Abstract
Background:
Cognitive impairment is common in patients receiving dialysis and is associated with morbidity and mortality. Existing approaches to administering face-to-face cognitive screening assessments like the Montreal Cognitive Assessment (MoCA) may be challenging to undertake in dialysis. Virtual reality (VR) technology may be a novel way to assess cognitive function in patients on dialysis.
Objective:
In a cohort of patients undergoing hemodialysis, the primary objective of this study is to evaluate the test-retest reliability, diagnostic performance, and agreement of an MoCA, generated using VR-based cognitive testing, to a face-to-face MoCA. Secondary objectives are to (1) evaluate changes in cognitive function over time using the VR-generated MoCA, (2) examine associations between cognitive impairment and mortality or hospitalization, and (3) assess the usability of VR-based cognitive testing.
Design:
This is a prospective cohort study (conducted from 2025-2028).
Setting:
Hemodialysis units affiliated with the Nova Scotia Health Renal Program.
Patients:
Incident (within 3 months of dialysis initiation) and prevalent patients receiving hemodialysis.
Measurements:
Cognitive function will be assessed using the React Neuro VR Headset and the paper-based MoCA. The VR cognitive assessment will include tests such as Smooth Pursuit, Trail Making A/B, Letter/Category Fluency, Boston Naming, Stroop, and Digit Span (Forward/Backward). The results of these tests will be used to generate an MoCA score using device software.
Methods:
The VR cognitive tests and face-to-face MoCA assessments will be conducted at baseline and week 2, with the order of assessments randomly determined. Subsequent VR cognitive assessments will be conducted once every 3 months (up to 12 months). Agreement will be assessed using Cohen’s kappa (dichotomizing the MoCA at <24), and existing approaches for continuous MoCA scores. Test-retest reliability will be assessed using a similar approach comparing baseline and 2-week scores. Associations between the VR-generated MoCA and outcomes will be analyzed using appropriate regression methods.
Results:
To date, we have recruited 84 patients, 75 of whom have completed at least their baseline assessment.
Limitations:
Potential challenges in VR implementation and patient adaptation, as well as the loud and distracting dialysis environment, could impact performance in cognitive assessments.
Conclusions:
This proposed study aims to evaluate test-retest reliability, performance, and agreement between a VR-generated and face-to-face MoCA. The VR technology may provide a reliable alternative to traditional cognitive testing in dialysis patients. The findings can be used to assist in the early identification of patients with cognitive impairment and may also pave the way for future research, including VR-delivered interventions to improve cognitive function and health outcomes in this population.
Introduction
Background
Dialysis is a form of kidney replacement therapy that may be used to support the more than 29 000 patients with kidney failure in Canada. 1 Patients receiving dialysis are at an increased risk of mortality and hospitalization.2-4 The rising burden of kidney failure may be due to a growing population of older adults with multimorbidity that predisposes kidney failure, including diabetes. Furthermore, many of these health conditions may independently associate with morbidity and mortality in the dialysis population. 5 Given the impact of dialysis on patients, it is crucial to identify key health conditions that, if addressed, may lead to improved outcomes for this vulnerable patient group.
A health condition of importance is cognitive impairment, which may be defined as difficulty in mental processing, including learning, memory, understanding, reasoning, and executive function. 6 In the general population, individuals with cognitive impairment are at risk for morbidity and mortality.7,8 Kidney failure is an independent risk factor for cognitive impairment.9-11 The prevalence of cognitive impairment is high across the spectrum of kidney disease,12-14 and patients with more severe kidney disease are at even higher risk.15,16 Patients initiating hemodialysis are especially vulnerable irrespective of chronological age;17-19 the prevalence of cognitive impairment is estimated to be 30% to 60% in this population. 20
Cognitive impairment has been shown in many studies to be an independent risk factor for all-cause mortality in hemodialysis.21-25 In contrast, less is known regarding the association between cognition and hospitalization, although hospitalization is a common outcome in dialysis. While not all studies have identified a link between cognitive function and morbidity/mortality, 26 this lack of association may in part be related to an interaction with frailty, a multi-determined state of vulnerability to poor health outcomes, highlighting the importance of capturing frailty severity in studies evaluating the impact of cognitive impairment on health outcomes.
Cognitive impairment is not just associated with poor health outcomes but is also important to dialysis patients. In a recent qualitative study, patients receiving dialysis were accepting of cognitive screening and had favorable experiences when administered the Montreal Cognitive Assessment (MoCA), 27 further emphasizing the value of cognitive screening in dialysis.
Current approaches to cognitive screening in dialysis involve the administration of cognitive screening tests
Early recognition of cognitive impairment may be valuable for identifying reversible or treatable causes. For example, changes in brain edema that may result from the dialysis procedure28,29 have been shown to be associated with cognitive impairment. 30 Despite this and other important impacts of dialysis on cognition, there remains an important knowledge gap regarding the prevalence of cognitive impairment in this population. In a scoping review of the dialysis population, only 46 studies evaluating the prevalence of cognitive impairment were identified over 15 years, and 70% only used the Mini-Mental Status Exam (MMSE). 20 This is of importance, as the MMSE as a stand-alone test has only modest accuracy 31 for diagnosing cognitive impairment and limited applicability for those from diverse culture and language groups. 20 In that same scoping review, most studies did not explicitly evaluate changes in cognition over time. In a recently completed systematic review 32 including studies up to 2022, 10 studies (out of 3644 screened abstracts) were identified that specifically evaluated the performance characteristics of cognitive screening tools in dialysis relative to a comprehensive neuropsychological battery. The MoCA had the highest diagnostic accuracy using a cut-off of 24 with a pooled sensitivity of 0.87 (95% confidence interval [CI] = 0.51-0.98), and the conclusion of this study was that the MoCA is recommended for cognitive impairment screening in dialysis. 32
In addition to the MoCA, 20 other screening tests have been used in the dialysis population. However, these tests are not without their limitations. Administration of a face-to-face test can limit sensitivity33,34 in detecting cognitive impairment in the dialysis unit, where other distractions (including other patients receiving dialysis and auditory or visual distractions due to the dialysis machine and environment) can prevail. In addition, face-to-face administration of screening tests requires additional personnel resources (ie, the test administrator) and may take longer to deliver while a patient is physically connected to a dialysis machine and receiving treatment. These tests can be administered outside of a dialysis unit to avoid these issues; however, dialysis patients prefer their care to be delivered in the dialysis unit to minimize their need for other health visits. 35 Furthermore, as recently identified in a qualitative study of care delivery to Canadian dialysis patients, providers emphasized the value of incorporating technology into dialysis care and information delivery. 36
Virtual reality technologies can reliably identify cognitive impairment in many different patient populations, and using virtual reality as a way of grading cognition may be advantageous in dialysis
The virtual reality (VR) technologies can be used to generate an artificial environment for its users, where visual and auditory stimuli are controlled by the software. Some VR technologies have been developed to provide stimuli and record user responses in the form of head and eye movements, and vocal responses. The VR technologies can be used to generate an artificial environment for its users, where visual and auditory stimuli are controlled by the software. The VR has been utilized as a tool that can provide assessment and assist with a clinical diagnosis of cognitive impairment and dementia. The VR has been utilized as a tool that can provide assessment and assist with a clinical diagnosis of cognitive impairment and dementia.37-40 In these studies, virtual scenarios and tasks administered using VR technology assess executive function, memory, attention, visuospatial function, and orientation. In addition, these virtual scenarios can help assess an individual’s ability to carry out activities of daily living. The benefits of VR may extend beyond diagnosis. Through cognitive training, VR technologies have also been suggested to help slow the progression of cognitive impairment in those with dementia.41-44 Using VR to assess cognition in hemodialysis patients may provide the necessary results for clinical management while using fewer resources than a face-to-face assessment. The virtual environments created through VR can allow for real-world assessments of cognition that directly relate to functional status.45,46 The programs available on VR platforms can also assess various domains of cognition not normally tested through current methods of cognitive assessment. For example, VR technology can assess ocular movements in response to stimuli, which provides additional information on cognitive function.47,48 In addition, VR technology can assess other aspects of cognition not typically measured, eg, spatial awareness and orientation 49 and adaptive tasks. 46 In this study, we will utilize a novel VR cognition device, created by REACT Neuro. This device captures direct information on eye-tracking, voice, and head and body movements. 50 REACT Neuro can capture multiple different brain function tasks through neuropsychological tests.51-54 The software completes the test within the VR headset and outputs results to a tablet device. REACT Neuro has 7 domains that are covered through cognitive testing (working memory, verbal fluency, oculomotor function, acquired knowledge, long-term memory, motor function, and processing speed). Data are captured through a series of tests that generate raw data, interpretable individualized test results, and most importantly, summative scores for established cognitive tests (specifically, the MoCA).
As mentioned, there exists a knowledge gap in the prevalence of cognitive impairment in dialysis patients. Part of this gap may be related to a lack of evidence-based interventions for cognitive impairment in patients undergoing dialysis. However, there is evidence that VR technology can be used in neurorehabilitation in patients with cognitive impairment. 55 In addition, a recent systematic review and meta-analysis showed improved physical and psychological outcomes in dialysis patients who received VR training. 56
To our knowledge, no prior study has used virtual reality platforms to comprehensively assess cognitive function in a cohort of dialysis patients
Given the potential limitations of face-to-face cognitive screening in dialysis (complexities of the dialysis environment and the potential negative impact of visual/auditory distractions) as well as the desire for dialysis patients to have care integrated into their routine dialysis treatment, VR-administered cognitive testing may be optimal for this population. However, thus far, no prior study has used VR to assess cognitive function in dialysis, with a focus on reliability, performance characteristics, predictive validity, and agreement between a VR and face-to-face cognitive screen. In this study, we will use software developed by REACT Neuro and delivered through a VR headset to assess cognitive function in a cohort of adult patients with kidney failure treated with maintenance hemodialysis. We will compare an MoCA generated using VR-based cognitive testing to standard face-to-face paper administration of the MoCA. We will also assess changes in cognitive function over time and assess the association between cognitive impairment (using the VR-generated MoCA) and health outcomes (mortality and hospitalization). We anticipate that the MoCA assessment delivered through the VR headset will identify patients with cognitive impairment, will be reliable, and will demonstrate good predictive ability for clinically important outcomes, thus paving the way for future interventional studies that evaluate VR-administered cognitive training and its impact on cognitive function in dialysis.
Objectives
In a cohort of maintenance hemodialysis patients:
Primary objective: to evaluate agreement between a VR-generated and face-to-face (paper-based) MoCA, diagnostic performance, and test-retest reliability of the VR-generated MoCA.
Secondary objectives: to determine how cognitive function (defined using the VR-generated MoCA) changes over time (objective 2), to determine if cognitive impairment (defined separately using the VR-generated and face-to-face MoCA) is associated with mortality and hospitalization (objective 3), and to assess patient perspectives toward usability of a VR-administered cognitive screen (objective 4).
Methods
Study Design and Population
We will analyze a prospective cohort of adult patients treated with maintenance hemodialysis recruited over 2 years (2025–2027), with an additional year of follow-up to capture longitudinal measures of cognition and outcomes.
Recruitment
Participants will be recruited from 4 urban sites within the Nova Scotia Health Renal Program. Approximately 200 to 250 patients receive chronic hemodialysis at these sites, and 100 patients initiate hemodialysis annually, estimating an eligible population of 400 individuals over 2 years. Patients will be consented at any dialysis treatment centre prior to study initiation. This approach has been successful in a prior observational study of chronic dialysis patients. 57
Inclusion Criteria
Incident/prevalent hemodialysis patient.
Ability to hear and understand simple English conversation.
Exclusion Criteria
Severe visual/auditory limitation preventing use of the VR device.
Unable or unwilling to consent.
Physical impairments limiting ability to conduct the tests (some tests that require no arm movement may still be administered).
Study Visits/Timing of Cognitive Assessments and System Usability
After consent (during which patients will receive information about the device), patients will undergo a baseline cognitive assessment using the React Neuro VR Headset (Figure 1 and Table 1 for study timeline). This assessment will be repeated 2 weeks after the initial assessment to assess test-retest reliability (see section “Data Analysis Plan and Sample Size”). In all patients at baseline and week 2, a face-to-face paper MoCA will be completed by a research assistant (RA) at the same time as the VR cognitive assessment. The order of assessment (face-to-face vs VR) will be randomly determined (https://www.randomizer.org/) for both time points by the RA. In addition, patient perceptions of usability of the VR cognitive assessment will be determined using the virtual reality system usability questionnaire (VRSUQ) after each VR cognitive assessment.
Patient recruitment and time of cognitive assessments and data collection.
Schedule of Events.
The cognitive assessments will occur during a routine hemodialysis session 58 on either the treatment following the Long Interdialytic Interval (LIDI) (Monday or Tuesday) or midweek treatment (Wednesday or Thursday). We chose to administer the cognitive testing during a hemodialysis session to avoid delaying dialysis initiation (which is unfavorable to patients) and to minimize the impact of transient cognitive changes that occur toward the end of a hemodialysis session. 59 This suggestion was based on patient feedback. We chose to do the assessment on a day where routine clinical rounds with a nephrologist were not being conducted (most commonly a midweek treatment) to minimize burden to patients. For incident patients, the baseline cognitive assessment will occur during the second week after dialysis initiation. We chose the second week based on patient feedback; the first week of hemodialysis may be overwhelming for patients and initiating a study on a treatment during this week may misrepresent cognitive function. No assessments will be conducted during hospitalization. After discharge, cognitive assessments will be delayed until patients have recovered from their hospitalization.
After the baseline and 2-week cognitive assessment, a VR cognitive assessment will be conducted 1 hour into a dialysis session using the same approach as noted above, once every 3 months until Month 12 (Table 1). Patients will be censored after modality transitions to peritoneal dialysis, home hemodialysis, or following transplantation.
Data Collection and Measurement
Cognitive Function Tests
The VR cognitive assessment will consist of a series of individual cognitive tests some of which will be collated using software to generate an MoCA score using internal algorithms. The face-to-face MoCA (Figure 2) is a screening test of cognition that consists of 8 sections and a total score out of 30. Impairment will be defined as a score <24 based on previously published data. 32 Face-to-face testing will be conducted by the same RA who will be fitting patients with the VR devices. The RA has been trained and certified in administering face-to-face MoCA testing.
Montreal cognitive assessment (MoCA).
Individual VR tests are summarized in Table 2 and include the following:
VR Tests Used to Assess Cognitive Function.
Smooth pursuit: this test involves participants tracking an object on the VR headset using their eyes. Changes in fundamental eye movements have been demonstrated in patients with cognitive impairment.60,61 Scoring will be based on individual eye positions, reaction time, phase of each eye, and eye movements.
Trail making A/B: part A of the Trail Making test involves connecting a series of numbers in numerical order. Part B involves connecting numbers and letters in an alternating fashion (1 → A→ 2 → B, etc). Images will be shown on the VR headset, and trail making will be performed using the handheld component of the VR device (using the arm without the patient’s dialysis access). The trail-making tests assess working memory, motor function, and processing speed54,62 and have been used in hemodialysis.33,34,59,63,64 Scoring will be based on the time taken to complete the test, adjusted for the number of errors.
Letter/category fluency: this test involves asking participants to produce words beginning with a specific letter within a specified period-of-time to test verbal fluency/acquired knowledge. This test has been previously validated.52,65,66 Verbal fluency may be impacted by literacy and vocabulary size; therefore, abnormal results will be interpreted with caution and in the context of education. 67
Boston naming: This test involves showing participants objects to name. The degree of difficulty will increase throughout the test. This test evaluates long-term memory, acquired knowledge, processing speed, and linguistic breakdown. 68 Scoring will be done based on the number of objects named correctly.
Stroop: this test shows participants color words that are also colored, after which they name the color of the words presented. This test is commonly used in neuropsychological assessment,69,70 and scoring will be done based on the time taken to complete the test, adjusted for the number of errors.
Digit span (forward/backward): this test involves presenting a series of numbers and then asking participants to repeat these numbers forwards or backwards. The digit span test will assess working memory and processing speed.52,71,72 Scoring will be based on the longest sequence of numbers successfully reached, adjusted for the number of attempts.
Baseline Characteristics and Other Measures
Data will be collected at baseline to characterize the population being studied. Demographic data will include age, self-reported sex, ethnicity, weight, height, education, occupation, and socioeconomic status. Comorbidities/laboratory data will be collected from an existing dialysis electronic medical record and include mild cognitive impairment/dementia, cerebrovascular disease, presence of visual and auditory impairments, history of mood disorder, diabetes, coronary artery disease, congestive heart failure, peripheral vascular disease, prior malignancy (including type), serum albumin, number and type of prescriptions, and over-the-counter medications. Dialysis characteristics will include length of time since hemodialysis start, hemodialysis frequency, urea reduction ratio, Kt/V, and dialysis access. Characteristics about functional status will be completed using a questionnaire that asks patients to rate their ability to perform instrumental and basic activities of daily living, developed for a prior study 57 and the clinical frailty scale, a well-established tool used in our dialysis population. 73
Outcomes
The primary outcome, cognitive impairment, will be defined by an MoCA score of <24, and separately reported for those who will undergo the VR-generated and face-to-face assessment (objective 1). An MoCA of <24 will also be used for objective 2. For objective 3 (risk of death and hospitalization), time to all-cause mortality will be censored at transplantation and end of follow-up. Notably, dialysis patients who die, no longer attend treatments which trigger an immediate clinical follow-up as part of the standard of care. All hospitalizations will be captured by the RA using electronic records, inclusive of the date of admission/discharge and reason for admission. For objective 4, patients will complete the VRSUQ immediately following each VR cognitive assessment. The VRSUQ was chosen as it is a recently developed, VR-specific, usability questionnaire that has been validated against longstanding usability scales, captures usability metrics specific to VR tools, and is not overly burdensome on patients (~3 minutes to complete). 74 The VRSUQ (Table 3) consists of 9 questions, each rated on a 5-point Likert scale, yielding a total raw score ranging from 5 to 45. To standardize the score on a 0 to 100 scale, the following formula is applied: VRSUQ score = [(raw total score / 9) − 1] × 25. 74 An interim analysis of VRSUQ results will be done at 6 and 12 months from first participant enrollment. Based on these results, patient focus groups may be organized with consenting participants to better enrich usability data to help inform future implementation work. Individual tests of cognition captured by the VR device are not the primary focus of this study but will be described. For individual tests, impairment will be defined as a score >1.5 SD below the general population mean, standardized to age, sex, and education, as has been done in prior studies. 73
VR System Usability Questionnaire (VRSUQ).
Data Storage, Management, and Access and Confidentiality
The REDCap web-based application run through the NS Health Research Methods Unit will be used for data entry. Anonymized patient data (stripped of identifiers) will be stored in the REACT Neuro storage database in Boston, Massachusetts, which is Health Insurance Portability and Accountability Act (HIPPA) compliant. The study will be performed in accordance with all Nova Scotia Health requirements; accordingly, ethics approval for this study has been received (REB # 1029376).
Data Analysis Plan and Sample Size
A detailed statistical analysis plan will be finalized prior to analysis; a summary of key planned statistical analyses is summarized below.
Baseline characteristics of the study cohort will be described using means and standard deviations for continuous, normally distributed variables, medians and Q1 to Q3 for continuous non-normally distributed variables, and N (%) for categorical variables, respectively:
Objective 1. Neuropsychological evaluation is the gold standard for diagnosing cognitive impairment; however, due to resource limitations, it will not be used in this study. Instead, the VR-generated MoCA will be evaluated against the face-to-face MoCA which serves as the reference. First, both the VR-generated and face-to-face MoCA will be summarized using means (± standard deviation), and the proportion of patients with scores <24 will be reported. Performance characteristics relative to the face-to-face MoCA (including sensitivity, specificity, negative and positive predictive values) and classification accuracy (using area under the receiver operating characteristic curve) will be reported at baseline and 2 weeks. Agreement between the VR-generated and face-to-face MoCA (dichotomized at <24) will be assessed using Cohen’s kappa (moderate agreement defined as 0.41-0.60; strong agreement as 0.61-0.80). For continuous MoCA scores, agreement will be assessed using the intraclass correlation coefficient (ICC). Bland-Altman plots will be used to assess measurement bias and limits of agreement between approaches. Test-retest reliability for each approach to the MoCA will also be assessed using ICC between baseline and 2-week scores and Cohen’s kappa for consistency over time (using the dichotomized threshold of <24). Alternate forms reliability (the consistency of test results between 2 different measures of the same construct) will also be assessed using ICC. Given that the VR-generated MoCA may classify more individuals as cognitively impaired compared to face-to-face testing, the net reclassification index will be calculated to assess the incremental classification value.
Objective 2. Changes in the average MoCA score over time and the proportion of patients with a MoCA <24 will be reported and graphically displayed. Trends in continuous scores will be assessed using ordinary least squares regression, and changes in categorical scores will be evaluated using the nonparametric Mann-Kendall test. If a non-monotonic trend is suspected, the Anderson-Darling test will be used to assess the distributional shift over time.
Objective 3 (predictive validity). The association between VR-generated and separately, face-to-face administered MoCA (using a threshold of <24) and time to death will be analyzed using a multivariable Cox proportional hazards regression and reported using hazard ratios and 95% confidence intervals. The baseline MoCA will be the primary exposure; however, in a sensitivity analysis, we will take the average of the baseline and 2-week MoCA. In addition, we will incorporate variables associated with mortality based on a prior dialysis study, 73 including but not limited to the clinical frailty scale score, dialysis vintage (to account for survivor bias), age, sex, diabetes, body mass index, and albumin. Hospitalization rate will be analyzed using adjusted negative binomial regression (using those variables noted above) to account for over-dispersion.
Objective 4. VRSUQ results will be tabulated and assessed for mean and median scores for each assessment category, as well as overall compiled scores. Subgroup analysis will be performed to identify any differences between patient groups exist, such as prevalent vs incident patients, age, and sex.
The current study will be undertaken with a primary objective to evaluate the agreement between 2 approaches to the MoCA and reliability of the VR-generated MoCA between 2 measurements. Prior studies comparing the performance of VR technology in screening for cognitive impairment have demonstrated high levels of accuracy and agreement (area under that curve [AUC] of 0.86 for those <70 and 0.97 for those ≥70 years 36 and weighted kappa statistics of 93% and 100%, respectively). 75 For this study, we assume a recruitment rate of 60% (N = 240), prevalence of cognitive impairment of 30% (the lower end of reported studies). Using a 2-tailed alpha of 0.05, power of 80%, and expected kappa of 0.8, we would need to recruit 149 and 239 patients to demonstrate kappa’s of 0.60 and 0.65, respectively (moderate agreement). This corresponds to recruitment rates of 37% and 60%.
Results
To date, 84 patients have been recruited to participate in this study and 75 have completed the first assessment.
Discussion
In those living with kidney failure requiring hemodialysis, cognitive impairment can result in worse health outcomes including hospitalization and mortality, and thus, to practitioners, knowing who may be at-risk for these outcomes is important for prognostication and may inform therapeutic strategies. Even more importantly, patients living with cognitive impairment have identified their own health care preferences. In a systematic review of 12 studies, it was identified that patients prefer to be informed about the diagnosis of cognitive impairment as early as possible for care planning. 76 Thus, studies aimed at accurately and efficiently capturing cognitive impairment early after dialysis initiation are of high priority.
This study will provide insights into how cognition changes in dialysis patients and whether there may be an unrecognized burden of cognitive impairment at baseline. Importantly, this is the first step toward interventions to improve cognitive function in what is clearly a vulnerable population. This information is especially important in patients undergoing hemodialysis if cognitive impairment results in worse health outcomes. It is also possible that VR-generated tests will identify impairment that is not identified when measuring the MoCA with a traditional approach.
In addition to the immediate benefits of this research, it is hoped that this project will be a launching pad for future research aimed at generalizability, assessing stakeholder perspectives toward cognitive impairment and ultimately, intervention. Regarding stakeholder perspectives, our patient partner identified the importance of understanding the patient experience and how it associates with cognitive function. In this regard, if the VR assessments of cognition prove to be useful in dialysis, future work will evaluate patients’ perspectives of cognitive function and considerations around usability of the VR devices. Identifying patients’ experience with respect to cognitive function will also allow a detailed analysis of the intersection of quality of life and cognition, something that was identified as a future area of importance (by our research team and patient partner).
Conclusions
Screening for cognitive impairment early on may be important in some patients undergoing hemodialysis. By using VR assessments, this study aims to provide a seamless but accurate method of detecting cognitive impairment in hemodialysis patients. This has the potential to help improve health outcomes and inform future interventions in this vulnerable population.
Footnotes
Acknowledgements
The authors would like to thank our patient partner for providing great insight into the patient experience on dialysis and using the VR device. The authors would like to thank REACT Neuro for providing further information regarding their virtual reality software and internal algorithm for assessing cognitive function. The authors gratefully acknowledge the outstanding generosity of Robert and Nancy Van Wart, the River Philip Foundation, and the Dartmouth General Hospital Foundation, whose significant contributions supported this work.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was financially supported by Robert and Nancy Van Wart, the River Philip Foundation, the Dartmouth General Hospital Foundation, and the Nova Scotia Health Research Fund.
Declaration of Conflicting Interests
The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: KT has a financial interest in REACT Neuro, has received unrestricted grant funding from Otsuka for investigator initiated research and has conducted advisory board work and CME work for Otsuka and Virtual Hallway.