implanted three sets of electrodes: epidural screw electrodes over each motor cortex, electrodes in each biceps muscle, and epidural electrodes over C5-C6 spinal cord. All recordings were done in freely moving adult rats. As a stimulation protocol, we initially stimulated the motor cortex and SC using 110% and 90% of their thresholds, respectively. We also tested whether subthreshold stimulation (90%) at each site would produce a muscle response when paired. Finally, we repeatedly paired brain and SC stimulation for 30 min to determine whether this produced plasticity.
Paired stimulation timed to converge in the SC was more effective than convergence in the motor cortex. SC potentials evoked by motor cortex stimulation were recorded at 9.3±0.3ms. The cortical potentials evoked by SC stimulation had an onset of 7.2±0.25ms with a peak potential at 20ms. Subthreshold SC stimulation augmented cortical muscle responses at multiple time points, with the highest peak when the cortex was stimulated 10ms before the SC (155%±27%, p < 0.0001 compared with baseline). There were also smaller increases in muscle responses when both stimuli converge in the cortex (17%±6%, p = 0.019). When compared directly, the spinal convergence was significantly stronger than the cortical convergence (p = 0.003). Targeting the SC with 30 min repetitive paired stimulation doubled cortical MEPs immediately after stimulation, and they were still ~50% elevated at 120 minutes. Neither the control conditions (motor cortex alone, SC alone, or paired with SC stimulation after a 100 ms delay) nor convergence at the cortex significantly increased cortical MEPs (p=0.20). Applying subthreshold stimulation to both sites also only evoked muscle responses when they were timed to converge in the cord.
Together, these results support the hypothesis that timing paired stimulation to converge in the SC was optimal. The specificity of the approach and the large effect size make this a good candidate for study as a therapeutic protocol, spinal cord associative plasticity.
Categories
Motor Rehabilitation
Feasibility of error augmentation feedback for upper limb rehabilitation in stroke survivors
Caroline Rajda1, Sigal Berman2, Shelly Levy-Tzedek2, Philippe Archambault1, Farnaz Ghazali Jahromi1, Mindy Levin1
1McGill University, Montreal, Canada. 2Ben-Gurion University, Negev, Israel
Abstract
Many people lose function of their upper limb (UL) after stroke. Despite intensive rehabilitation, many patients do not attain their maximal potential recovery of function.1 Treatment interventions aimed at helping patients regain function rely on motor learning principals including different forms of feedback to the learner.2
Error augmentation (EA) is a feedback modality, based on implicit error information. It may improve motor learning by providing enhanced intrinsic feedback that is used by the nervous system to adapt movement to motor errors.3 The enhanced intrinsic feedback magnifies the participant’s motor errors so that it appears as if the limb or joint is moving less than in reality. Therefore, participants are encouraged to use larger movement ranges to attain the motor goal. EA has been implemented in UL training for patients with chronic stroke resulting in improved movement range, precision, and speed. However, these changes have not been accompanied by full UL functional improvement.4,5 One explanation is that patients with stroke do not incorporate all of their available elbow extension range during reaching movements and thus, may have limitations in functional UL tasks.6
We implemented an EA training paradigm in virtual reality (VR) to provide a challenging and motivating training environment and to promote motor learning aimed at increasing the functional range of elbow extension during reaching. The objectives were to: 1) develop a personalized virtual rehabilitation program incorporating EA feedback, and 2) evaluate the feasibility of the EA feedback training protocol. The EA program consists of an interactive game that involves reaching for targets that appear randomly within a calibrated workspace.
Twelve healthy participants (≥ 18 y/o) trained with the EA program 3x in one week. The EA feedback caused participants to perceive their hand position as undershooting the target, encouraging them to incorporate more elbow extension into the whole arm reaching task. Reaching kinematics were assessed at 3 time points using a standardized reaching task. Outcome measures were the index of performance (IP), a measure of movement quality based on endpoint speed and precision, the level of fatigue, and the degree of motivation in performing the task assessed by questionnaires.
The EA program was successfully implemented in a 2D VR environment with appropriate depth cues. The IP of the standardized reaching task improved following training. Participants reported that the EA program was not fatiguing and that it held their interest throughout the training, demonstrating the program’s feasibility. We continue to test the feasibility and effectiveness of the EA program in individuals with stroke and to investigate the role of cognition on motor learning. This project will inform the design of effective personalized training protocols that use enhanced feedback for people with UL sensorimotor deficits following stroke.
Categories
Motor Rehabilitation
Enabling Unsupervised Closed-loop Vagus Nerve Stimulation During Rehabilitation for Stroke or Spinal Cord Injury
Joseph Epperson1,2, Eric Meyers1, David Pruitt1, Joel Wright1, Rachael Hudson1,3, Emmanuel Adehunoluwa1,3, Y-Nhy Nguyen-Duong1,3, Chad Swank4, Christi Stevens4, Jaime Gillespie4, Dannae Arnold4, Jane Wigginton1,4, Robert Rennaker1,2,3, Michael Kilgard1,3, Seth Hays1,2
1Texas Biomedical Device Center, Richardson, USA. 2Erik Jonsson School of Engineering and Computer Science, Richardson, USA. 3School of Behavioral and Brain Sciences, Richardson, USA. 4Baylor Scott and White Institute for Rehabilitation, Dallas, USA
Abstract
Many stroke and spinal cord injury survivors are left with chronic motor disability. Paired VNS was recently approved by the U.S. Food and Drug Administration for the treatment of upper extremity motor deficits associated with chronic ischemic stroke. Three clinical trials indicate that vagus nerve stimulation (VNS) combined with supervised physical rehabilitation after stroke leads to increased recovery of upper limb motor function, but additional performance gains are not observed with continued unsupervised therapy. Development of tools to increase repetitions and enhance VNS timing may lead to improved outcomes. To address this, we developed RePlay, a novel mobile system to facilitate unsupervised rehabilitation exercises, and incorporated a dynamic algorithm that can automatically trigger stimulation based on selected parameters of movement.
RePlay leverages accessible technology to provide a simple tool that allows users to perform common rehabilitative exercises in a gameplay environment. RePlay collects quantitative time-series force and movement data from handheld devices and inputs the information to the dynamic algorithm for real-time VNS triggering based on selected parameters of movement. The dynamic algorithm produces reliable triggering near a desired interval while triggering during movements of the largest amplitude. The dynamic algorithm continually adapts over time to adjust for intermittent periods of rest and person-to- person variability and can flexibly be applied to signals from various controllers (handheld sensors, touchscreen, and keyboard) while maintaining high triggering selectivity.
The prospective benefits of RePlay are being studied in two ongoing double-blind randomized placebo controlled clinical studies in chronic stroke and spinal cord injury participants. Preliminary evidence indicates RePlay combined with VNS may reduce upper limb deficits in spinal cord injury. The RePlay system and its dynamic algorithm represent novel tools to facilitate access to unsupervised rehabilitative exercises with closed-loop neuromodulation that may improve overall patient outcomes.
Categories
Motor Rehabilitation
Aphasia outcomes are modulated by lesion size and race in chronic stroke survivors
Davetrina Gadson, Candace van der Stelt, Elizabeth Lacey, Andrew DeMarco, Sarah Snider, Peter Turkeltaub
Georgetown University School of Medicine, Washington DC, USA
Abstract
Black Americans experience greater disparities in stroke and aphasia outcomes than any other ethnic group. Previous studies have compared Black and White stroke survivors with aphasia (SWA) and found differences in language severity and functional independence at stroke onset. These racial differences may reflect biological stroke severity, potential measurement issues, or interactions between these factors and race in aphasia outcomes. In order to identify the most effective ways to promote health equity for people with aphasia, we must first understand the origin of disparities in stroke recovery. In this study, we examined relationships between the Western Aphasia Battery- Revised (WAB-R), race, SES (income and education), and neurological factors (leukoaraiosis and lesion size).
Eighty-seven chronic left-hemisphere SWA (31 Black, 56 White) participated in the study. The primary outcome measure was the WAB-Revised. The WAB-R categorizes language severity by assessing four modalities to provide an aphasia quotient (AQ): spontaneous speech, auditory verbal comprehension, repetition, and naming and word finding. Lesion size was measured based on manual lesion segmentations. FLAIR and T2 images were scored for the severity of leukoaraiosis (Manolio, 1994).
Independent sample t-tests were used to determine differences in SES, WAB-R, lesion size, and leukoaraiosis between Black and White SWA. A linear regression model was used to explore factors that predicted aphasia severity on the WAB-R.
Preliminary findings of our study found significant SES racial differences in education t(85)=3.76, p<.001 and median income t(85)=3.48, p=.001. For predictors of aphasia severity, the regression model revealed significant effects of lesion size on WAB AQ (β = -0.71, t(67) = -6.9, p < .001) and an interaction of race x lesion volume (β = -0.22, t(67) = -2.2, p = .035). We found an interaction between race and lesion size such that Black and White participants with small lesions have similar WAB scores, but in individuals with larger lesions, Black participants have lower WAB scores than White participants. The overall model fit was R2 = .49, F(9,68) = 7.33, p < .001.
In conclusion, we suggest two explanations for the difference between Black and White SWA. First, disparities in access to care at stroke onset may be more evident in larger SWA and subsequently impact rehabilitation. Or there is an assessment bias in aphasia outcomes that may not consider behavioral language performance unique to Black Americans, like African American Vernacular English. In this preliminary study, we are unable to definitively determine the exact reason for the difference between Black and White SWA. Therefore, future studies should aim to better discern the reasons for the disparities in aphasia outcomes.
Categories
Stroke
Persistent asymmetry of aperiodic resting-state neural activity in both cortical and sub-cortical strokes
Richard Hardstone1, Lauren Ostrowski1, Aliceson N. Dusang2, Catherine Chu1, Sydney S. Cash1, Steven C. Cramer3,4, Leigh R. Hochberg1,2, David J. Lin1,2
1Center for Neurotechnology and Neurorecovery, Department of Neurology, Massachusetts General Hospital, Boston, MA, USA. 2VA RR&D Center for Neurorestoration and Neurotechnology, Department of Veterans Affairs Medical Center, Providence, RI, USA. 3Department of Neurology, University of California, Los Angeles, CA, USA. 4California Rehabilitation Institute, Los Angeles, CA, USA
Abstract
Background: Stroke is a leading cause of acquired adult disability. Understanding neuroanatomical and neurophysiological changes post-stroke will be key to developing new approaches to neurorehabilitation. In addition to the neural tissue that is directly lesioned by a stroke, there is also surrounding pathological physiology that impacts recovery and functional deficits. One outstanding issue in understanding this abnormal physiology is the influence of stroke lesions on the relative balance between excitation and inhibition (E/I). E/I can be estimated from EEG using the slope of the aperiodic component, where a steeper slope (relatively more power in lower frequencies than higher) implies decreased E/I. Here we investigate how stroke lesion characteristics affect E/I during the acute and sub-acute periods of recovery in a cohort of patients with heterogeneous cortical and subcortical lesions.
Methods: 16 patients with unilateral stroke (Age: 59.5+/-12.0 years mean+/-std. NIHSS: 6.5(7) median(IQR)) were enrolled (Acute: 2.6+/-1.2 days). Nine of these subjects returned within 3 months (Sub-Acute: 49.4+/-13.0 days. NIHSS 3(4.5)). Stroke lesion masks (MRI) were derived from acute stroke images and characterized by their total size (44.5+/-35.5 cm3) and cortical involvement (% of voxels in cortex, 42.9+/-27.8%). HD-EEG (129 channels) was recorded during eyes-closed rest (185.2+/-82.8 seconds). The 1/f slope (3-40Hz) was calculated for electrodes on the lesioned and intact hemispheres and compared.
Results: Unilateral stroke lesions showed substantial variability in terms of size (44.5+/-35.5 cm3) and cortical involvement (42.9+/-27.8%). The aperiodic slope was significantly steeper on the lesioned side compared to the intact side (Acute: z=2.38, p=0.02, Sub-acute: z=2.43, p=0.02) with no difference in this asymmetry found between the 2 timepoints (z=0.06, p=0.95). We found no significant relationships between asymmetry and lesion size (Acute: rho=-0.11, p=0.67, Sub-Acute: rho=0.28, p=0.46), % of lesion that affected the cortex (Acute rho=0.20, p=0.45, Sub-Acute: rho=0.42, p=0.27) or change in NIHSS (rho=0.14, p=0.73).
Discussion: Despite large heterogeneity in stroke lesion size and location, we found persistent increases in asymmetry of aperiodic neural activity in the first 3 months after stroke, indicating a persistent decrease in Excitation/Inhibition on the lesioned compared to intact hemisphere. Local EEG activity is a combination of many factors, including input received from intra-and inter-cortical regions, subcortical regions (e.g. thalamus, basal ganglia) and excitability changes caused by ascending neuromodulatory pathways (e.g. acetylcholine). Understanding how different aspects of the EEG signal relate to underlying pathophysiology, will increase the utility of EEG to monitor recovery and guide personalized neurorehabilitation.
Categories
Stroke
Predicting a functional rehabilitation outcome in chronic stroke survivors via a hierarchical Bayesian model of motor learning
Nicolas Schweighofer1, Dongze Ye2, David, Z. D’Argenio3, Carolee Winstein1
1Biokinesiology and Physical Therapy, University of Southern California, Los Angeles, USA. 2Computer Science, University of Southern California, Los Angeles, USA. 3Biomedical Engineering, University of Southern California, Los Angeles, USA
Abstract
Recent modeling work has sought to predict the long-term “spontaneous” recovery post-stroke. Whereas such predictions are useful for clinical and research stratification, what the neurorehabilitation clinician needs is to be able to determine the best course of motor therapy for each stroke survivor, i.e., to personalize based on predictions of long-term outcomes. Functional improvements in the chronic phase post-stroke are thought to be due to the learning of compensatory movements, in which the individual uses different joints and muscles from those originally used before the injury to achieve a movement goal.
Here, we therefore propose to predict long-term changes in a functional motor outcome that results from rehabilitation training in the chronic phase post-stroke with a dynamical (state-space) model of motor learning. Our model incorporates the effects of external training, self-training, and forgetting (Hidaka, Han et al. 2012) (Wang, Winstein et al. 2020). To improve predictions early in treatment, when training data is sparse or even non-available, we incorporate co-variates in the model and use a Bayesian hierarchical structure, which allows for the incorporation of prior information from similar patients: the prior parameter distributions for each individual are derived from a higher-level population analysis that captures similarities between individuals.
We use this dynamical model to re-analyze Motor Activity Log (MAL) data from the DOSE trial (Winstein, Kim et al. 2019), in which participants were assigned to 0, 15, 30, and 60-hour dose groups. The MAL, which measures the participant’s perception of their amount and quality of movements by asking them to recall and rate the quality of movement of the paretic arm for 28 activities of daily living (Uswatte, Taub et al. 2006), was administered in 14 longitudinal assessments before and after practice in each of the three 1-week training bouts and then monthly for 6 months following the last training bout. The model well accounts for individual trajectory in the MAL during and outside of training. The model better fits the data than other simpler models without effects of either external training, self-training, or forgetting. We then show how the model can be used to predict long-term outcomes of new participants – we incorporate baseline and initial training data for a new participant and predict the changes in MAL up to 8 months ahead. We verify the accuracy of these predictions based on actual data for these “new” participants and show how additional data improves the mean and the precision of the predictions. In future work, such models can be simulated for different doses and schedules to optimize rehabilitation for each stroke survivor.
Categories
Motor Rehabilitation
Intraspinal microstimulation intended for motor rehabilitation modulates spinal nociceptive neural transmission
Maria Bandres, Jefferson Gomes, Jacob McPherson
Washington University in St. Louis, St. Louis, USA
Abstract
Spinal cord injury (SCI) leads to radical and detrimental physiological changes at and below the level of the lesion. For example, SCI often results in dysregulation of spinal reflexes, impaired bowel and bladder function, respiratory distress, reduction of voluntary motor output, and neuropathic pain. Spinal stimulation-based therapies are a promising approach to treating each of these issues. However, spinal stimulation can also cause off-target effects.
Spinal stimulation for motor rehabilitation relies on the recruitment of sensory afferent pathways and interneuron networks that provide excitatory synaptic drive to spinal motor pools. Some of these pathways are also implicated in the development and persistence of SCI-related neuropathic pain. Yet, spinal stimulation-based therapies intended to restore motor function after SCI often fail to consider these potential off-target effects.
Here, we used spike train analyses to characterize the effects of intraspinal microstimulation (ISMS) intended to enhance voluntary motor output on spinal nociceptive neural transmission. All experiments were approved by the Institutional Animal Care and Usage Committees of Florida International University and Washington University in St. Louis and were conducted in adult male Sprague-Dawley rats under urethane anesthesia (n=14). After T13-L2 laminectomy, microelectrode arrays were implanted at the L5 dorsal root entry zone. Electrode locations for spinal stimulation targeted motor pools of the ventral horn, whereas their locations for quantifying neural transmission in sensory pathways spanned the superficial and deep dorsal horns. Prior to, during, and after ventral ISMS (vISMS), we mechanically stimulated the L5 dermatome by applying controlled forces ranging from non-painful to painful. Our primary outcome measure was change in maximum firing rate [Hz] of nociceptive specific (NS) and wide dynamic range (WDR) neurons during nociceptive transmission. These changes were quantified by comparing the average maximum instantaneous firing rates prior to, during, and after 2, 10, or 30min of vISMS.
We found that even short periods of vISMS (≤10min) modulated neural transmission in nociceptive pathways of the dorsal horn. Indeed, ~58% of NS (n=94) and ~60% of WDR (n=349) neurons reduced their maximum firing rate during induced nociceptive transmission after vISMS compared to before vISMS. And surprisingly, in neurons that were depressed by vISMS, the magnitude of depression continued to increase after vISMS was discontinued. For example, NS neurons exhibited a 40.86% decrease in firing rate after vISMS (relative to baseline) compared to a 26.76% decrease during vISMS.
Our results demonstrate that vISMS for motor rehabilitation simultaneously modulates transmission in spinal nociceptive pathways, with off-target effects persisting beyond the duration of vISMS. Although future work is required to elucidate the mechanisms underlying these neuromodulatory actions, our results suggest that it may be possible to optimize the stimulation paradigm to deliver multi-modal therapeutic benefits.
Categories
Spinal Cord Injury (SCI)
Examining the Relationships between Measures of Activity Behavior and Physical Health in Individuals with Chronic Stroke
Allison Miller, Zachary Collier, Darcy S. Reisman
University of Delaware, Newark, USA
Abstract
Steps per day (SPD) is the most commonly used metric to quantify real-world walking activity in stroke rehabilitation clinical trials. While SPD captures an individual’s overall stepping volume for a particular day, studies suggest that other domains of activity, such as sedentary time, the intensity of stepping activity, and the frequency (bouts) of activity may also be important to fully understand a person’s overall activity behavior. However, there is no consensus or gold-standard battery of metrics that exist that could be used to comprehensively measure activity behavior in stroke. Therefore, the purpose of this work was to examine the importance of specific activity behavior metrics that reflect the domains of Activity Volume, Activity Frequency, Sedentary Time, and Activity Intensity by analyzing their relationships with measures of physical health (body mass index (BMI), systolic blood pressure (SBP), Charlson Comorbidity Index) in individuals post stroke. We hypothesized that specific measures of Activity Volume (SPD), Activity Frequency (average number of walking bouts/day), and Sedentary Time (percent sedentary time and the fragmentation index) would be significantly related to measures of physical health in individuals with stroke. This was a cross-sectional analysis of 280 individuals with chronic (≥6 months) stroke. Twelve activity metrics reflecting each domain were examined: Activity Volume (SPD, average time walking/day, percentage of time walking/day), Activity Frequency (average number of long bouts ≥300 steps, average number of short bouts <40 steps, average number of walking bouts/day), Activity Intensity (peak 30, maximum 20-minute cadence, average bout cadence), Sedentary Time (percent sedentary time, number of sedentary bouts ≥30 minutes, fragmentation index). Three lasso regression models were used to distinguish the activity metrics most strongly associated with measures of physical health. The dependent variable in Models 1, 2 and 3 were SBP, BMI, and the Charlson Comorbidity Index, respectively. The optimal value of lambda was obtained using 10-fold cross- validation. Variables whose coefficients were >0 were interpreted as strongly related to the outcome.
The results for Model 1 are reported; however, the results of all models will be presented if accepted. In Model 1, the coefficients for peak 30 (β= -0.008), average bout cadence (β= -0.06), and number of sedentary bouts ≥30 minutes (β= 0.06) were >0, suggesting that these measures were most strongly related to SBP. The results of this work support physical activity guidelines and suggest that the intensity of real-world walking activity and sedentary time are strongly related to physical health outcomes. This also suggests that rehabilitation professionals should consider measuring the intensity of real-world activity and sedentary time in addition to activity volume. Completion of this work will result in a battery of metrics that could used in subsequent studies in stroke targeting activity.
Categories
Stroke
A system to test forepaw touch in rodents shows stability in health and loss of function with injury
Derrick Yoo1, Aditya Ramamurthy1, Justin Lee1, Tong Chun Wen1, Andrew Sloan2, Jason Carmel1
1Columbia University, New York, USA. 2Vulintus Inc, Lafayette, USA
Abstract
Somatosensory function is routinely measured in humans by determining the threshold for detection of a vibrotactile stimulus applied to the hand, and this measure correlates with dexterity. There are currently no turnkey systems to assess touch in rats despite the need for touch in dexterity since behaviors that signal stimulus detection were challenging to train. We thus developed SensiTrak, to deliver graded tactile stimuli (vibration or texture) and measure rat responses by a rat’s release of a handle (go/no go) or a nose poke to one side or the other (2 alternative forced choice). Rats were trained to reach through an aperture in a clear plastic box to grab a handle attached to a force sensor and a vibrator. The force sensor was used to determine when a rat contacted the handle to trigger vibration 1 to 9 seconds later. Rats were trained to release the handle when they feel the vibration by releasing the handle, and a timed release results in a food reward. The ability to correctly release the handle across a range of stimulus durations determines their sensory ability. We tested the time to train rats, the variability between and within rats, and change in performance with two injury models. These models include a cut injury to the dorsal column of the spinal cord which severs a key sensory pathway, and a local anesthetic applied at the wrist. On average, training rats on SensiTrak took 27 ± 2.25 days, down from a previous training regimen of 36.3 ± 2.3 days. To determine the sensory ability, we plotted correct responses against the duration of the vibration stimulus. We fit a sigmoidal shaped response curve to determine the stimulus duration needed to perform 50% successful trials. Rats achieved a 50% success rate at a stimulus duration of 21.8 ± 2.76ms, (n=9). Stability of this measurement was determined between trials of a single subject and between subjects using the intraclass correlation method. We used these methods to quantify the loss of sensory function with two types of perturbations. In our first model, a dorsal column lesion was performed. The performance on the vibration task changed from a baseline of 29.1ms, to 71.4ms 3 weeks after injury, to 28.95ms ~5 weeks after injury. In our second model, we administered bupivacaine injections in the wrist and observed a shift in baseline from 17.2 ± 0.91ms to 42.8 ± 8.21ms during inactivation, fully washing out after 12 hours to 20.6 ± 5.01ms (n=3). SensiTrak, thus, is a reliable measure of somatosensory thresholds in rats in health and after injury.
Categories
Sensory Rehabilitation
Functional Implications of Lower Extremity Transcortical Reflex Responses Post-Stroke
Caitlin Banks1,2,3,4, Elliott Perry1,2,4, Wandasun Sihanath1,2, Theresa McGuirk1,2,4, Carolynn Patten1,2,3,4
1Biomechanics, Rehabilitation, and Integrative Neuroscience Lab, Department of Physical Medicine & Rehabilitation, UC Davis Health, Sacramento, CA, USA. 2UC Davis Center for Neuroengineering & Medicine, Davis, CA, USA. 3UC Davis Biomedical Engineering Graduate Group, Davis, CA, USA. 4VA Northern California Health Care System, Martinez, CA, USA
Abstract
The long-latency reflex (LLR) in the distal leg occurs 80-120ms following muscle stretch and is known to involve cortical structures [1]. Our prior work revealed absent LLRs in some individuals with chronic stroke and LLRs served as an unambiguous distinction between higher- and lower-functioning hemiparetic individuals [2]. We proposed that LLR presence or absence may therefore serve as a biomarker for lower extremity function and recovery in individuals with stroke. To establish this as a biomarker, it is important to confirm that LLRs are physiologically relevant to the pathology of stroke (i.e., if the stretch response measured previously involves cortical structures). The standard experiment involves timing transcranial magnetic stimulation (TMS) to align with the reflex response. If the LLR is facilitated by TMS, one can reasonably conclude it involves a transcortical component [3]. Here, we sought to confirm the cortical contribution of LLRs in individuals with chronic stroke, and to validate the previously observed relationship between LLR presence and clinical assessments of motor function.
Six individuals with unilateral, hemispheric stroke (age: 49±15 years, chronicity: 6.3±3.4 years) and six healthy individuals (age: 57±16 years) were studied. LLRs were recorded from the paretic tibialis anterior while participants performed dynamic plantarflexion contractions. TMS was timed to coincide with LLR arrival and in increments of 10ms before and after LLR onset. An individual was classified LLR present (LLR+) if there were measurable LLRs in at least 5/10 trials, or LLR absent (LLR-) otherwise.
All healthy individuals revealed LLRs as expected, while three individuals with stroke were LLR+ and three were LLR-. Stretch responses varied by TMS Timing (p<0.001). Additionally, there was a significant effect of LLR Status (i.e., control, LLR+, LLR-; p=0.002) and a significant Timing by Status interaction (p=0.002). The greatest facilitation occurred when TMS was delivered 10 and 20ms after LLR onset.
Lower extremity FMA score was greater in LLR+ than LLR- (p=0.038). SPPB score, DGI, and self-selected overground walking speed were all significantly different by LLR Status (p=0.005, p=0.019, and p<0.001, respectively).
These results confirm the reflex recorded using this paradigm involves a transcortical component. Consistent with our prior work [2], we observed absence of dorsiflexor transcortical reflex responses in some individuals with chronic stroke. LLR- individuals revealed more severe functional impairment, while LLR+ individuals were not different from controls in any aspect measured. These data show promise for LLR status as a biomarker of a physiologic process that can differentiate functional levels of motor impairment which, unlike many clinical measures, is objective and unambiguous. It is yet to be determined whether LLRs are a biomarker of stroke recovery. Further research will investigate the mutability of LLRs and how changes in LLRs contribute to improvements in motor function.
Categories
Stroke
Characterizing upper extremity movement smoothness in patients with acute stroke
Sarah Cavanagh1,2, Taya Hamilton2, Nicole Dusang1,2, Perman Gochyyev2, Julie DiCarlo1,2, Sydney McKiernan2, Hannah Jacobs2, Rashida Nayeem3, Steven Kautz4,5, Dagmar Sternad3, Leigh Hochberg1,2, David Lin1,2
1VA Medical Center, Providence, USA. 2Massachusetts General Hospital, Boston, USA. 3Northeastern University, Boston, USA. 4Medical University of South Carolina, Charleston, USA. 5VA Medical Center, Charleston, USA
Abstract
Robot kinematics provide a standardized, objective, and high-resolution assessment of an individual’s motor control and movement quality. After stroke, goal-directed upper limb movements are often slow and exhibit spatial and temporal discontinuities (i.e., abnormalities in movement smoothness). Few studies have characterized smoothness in upper limb movement in the acute stage of stroke when individuals have the most motor impairment and have yet to receive rehabilitative therapy. Here we aimed to characterize upper extremity movement smoothness and examine its relationship to motor impairment in a population of acute stroke patients compared to able-bodied controls.
Eighteen acute stroke survivors with upper extremity impairment (4.5 ± 1.9 days poststroke, Fugl-Meyer range [7-65]) and fifteen able-bodied control subjects completed 80 trials of a planar reaching center-out task (8-directions) on the Bionik InMotion2 Arm Therapy System, a clinical rehabilitation robotic system. For able-bodied controls, the dominant arm was assessed, and for stroke patients, both “affected” and “unaffected” arms were assessed. Dimensionless Squared Jerk (DSJ), the time-derivative of acceleration without dependency on amplitude or duration, was calculated on single trials to quantify the movement smoothness of planar reaching. One-way ANOVA was used to analyze mean differences of DSJ between able-bodied control, acute stroke “affected,” and acute stroke “unaffected” arm reaching movements.
For stroke reaching trials, we calculated the percent of trials that fell within the two standard deviations of the movement smoothness distribution of our control population. Finally, in stroke patients, we investigated the relationship between DSJ (average, variability, difference from able-bodied) of reaching movements and upper-extremity Fugl-Meyer score.
There were statistically significant differences between the means of able-bodied control dominant arm, acute stroke “unaffected” arm, and acute stroke “affected” arm DSJ during planar reaching (F(2)=559.2, p<0.001). We found significant correlations between upper extremity Fugl-Meyer scores and mean DSJ (r=-0.516, p=0.017), standard deviation DSJ (r=-0.525, p=0.015), and percent of trials with DSJ in the able-bodied range (r=0.435, p = 0.049). Surprisingly, even for patients with moderate-severe motor impairment (Fugl-Meyer less than 44), over 50% of “affected” arm trials demonstrated DSJ within the able-bodied range.
In this preliminary analysis, individuals with greater upper extremity motor impairment in the acute stage of stroke have reduced and more variable movement smoothness in both the “affected” and “unaffected” upper extremity. Nevertheless, there was also a considerable number of “affected” arm trials, even in patients with moderate-severe motor impairment, that fell within the range demonstrated by able-bodied control participants. Future studies will examine synergy patterns, cortical and subcortical injury, and real-time neurophysiology associated with these results. These findings may have implications for targeted neurorehabilitation for individuals after acute stroke.
Categories
Stroke
Pre-training neural correlates for predicting gains from robot-assisted finger training after stroke
Sebastian Rueda Parra1, Joel C. Perry2, Eric T. Wolbrecht3, David Reinkensmeyer4,5, Disha Gupta6,7
1Electrical Engineering, University of Idaho, Moscow, ID, USA. 2Mechanical Engineering, University of Idaho, Moscow,ID, USA. 3Mechanical Engineering, University of Idaho, Moscow, ID, USA. 4Biomedical Engineering, University of California, Irvine, CA, USA. 5Anatomy and Neurobiology, University of California, Irvine, CA, USA. 6National Center for Adaptive Neurotechnology, Stratton Veterans Affairs Medical Center, Albany, NY, USA. 7Electrical and Computer Engineering, University of Albany, State University of NY, Albany, NY
Abstract
Objective: The prevalence of stroke is one of the most important public health problems in the US and worldwide. In the chronic stage, most post-stroke survivors still present impairment, specially of upper extremities affecting performance in activities of daily living (thus impacting their independence).
Training paradigms aimed at reestablishing neural pathways usually involve repeated arm movements to stimulate neuroplastic changes. Robotic devices are being actively explored for movement rehabilitation, but results have been variable, highlighting the need for personalized robotic training. This may be possible with appropriate predictors of robotic therapy gains. Metrics derived from electroencephalogram (EEG), such as delta alpha ratio (DAR), and Power Patio Index (PRI), that have been useful in predicting gains in standard movement therapy [1], are yet to be fully explored for robotic-assisted movement therapy (RAMT).
Method: We analyzed 256-channel scalp EEG resting state data collected pre- and post-therapy from 27 chronic stroke survivors who underwent RAMT (3 sessions of one hour a week for 3 weeks, completing around 8000 repetitions) of the index and middle fingers with the Finger Individuating Grasp Exercise Robot [2]. The aim was to assess whether pre-training EEG features were useful in predicting RAMT gains for 1) the distal joint that received training (i.e. the fingers), as measured by standard assessments of dexterity (Box and Blocks, B&B); and 2) the proximal joints (such as elbow/shoulder), as measured by Flugl-Meyer Assessment, FMA-proximal score.
Results: We found that interhemispheric DAR (i.e. affected vs unaffected hemisphere), was significantly correlated with B&B (rho= 0.59, p< 0.05), but not with the proximal FMA (rho=0.16, p> 0.05). The PRI, previously shown to be related to motor gains in a robotic paradigm for elbow [1], did not show a significant correlation with distal (B&B, rho= 0.23, p> 0.05), or proximal (FMA proximal, rho= -0.026, p> 0.05) function gain. Global alpha power was significantly correlated with proximal FMA (rho= 0.47, p< 0.05) and not with B&B (rho= 0.19, p> 0.05).
Conclusion: Our results suggest that EEG power features can be useful for predicting RAMT gains in chronic stroke survivors. We also highlight how a distal (finger based) RAMT may have an impact on upper-extremity movement, like standard rehabilitation protocols.
Categories
Stroke
Ecological momentary assessment of post-amputation pain as an accurate and complementary alternative to traditional pain assessment
Kelli Buchanan, Binal Motawar, Scott Frey
University of Missouri, Columbia, USA
Abstract
Introduction: Post-amputation pain (PAP) is a chronic, difficult-to-treat condition comprised of both phantom (PLP) and residual limb (RLP) components. Due to the fluctuating severity of PAP, we hypothesize that data from standard, retrospective, self-report PAP measures are distorted by memory biases. We test this possibility by comparing estimates gathered with traditional measures and Ecological Momentary Assessment (EMA), a methodology that uses repeated sampling of pain severity in real-time.
Methods: 45 individuals with amputation(s) completed 7 days of EMA using a mobile phone application. Participants were asked to report current PLP and RLP on separate visual analog scales (VAS) from 0 (“no pain”) to 100 (“worst possible pain”) 6 times daily (i.e., notification-initiated surveys delivered once at a randomized time within each two-hour window from 8am-8pm). Additionally, participants were encouraged to enter intense or unusual pain events immediately (i.e., user-initiated surveys). Prior to starting the week-long EMA protocol, participants reported average PLP and RLP levels on an electronic retrospective pain measure with identical VASs.
On average, individuals responded to 62% notification-initiated surveys (SD = 23%, range = 12% - 98%). The compliance rate for daily responders (n=30) was 73%, SD = 15%, range = 31% - 98%. The minimum number of randomized surveys taken in this group was 13, M = 30, SD = 6. We utilize the subgroup of daily responders for all analyses.
Results: Pre-test estimates of average PLP on traditional, cross-sectional VASs were significantly higher than levels reported by EMA in notification-initiated surveys (p < .001), user-initiated surveys (p < .001), and these scores combined (p < .001). Similarly, pre-test RLP pain averages were significantly higher than levels reported by EMA in notification-initiated surveys (p = .002), user-initiated surveys (p = .049) and these scores combined (p = .003).
Conclusion: Retrospective estimates of average PLP and RLP severity are prone to overestimation, likely due to memory biases. EMA measures may provide more valid and reliable estimates of PAP, as well as a more nuanced representation of the patient experience. This technique may better inform interventions that aim to increase the quality of life of individuals who live with amputation.
Categories
Other
A gamified electromyographic computer interface to measure specific motor control abnormalities in healthy controls and individuals with arm impairment due to stroke
Danielle Marouni1, Yiyun Wang1, David Cunningham2, Ania Busza3
1Univeristy of Rochester, Rochester, USA. 2Case Western, Cleveland, USA. 3University of Rochester, Rochester, USA
Abstract
Background: Upper extremity (UE) disability is common after stroke. Prior studies in individuals with hemiparesis due to stroke have identified four distinct impairments of motor control: (1) decreased maximal muscle activation, (2) delayed muscle activation, (3) motor fatigue, and (4) abnormal co- activation of antagonistic muscle groups. In order to develop impairment-oriented interventions for motor recovery, more information is needed about the emergence and prevalence of these impairments. We have developed an electromyographic (EMG) computer interface to collect EMG signals from subjects performing repetitive muscle activations, with the goal of characterizing the evolution of each impairment over the first 3 months post-stroke.
Objective: To use our previously-developed EMG computer interface to collect EMG data from able- bodied participants and participants with UE hemiparesis after stroke in order to measure the intra- subject variability in able-bodied controls across multiple time points, to assess whether the system can be used to compare relative motor impairments in a longitudinal study of individuals with recent stroke.
Design/Methods: Our EMG computer interface uses surface EMG signals from the wrist flexor and extensor muscle groups to control a simple computer game. Successful gameplay requires multiple isometric muscle contractions at precise time points, each lasting 3 seconds. EMG data are analyzed to identify maximum EMG values, muscle activation delay, and co-activation of antagonist muscle groups. EMG data from able-bodied controls are recorded multiple times, separated by several days, to measure intra-subject variability. In parallel, the system is being used to collect EMG data from subjects with hemiparesis due to recent (<4 weeks) stroke.
Results: Our system is well-tolerated and collects relatively large amounts of data from able-bodied controls and participants with acute stroke. Preliminary data collected from participants with hemiparesis due to recent (<4 weeks) stroke show increased delay in activation of wrist flexor and extensors in their affected UE as compared to their unaffected arm. Some participants show increased co-activation and increased motor fatigue in their paretic arm, especially when the required activation threshold is increased.
Conclusions: Our EMG-controlled computer interface is well tolerated and can be used to collect information about relative motor impairment levels in an individual across multiple time points. Our preliminary results are consistent with prior studies’ that show some individuals with hemiparesis due to stroke have high levels of specific motor impairments such as delay in muscle activation and abnormal co-activation. Further studies will assess the prevalence and individual trajectory of each motor impairment over the first 3 months post-stroke.
Categories
Stroke
Identifying Cognitive Predictors to Reactive Step Training in People with Parkinson’s Disease
Andrew Monaghan1, Jessica Trevino1, Jordan Barajas1, Lee Dibble2, Shyamal Mehta3, Daniel Peterson1,4
1Arizona State University, Phoenix, USA. 2University of Utah, Salt Lake City, USA. 3Mayo Clinic, Scottsdale, USA. 4Phoenix VA Health Care Center, Phoenix, USA
Abstract
Introduction: Falls often occur due to an unexpected challenge to balance, requiring rapid postural responses such as reactive stepping. Significantly, reactive stepping is impaired in Parkinson’s Disease (PD) and is associated with falls. Despite this, reactive stepping is adaptable through training. However, the responsiveness to balance training in people with PD (PwPD) is highly variable. Evidence suggests that a person’s cognitive function, particularly visuospatial, correlates with motor learning. Given the relationship between cognition and learning, and the cognitive deficits observed in PD, this study aimed to determine whether global cognition or visuospatial function predicts responsiveness to reactive step training in PwPD.
Method: 24 PwPD were recruited to an 18-week multiple baseline study. Participants attended 2 baseline assessments (B1 and B2) before training, a 2-week, 6-session step training protocol, and 2 post-training assessments (P1 and P2) to assess the acute (B2-P1) and retained (B2-P2; 8 weeks post-training) effects of training. Assessment visits consisted of 3 backward and forward reactive step trials at constant participant-specific perturbation intensity. Training consisted of 32 (8 each in forward, backward, left, rightward directions) stepping trials per day with increasing perturbation intensity throughout training. Steps were elicited via support-surface translations using an instrumented split-belt treadmill. 3D motion capture and force plate data were captured. The primary outcomes were the anterior-posterior margin of stability (MOS) and step length during backward stepping. Backward steps were prioritized due to the particular difficulty of backward loss of balance in PwPD. Measures of global cognition were measured at baseline via the Montreal Cognitive Assessment (MoCA) and the Scales for Outcomes in Parkinson’s Disease- Cognition (SCOPA-COG). Visuospatial functioning was assessed from the Block Design instrument of the WAIS-IV and the Judgement of Line Orientation (J-Lo) test. Multiple regression was used to determine if cognitive measures could predict training (B2 - P1; n = 24) and retention (B2 – P2; n = 20) effects.
Results: The MoCA significantly predicted training and retention effects in MOS, such that a 1-point increase in MoCA score reflected a 1.04 cm and 1.21 cm larger MOS following training and at the follow-up, respectively. The MoCA predicted improved (i.e., longer) step lengths following training (1.21 cm) and at the follow-up assessment (1.89 cm). The SCOPA-COG also significantly predicted training improvements in MOS but not the retention effects. Neither the J-LO nor the Block Design significantly predicted training or retention effects for MOS or step length.
Discussion: Global cognitive measures obtained from the MoCA and SCOPA-COG, rather than visuospatial functioning, predicted reactive step training responsiveness in PwPD. Therefore, measuring cognitive status through quick and inexpensive standardized cognitive assessments such as the MoCA and SCOPA- COG may be useful for predicting responsiveness to training in people with PD.
Categories
Motor Rehabilitation
Relating Reactive Step Length and Step Latency to Falls in People with Multiple Sclerosis
Andrew Monaghan1, Avril Mansfield2,3,4, Jessie Huisinga5, Daniel Peterson1,6
1Arizona State University, Phoenix, USA. 2KITE- Toronto Rehabilitation Institute, Toronto, Canada. 3University of Toronto, Toronto, Canada. 4Sunnybrook Research Institute, Toronto, Canada. 5University of Kansas Medical Center, Kansas City, USA. 6Phoenix VA Health Care Center, Phoenix, USA
Abstract
Introduction: Although reactive stepping is necessary to prevent falls, which aspects of stepping relate to falls remains poorly understood. Identifying which outcomes are related to falls in people with multiple sclerosis (PwMS) is an important first step towards improving fall-prevention rehabilitation. For example, compared to non-MS age-matched controls, reactive backward steps are later (but not necessarily smaller) in PwMS. The purpose of the current study is to determine whether reactive step latency or length during forward and backward losses of balance relates to fall history in PwMS.
Methods: 111 people with relapsing-remitting MS were recruited. 36 reported 1 or more falls, and 76 reported no falls in the previous 6 months. Participants completed treadmill-induced reactive steps (3 forward and 3 backward) from stance administered via support surface translations. 3D motion capture assessed step length (cm) and step latency (ms) and were analyzed via customized MATLAB software. Independent t- tests and Mann-Whitney U tests were used to determine whether step length or latency differed across PwMS with and without a fall history. Further, binary logistic regression with Firth’s penalized maximum likelihood estimation (PMLE) examined whether step latency or step length predicted fall history.
Results: PwMS with a history of falls exhibited significantly slower step latencies during backward (p=0.002) but not forward stepping than those without a history of falls. Slower backward step latencies significantly increased the odds of having experienced a fall. For a 10 ms increment in step latency, the probability of having a fall history increases by a factor of approximately 2.479 (OR= 2.479, 95% Wald CI: 1.125 – 5.464). Step length was not significantly different between groups and did not predict fall history.
Discussion: PwMS and a history of falling exhibited later step onsets during backward reactive stepping. This work suggests that latency of steps may be a relevant target for fall prevention rehabilitation; however, additional fall-prevention clinical trials are necessary to 1) determine the adaptability of reactive step length and 2) whether changes reduce the incidence of falls.
Categories
Motor Rehabilitation
Taking the Assessment of Freezing of Gait from the Lab into the Clinic and the Real World
David May, Gammon Earhart, Pietro Mazzoni
Washington University in St. Louis, St. Louis, MO, USA
Abstract
Introduction: Freezing of gait (FOG) is a debilitating symptom of Parkinson’s disease (PD) that remains difficult to assess due to its variability. Currently, clinicians primarily rely on the New Freezing of Gait Questionnaire (NFOG-Q) to assess FOG. The NFOG-Q is a useful screening tool but is of limited utility for accurately and reliably assessing FOG. Recent advances in wearable movement sensors make it possible to study gait outside the lab. We aimed to enhance ability to measure FOG objectively by using wearable sensors to detect FOG episodes in two settings: 1) a highly structured clinic-like lab environment, and 2) the real-world. Should these methods prove reliable, they could considerably improve our ability to measure severity of FOG and assess the true impacts of interventions on FOG.
Methods: We recorded foot and hip movements with three commercially available wearable movement sensors in 14 people with PD as they performed structured gait tasks in a clinic-like lab environment and during three subsequent days at home. The lab-based testing session included a set of structured gait tasks suitable for the clinic and a set of daily-life activities to be compared to data from home recordings. We calculated FOG severity through video review of the lab session by human experts and through analysis of sensor data from the lab and the home using a detection algorithm recently developed in collaboration with engineering colleagues. We compared video vs. sensor-based measures for the lab- based tasks and compared sensor data from the lab- vs real-world. All 14 participants completed the NFOG-Q.
Results: Preliminary analysis of 5 participants’ data indicates good agreement between FOG detection via sensor data and via human expert review (r=0.70). Preliminary analysis also shows a correlation between sensor-based FOG severity in the real-world and in the lab (r=0.73). Preliminary analysis of NFOG-Q results, however, shows a negative correlation with human expert review (r= -0.29) and with sensor-based FOG severity in the real-world (r= -0.82) and lab (r= -0.55). The lab protocol proved feasible for patients with a wide range of severity of freezing and gait difficulty. Data analysis is ongoing and the full results from the 14 participants will be presented at the meeting.
Conclusions: Thanks to technological advances, it is now realistic to add wearable sensors and signal processing techniques to the repertoire of motor symptom assessment strategies, as the approach we used can potentially be deployed in clinical practice. The negative correlation between the NFOG-Q and observed measures of FOG severity further demonstrates the need for these new methods. Objective, real-world assessments of FOG will allow clinicians and researchers more accurately to determine the severity of this often-elusive symptom and to measure the effects of potential treatment strategies.
Categories
Other
Targeted Plasticity Therapy for Upper Limb Rehabilitation in Spinal Cord Injury
Emmanuel Adehunoluwa1,2, Joe Epperson1,3, Chad Swank4, Christie Stevens4, Dannae Arnold4, Jaime Gillespie4, Erina Sarker4, Jane Wigginton1, Michael Foreman4, Richard Naftalis4, Rita Hamilton4, Amy Porter1, Robert Rennaker1,2, Seth Hays1,3, Michael Kilgard1,2
1Texas Biomedical Device Center, University of Texas at Dallas, Richardson, USA. 2School of Behavioral and Brain Sciences, University of Texas at Dallas, Richardson, USA. 3Erik Jonsson School of Engineering and Computer Science, University of Texas at Dallas, Richardson, USA. 4Baylor Scott & White Institute for Rehabilitation, Dallas, USA
Abstract
Spinal cord injury (SCI) has widespread consequences, including partial or total loss of upper limb function. The goal of rehabilitation is to help the affected individuals optimize their function and improve their quality of life. Recent animal studies showed that pairing vagus nerve stimulation (VNS) with physical rehabilitation improves upper limb function after spinal cord injury. Additionally, evidence from clinical studies shows that VNS paired with rehabilitation improves recovery in individuals with chronic stroke. Building on these findings, we performed a double-blinded randomized controlled early feasibility study to investigate the safety, feasibility, and potential benefit of VNS paired with rehabilitation after chronic incomplete spinal cord injury.
Participants in this study are individuals between 18-64 years with chronic motor deficits resulting from an incomplete traumatic spinal cord injury that occurred at least 12 months before enrollment. All the participants were implanted with a novel VNS device on the left cervical vagus nerve. They were randomly assigned to either receive active VNS paired with rehabilitation or sham stimulation with rehabilitation for 6 weeks of in-clinic therapy. Thereafter, participants could elect to receive an additional 6 weeks of in-clinic therapy with active VNS, regardless of previous group assignment. Physical rehabilitation sessions occurred three times per week for 90 minutes each and consisted of standard exercises and game-based training. Assessments (including the Graded Redefined Assessment of Strength, Sensibility, and Prehension) were performed before therapy, after the conclusion of the first six weeks of therapy, and after the conclusion of the second six weeks of therapy.
To date, no serious or unexpected device-related adverse events have been observed. Preliminary analysis of functional outcomes indicates that GRASSP scores are increased in individuals that receive active VNS paired with rehabilitation. These initial findings suggest that VNS paired with physical rehabilitation is safe and has potential benefits in individuals with incomplete spinal cord injury.
Categories
Motor Rehabilitation
Bilateral upper extremity motor priming (BUMP) plus task specific training for severe, chronic upper limb hemiparesis: Study protocol for a randomized clinical trial
Mary Ellen Stoykov1,4, Olivia M. Biller2, Alexandra Wax4, Erin King1, Jacob M. Schauer1, Louis F. Fogg5, Daniel M. Corcos3
1Northwestern University, Chicago, USA. 2Patient Reported Outcomes, Global Market Access, Johnson & Johnson Innovative Medicine, Global Commercial Strategy Organization (GCSO). 3Northwestern University, Chicago. 4Shirley Ryan Ability Lab, Chicago, USA. 5University of Illinois at Chicago
Abstract
Background. Various priming techniques to enhance neuroplasticity have been examined in stroke rehabilitation research. Most priming techniques are costly and approved only for research. Here, we describe a priming technique that is cost-effective and has potential to significantly change clinical practice. Bilateral motor priming uses the Exsurgo priming device (Exsurgo Rehabilitation, Auckland, NZ) so that the less affected limb drives the more affected limb in bilateral symmetrical wrist flexion and extension. The aim of this study is to determine the effects of a five-week protocol of bilateral motor priming in combination with task specific training on motor impairment of the affected limb, bimanual motor function, and interhemispheric inhibition in moderate to severely impaired people with stroke.
Methods. Seventy-six participants will be randomized to receive either 15, 2-hour sessions, 3 times per week for 5 weeks (30 hours of intervention) of bilateral motor priming and task specific training (experimental group) or the same dose of control priming plus the task specific training protocol. The experimental group performs bilateral symmetrical arm movements via the Exsurgo priming device which allows both wrists to move in rhythmic, symmetrical wrist flexion and extension for 15 minutes.
The goal is one cycle (wrist flexion and wrist extension) per second. The control priming group receives transcutaneous electrical stimulation below sensory threshold for 15 minutes prior to the same 45- minutes of task specific training. Outcome measures are collected at pre- intervention, post-intervention, and follow-up (8 weeks post-intervention). The primary outcome measure is the Fugl-Meyer Test of Upper Extremity Function. The secondary outcome is the Chedoke Arm and Hand Activity Index-Nine, an assessment of bimanual functional tasks.
Discussion. To date, there are only 6 studies documenting the efficacy of priming using bilateral movements, 4 of which are pilot or feasibility studies. This is the first large-scale clinical trial of bilateral priming plus task specific training. We have previously completed a feasibility intervention study of bilateral motor priming plus task specific training and have considerable experience using this protocol.
Categories
Stroke
Comorbid anxiety disorder as the strongest predictor of post-stroke depression
Amber Criswell, BA, Timea Hodics, MD, Camila Quintero, BS, Mario Dulay, PhD
Houston Methodist Neurological Institute, Houston, USA
Abstract
Background: Post-stroke depression (PSD) is common in a subset of individuals after cerebrovascular accident. Studies have found that anywhere between 20-60% of stroke patients develop PSD, most commonly within the first year of injury (Towfighi, et al., 2017). PSD may occur due to environmental factors such as functional limitations in daily activities and lower quality of life or biological factors such as damage to areas in the brain involved in emotion regulation, including the prefrontal cortex or temporal lobe. Although many factors have been proposed to be associated with increased risk of PSD, the relative contribution of these factors is not well understood. We evaluated which cross-sectional variables are the strongest predictors of PSD in our outpatient neuropsychology clinic population.
Methods: Three-hundred and thirty-five patients (49.3% female; mean age of 59.3 years-old) who sustained an ischemic or hemorrhagic stroke were included at an average of seven-and-a-half months after stroke in this prospective single center study. Stroke-related (side and location of stroke, time since stroke), demographic (gender, race, ethnicity, chronological age), psychosocial (employment status, marital status), comorbid medical (sleep difficulties, fatigue), psychiatric (history of treatment for depression prior to the stroke, current anxiety disorder), and neuropsychological (memory or executive difficulties) variables were recorded. The Mini International Neuropsychiatric Inventory (MINI) was used to determine DSM-IV-TR diagnosis of post-CVA major depressive disorder or an adjustment disorder with depressed mood (referred to as PSD) and anxiety disorders (mainly panic disorder). A standard neuropsychological test battery was administered. SPSS statistical software was used for all calculations.
Results: PSD occurred in 31% of the sample. Logistic regression using depression status as the dependent variable was significant (p < 0.001) revealing five (5) predictors of PSD. The strongest predictor of PSD was having a comorbid anxiety disorder, with these patients being 6.5 times more likely to suffer from PSD (p < 0.001). Other significant predictors of PSD included a history of depression before stroke (3.1 times more likely to suffer from PSD, p < 0.001), experiencing significant fatigue (2.9 times more likely to suffer from PSD, p = 0.002), having memory impairment (2.2 times more likely to suffer from PSD, p = 0.019), and older age at stroke (p = 0.006).
Discussion: Our findings were considered consistent with previous studies that demonstrated depression occurs at a high prevalence after stroke. Results extend the literature by demonstrating that the risk for PSD is multifactorial and includes demographic (age at stroke), comorbid medical (sleep difficulties, fatigue), psychiatric (history of treatment for depression prior to the stroke, current anxiety disorder), and neuropsychological variables (memory difficulties), however, comorbid anxiety disorder was the strongest predictor of PSD. Stroke specialty and rehabilitation centers should target patients with these neuropsychological comorbidities for intervention.
Categories
Stroke
Corticomuscular Coherence and Corticospinal Tract Injury Associations During Early Stroke Recovery
Rachana Gangwani, Jasper Mark, Rachel Vaughn, Jessica Cassidy
University of North Carolina at Chapel Hill, Chapel Hill, USA
Abstract
Introduction: Heterogeneity in stroke has propelled the development of neuroimaging-based biomarkers to enhance the prediction of functional outcomes and treatment response. Structural neuroimaging biomarkers of corticospinal tract (CST) injury are well-established in the literature as significant predictors of post-stroke motor recovery. However, biomarkers derived from functional neuroimaging require additional development. Corticomuscular coherence (CMC) is a functional measure representing synchrony between brain and muscle signals recorded during simultaneous electroencephalography (EEG) and electromyography (EMG) collection. Given the purported role of CST in mediating communication between brain and muscle as captured by CMC, this study determined the relationship between CMC and CST injury in early stroke recovery.
Methods: Participants with hemiparesis resulting from ischemic and hemorrhagic stroke hospitalized in an inpatient rehabilitation facility (IRF) completed motor testing (Upper Extremity Fugl-Meyer, UEFM; Action Research Arm Test, ARAT), a diffusion MRI, and CMC testing using a 256-lead EEG system at around the time of IRF admission. During CMC testing, muscle activity from bilateral biceps brachii (BB), flexor and extensor digitorum (FD, ED), and first dorsal interossei (FDI) was recorded while participants performed a submaximal isometric grip task using their affected hand with visual feedback provided.
CMC measurements involving leads overlying ipsilesional primary motor cortex (iM1) with the above muscles were computed across alpha-mu (8-12 Hz), low beta (13-19 Hz), and high beta (20-30 Hz) frequency bands. CST injury was determined by calculating a ratio of fractional anisotropy values from ipsi- to contralesional cerebral peduncles with higher ratios indicating less injury.
Results: Seven individuals (3 females, 68±7.5 years, 9.3±4.4 days post-stroke) with mild to moderate motor deficit (UEFM=54.3±8.7; ARAT=42.4±11.2) completed all study procedures. Reduced CST injury correlated with greater alpha-mu CMC between EEG leads overlying iM1 and EMG leads over affected BB (r=0.65, p=0.04) and between EEG leads overlying iM1 and EMG leads over affected FDI (r=0.64, p=0.04).
Conclusion: Preliminary findings of significant associations between alpha-mu CMC and CST injury measurements from this ongoing study support the premise of CST facilitating brain-muscle communication in a frequency band subserving both cognitive and sensorimotor aspects of movement during early stroke recovery. Determination of how these associations change over time and with motor recovery are important next steps.
Categories
Stroke
The fastest may not be the best: Analysis of the effects of gait speed on multiple biomechanical gait variables post-stroke
Michael Rosenberg1, Justin Liu1, Taniel Winner1,2, Gordon Berman1, Lena Ting1,2, Trisha Kesar1
1Emory University, Atlanta, USA. 2Georgia Institute of Technology, Atlanta, USA
Abstract
Background: Restoring walking speed and function following a stroke is a major rehabilitation goal. High-intensity treadmill training is an evidence-supported stroke gait intervention, with the perspective that faster training speeds yield better therapeutic outcomes [1]. However, studies underlying the faster- is-better perspective evaluated a limited number of speeds and biomechanical variables [2,3]. Here, we addressed the question: is faster always better (with respect to gait quality)? We hypothesized that changes in post-stroke gait biomechanics with speed are complex and multidimensional, such that personalizing gait training speeds requires a comprehensive understanding of how gait biomechanics vary with speed.
Methods: We collected data on 14 stroke survivors (4F; age: 59.8 ± 10.7 years; 41.1 ± 31.7 months post-stroke) walking on an instrumented split-belt treadmill at 6 evenly-spaced speeds ranging from their self-selected (SS) to fastest comfortable speeds. We recorded 3D marker trajectories and ground reaction forces during 15-second walking bouts at each speed. We computed the magnitudes of 9 discrete biomechanical gait variables of the paretic limb (e.g., peak propulsion; peak ankle moment) and the corresponding asymmetries between the paretic and non-paretic limbs. To determine if the fastest walking speed was best across these 18 variables, we identified the speed that optimized each variable (i.e., moved a variable toward target values, such as increasing propulsion or minimizing asymmetry).
Results: Across our 18 biomechanical variables, 78% of paretic-limb magnitude variables were optimized at the fastest two speeds, with large effects of the fastest speed compared to SS (Cohen’s d > 0.7). These variables included paretic-limb propulsion, trailing limb angle, step length, and peak hip and ankle power. Conversely, peak circumduction and ankle angle at heel-strike were optimized at the slowest two speeds (17% of variables). The detrimental effects of the fastest speed compared to SS were moderate (Cohen’s d < 0.3). While the fastest two speeds minimized biomechanical asymmetry in 35% of variables, 43% of asymmetry variables were optimized at the slowest speed. Unlike magnitude variables, which were generally optimized at 1-2 speeds across participants, the speeds at which asymmetry variables were optimized were individual-specific.
Discussion: By evaluating gait biomechanics at a range of speeds spanning participants’ walking capacity, we found that training at the fastest walking speed may not be best for optimizing important biomechanical variables, particularly during swing phase and with respect to gait symmetry. Because a major goal of stepping practice during rehabilitation is to reinforce improved walking patterns, our results indicate that 1) a range of speeds should be tested to find an individual-specific optimal balance of beneficial and detrimental changes in gait biomechanics with speed, and 2) holistic metrics that resolve disagreements between biomechanical variables are needed to identify individualized speeds that maximize biomechanical gait quality.
Categories
Stroke
Unraveling neuro-motor control deficits in healthy aging: Implications for neurorehabilitation
Daniele Piscitelli1, Rachael Walton-Mouw2, Stanislaw Solnik2,3
1School of Physical and Occupational Therapy, McGill University, Montreal, Canada. 2University of North Georgia, Dahlonega, USA. 3University of Health and Sport Sciences in Wrocław, Wrocław, Poland
Abstract
Introduction: People frequently interact with another person, and stability of such actions is crucial (patient-caregiver interactions, etc.). When a person executes a task alone, the central nervous system (CNS) exploits the variability of sensory-motor elements. However, when two persons share a task, their CNS needs to account for the extra-personal source of movement variability. The neurological effects of healthy aging may impair modulation of motor control processes to stabilize two-person motor tasks. We used the Uncontrolled Manifold (UCM) to compute the index of stability (DV), which quantifies how CNS utilizes movement variability to stabilize the performance of a motor task. Recently, the UCM was proposed as a sensitive and individualized biomarker of movement quality – a significant component in precision neurorehabilitation. This study explored neural motor control deficits in old adults while performing interpersonal prehension tasks with younger and age-matched counterparts.
Methods: Ten young (24.6±0.8yrs) and ten older adults (73.5±3.1yrs) were randomly assigned to 10 matched-age and mixed-age pairs for two-person conditions. Subjects sat on a chair holding the handle instrumented with force sensors vertically using one hand (in two-person conditions) or two hands (in one-person conditions). In one-person condition, each subject held the handle with both hands. Subjects held the handle with matched-aged or different age partners in two-person conditions. Subjects used only finger flexion and extension to cyclically transfer the handle from the left to right hand while keeping the handle’s orientation and position steady. We computed grip forces and UCM-parameters for the vertical and horizontal handle stability.
Results: We found no differences in clinical hand dynamometer measures between age groups (p=0.4). Old subjects had higher grip forces than young in all conditions (p=0.01). Only older adults increased their grip forces when paired with another older adult (p=0.05). Grip forces of young participants did not vary between conditions (p=0.6). For horizontal stability, young subjects decreased DV values when matched with another young person (p=0.03), compared to a task performed alone. Old adults showed no difference in DV between one-person and two-person condition with another older adult; however, they showed a decrease in DV when matched with a young person (p=0.04). For vertical stability, young and old participants behaved similarly, with one-person conditions having the largest DV values, than two- person conditions.
Conclusions: Results show changes in multi-digit coordination with age at the neural control level. Notably, young adults showed the ability to modulate the performance stability as an adaptation for tasks performed with another person of the same age. Old adults modulated stability of action only when sharing the task with younger counterparts. These findings suggest the potential role of motor control deficits in healthy aging. Future studies should clarify if targeted therapies could improve these deficits.
Categories
Motor Rehabilitation
Improved Post-Stroke Motor Recovery with Alternate Day Fasting in Mice
Mahlet Mersha, Robert Hubbard, Steven Zeiler
Johns Hopkins, Balitmore, USA
Abstract
Background. Caloric restriction promotes neuroplasticity and post-neurological injury recovery. In mice, we tested the hypothesis that caloric restriction can act post-stroke to enhance training-associated motor recovery.
Methods. Mice were trained to perform a skilled prehension task. We then induced a photothrombotic stroke in the caudal forelimb area, after which we retrained animals on the prehension task after an 8- day delay. Mice underwent either ad libitum feeding or alternate day fasting beginning 1-day after stroke and persisting for either 7 days or the entire post-stroke training period until sacrifice.
Results. Prior studies showed that post-stroke recovery of prehension can occur if animals receive rehabilitative training during an early sensitive period but is incomplete if rehabilitative training is delayed. In contrast, we show complete recovery of prehension, despite a delay in rehabilitative training, when mice underwent alternate day fasting beginning 1-day post-stroke and persisting for either 7 days or the entire post-stroke training period until sacrifice. Recovery was independent of weight loss. Stroke volumes were similar across groups.
Conclusions. Poststroke caloric restriction led to recovery of motor function without a protective effect on stroke volume. Since prehension recovery improved even after ad libitum feeding was reinstituted suggests that the observed motor recovery was not merely a motivational response. These data add to the growing field that post-stroke caloric restriction can enhance recovery.
Categories
Motor Rehabilitation
Age Related Differences in Kinematic Responses While Walking Over A Compliant Surface
Nesreen Alissa, Woohyoung Jeon, Ruth Akinlosotu, Kelly Westlake
University of Maryland, Baltimore, USA
Abstract
Background: Maintaining dynamic stability can be challenging when walking over compliant surfaces, especially in older adults where sensory input may be compromised. Anticipatory (APAs) and compensatory (CPAs) postural adjustments are important factors contributing to dynamic stability. These postural responses rely on sensory input, which can diminish with age.
Objective: To examine differences between young (YA) and older (OA) adults in limb and trunk kinematic responses during first and repeated trials of walking over a compliant surface.
Methods: YA (22 – 29yrs) and OA (68 – 84yrs) completed 4 trials of walking over a high-density foam mat. Participants were instructed to walk over the foam mat starting with their preferred leg and at their normal speed. Motion capture was used to record upper limb, lower limb, and trunk movements and angles at the time of first foot touch down (FFTD) on the foam. We determined anteroposterior (COMAP) and mediolateral center of mass (COMML) displacement during anticipatory and compensatory adjustments (500 ms before and after FFTD, respectively). Anteroposterior margin of stability (MOSAP) and mediolateral margin of stability (MOSML) were calculated at FFTD.
Results: YA had greater MOSAP (p<0.02), right shoulder flexion (p<0.018), and trunk flexion (p<0.05) at FFTD during the first trial compared to OA. OA had greater left hip flexion at FFTD during the fourth trial (p<0.019) and greater hip abduction at FFTD over all trials (p<0.019). OA also showed within-group increase in right elbow flexion (p<0.034) at FFTD from the second to third trial and decrease in right hip flexion (p<0.019) at FFTD from the third to the fourth trial.
Conclusions: There are age-related differences in dynamic stability strategies while walking over a compliant surface. OA showed habituation in right elbow flexion and right hip flexion angles over the four trials indicating the potential to improve dynamic stability with training.
Categories
Motor Rehabilitation
Different cortical oscillatory signatures during reactive balance are associated with distinct aspects of balance control post-stroke
Jasmine Mirdamadi1, Jacqueline Palmer1, Aiden Payne1, Lena Ting1,2, Michael Borich1
1Emory University, Atlanta, USA. 2Georgia Tech, Atlanta, USA
Abstract
Introduction: Increased cortical engagement for balance control post-stroke has been inferred through dual-task interference (DTI). More direct measures of neural activity during balance control may provide insight into the mechanisms underlying balance impairments post-stroke. In neurotypical adults, balance perturbations evoke increases in oscillatory brain activity at multiple frequencies over central sensorimotor areas. Beta (13-30 Hz) and gamma (30-50 Hz) oscillations represent signatures of sensorimotor integration. In contrast, theta (4-7 Hz) and alpha (8-12 Hz) oscillations are more related to cognition. Here, we tested whether different brain frequencies during reactive balance recovery are associated with distinct balance impairments post-stroke. We hypothesized that greater perturbation- evoked beta and gamma power would be related to lower balance ability, whereas greater perturbation- evoked theta and alpha power would be related to more DTI.
Methods: In 18 people with chronic stroke, electroencephalography (EEG) was measured during support surface perturbations to standing balance. Cortical activity from the central midline electrode (Cz) overlying central sensorimotor areas was analyzed in the time-frequency domain to quantify changes in oscillatory power in the theta, alpha, beta, and gamma bands pre- (-500-0 msec) and post-perturbation (150-500 msec). Clinical balance ability was assessed with the miniBEST. DTI was assessed using the Timed-Up-And-Go Test with and without a secondary cognitive task. We analyzed associations between the change in perturbation-evoked power (post - pre), balance ability, and DTI.
Results: Stroke survivors demonstrated a wide range of balance ability (miniBEST score range: 2-25) and DTI (range: -11% - 50%). MiniBEST performance was not associated with DTI (r = 0.20; p = 0.43). Worse balance ability was correlated with smaller evoked theta and alpha power during early balance recovery (150-300 msec; theta: r = 0.46, p = 0.055; alpha: r = 0.46, p = 0.056). In contrast, more DTI was correlated with greater evoked beta and gamma power throughout balance recovery (300-500 msec; beta: r = 0.66, p = 0.003; gamma: r = 0.69, p = 0.001).
Discussion: It was surprising that there was little to no DTI in stroke survivors with the worst balance. These distinct aspects of balance control were associated with different frequencies during reactive balance recovery. People with the worst balance had the lowest evoked theta and alpha power during the initial portion of balance recovery, suggesting potential deficits in sensory processing and/or appraisal. Contrasting with previous work in older adults, beta power was not associated with balance ability. Instead, individuals with higher beta power during balance recovery had more dual-task interference; this could be due to increased prefrontal-motor connectivity previously shown in older adults. These findings suggest that time-frequency EEG analyses can provide insights into differential contributions of sensorimotor and cognitive information processing to distinct aspects of balance control post-stroke.
Categories
Stroke
Palliative Care -Recent Perspectives
Srinivasan Avathvadi Venkatesan
Emeritus Professor The Tamilnadu Dr. MGR Medical University, Chennai, India
Abstract
Introduction: Palliative care is a specialist medical care for people with serious illness. The goal of palliative care is to improve quality of life of patient and the family,Group of doctors, nurses and other specialists form the core group of people who achieve this goal by working for need of the patient not on the progression of diseases. It is given with the Curative treatment. Types of illness. Alzhiemer’s disease, 2)Parkinson’s disease 3)Amyotropic lateral sclerosis 4)Cancer 5)Congestive cardiac failure 6)Chronic kidney disease.7)Chronic Obstructive Airway Illness 8)Pulmonary disease. Quality of life is the central theme in Palliative care.Palliative care in the people has increased the quality of life than with the treatment of the diseases alone. This palliative care must be provided directly to the patients and the family
The PNEUMONIC for Palliative Care is I AM HAPPY
I-If it is right for you A- Access the provider directly M-Multi disciplinary team
H-Handouts for patients and family A-Always match with options to the goal
P- Podcasts P- Palliative care Y-Yes
Six Pillars of palliative care:1) medical 2) psychological 3) cultural 4) social 5) spiritual 6) nursing Geriatric palliative care(GPC)
GPC requires integrating the disciplines of Hospital medicine and the palliative care 2) Domain of Palliative care;related geriatric specific concern 3) Care of the patient Nearing the End of Life 4) Ethical and legal aspects of care 5) Social aspects of the care
Friendly health system -4 M
1) M-what Matters 2) M-Mobility 3) M-Mentation 4) M-Medication
Geriatric-5M 1) M-Mutli complexity 2) multiple chronic conditions 3) Advanced medical conditions 4) Bio psychosocial 5) needs
Mind 1) Mentation 2) Dementia 3) Delirium 4) Depression
Mobility 1) Impaired waste and balance 2) Fall and injury prevention 3) Amount of mobility –Function
Medication- 1) poly pharmacy 2) De-prescribing 3) Optimal prescribing 4) Medication effects and burden Domains and Key Recommendations
1) Structure and processes of care.2) physical aspects .3) psychological and psychiatric aspects of care.4) Social aspects 5) Spiritual, religious, and existential aspects 6) Cultural aspects 7) Care of the imminently dying patient.8) Ethical and legal aspects of care
Goals: 1) Advance directives 2) Prognosis 3) functional status 5) Care plan 6) Treatment and therapy and Core values 7) Medical condition 8) Office visit or inpatient family meeting. 9) Challenges in GPC 10) Challenges in Clinical cases: GPC view points
Palliative care. Symptom management of a life limiting illness
Hospice care. Symptom management and comfort care at the end of life
Conclusion: WHO -Palliative care for older people; better practices are in five stages.1) Prevention 2) Early Identification.3) Comprehensive assessment.4) Management of Physical issues,including pain and other distressing symptoms.5) Psychological issues
Categories
Other
Daily life upper limb use asymmetric in below-elbow amputees
Binal Motawar, Kelli Buchanan, Scott Frey
University of Missouri, Columbia, USA
Abstract
Introduction: Healthy adults rely on their dominant side slightly more than their nondominant side in accelerometry measurement of their daily life movements (1). Prior accelerometry research found that below-elbow unilateral amputees rely more on their anatomical upper limb more than their prosthetic limb, based on data collected from sensors placed on prosthetic or anatomical forearm (2). Body may compensate for the lost upper limb below-elbow by altering movement of the proximal segment or upper arm. The goal of this study was to examine the effects of a below-elbow amputation on upper limb movement symmetry at the forearm and upper arm levels, in unilateral and bilateral amputees during daily life.
Methods: We recorded accelerometry data using a 4-sensor montage (forearm and upper arm level each on both sides) over a 3-day period during participants’ daily lives. A total of 35 amputees (6 bilateral, mean±SD 55.3±15.9 years old) with acquired below-elbow amputations and 46 healthy adults (53.5±14.5 years old) participated in this research. Accelerometer data was processed to computer %median reliance on a limb as a measure of symmetry.
Results: Unilateral amputees relied on their anatomical forearm and upper arms more than their prosthetic forearm and upper arm respectively (forearm: %median reliance on the affected limb of 12.6±8.0% for dominant amputations, 11±9.2% for nondominant amputations, upper arm: %median reliance on the affected limb of 31.5±6.1% for dominant amputations, 29.8±7.4% for nondominant amputations, p<0.05), regardless of the dominance of their amputated limb. This one-handedness of upper arm use in unilateral amputees increased by 11% when the prosthesis was not being worn (p<0.05). Bilateral amputees relied more on their dominant vs. nondominant forearm and upper arm levels (%median reliance of dominant forearm: 67.8±15.6%, upper arm: 61.4±11.5%, p<0.05). The degree of single- handedness was the greatest in unilateral amputees, followed by bilateral amputees and healthy adults for both levels (p<0.05 for all comparisons). The one-handedness did not correlate with the degree of prosthesis wear time in this study. The one-handedness did not differ between the two main prosthetic device types (mechanical vs myoelectric).
Conclusions: Upper limb use symmetry is altered resulting in more usage of the intact or dominant limb more after below-elbow amputation of a single or both upper limbs. This single-handed nature of the daily life accelerometry in upper limb amputees is more pronounced in unilateral amputees than bilateral amputees. More research is needed to examine the relationship between the one-handedness of upper limb use and functional outcomes after amputation.
Categories
Other
