Cortical Activation During Split-Belt Treadmill Walking in People With Parkinson’s Disease and Healthy Controls

Abstract

Background

People with Parkinson’s disease (PwPD) have difficulty adapting their gait to asymmetrical conditions. Objective. We investigated cortical activity between 42 PwPD (HY 2-3) and 42 healthy controls using functional near-infrared spectroscopy during tied-belt (TB) and split-belt (SB) treadmill walking.

Methods

Oxygenated hemoglobin (HbO2) was measured in the prefrontal cortex, supplementary motor area (SMA), premotor cortex (PMC), and posterior parietal cortex (PPC) during 3 blocks of treadmill walking: (1) with the belts moving at the same speed (TB) and (2) when the speed of 1 side was reduced by 50% (SB; 2 blocks). The ability to adjust gait to asymmetric conditions was quantified by step length asymmetry and its variability.

Results

Adaptive gait was worse during the last 5 steps of SB versus TB in PwPD compared to controls. PwPD showed higher HbO2 in the PMC (P = .005) and PPC (P = .004) relative to controls, regardless of condition. However, an increase in HbO2 in the SMA during SB was shown relative to TB in PwPD, a change not observed in controls (group × condition interaction P = .048; pairwise post hoc P = .032). Interestingly, increased PPC activity in PwPD was associated with poorer adapted gait.

Conclusions

Both regular and adaptive gait required enhanced cortical processing in PwPD, as evidenced by the increased activation in the PMC and PPC. However, this heightened cortical activity did not correlate with a reduction in gait asymmetry, suggesting that these changes might be maladaptive. Instead, the elevated cortical activity may reflect the challenges PwPD face in adapting to asymmetrical walking conditions. Careful interpretation is warranted given the relatively small sample of mildly affected PwPD, limiting generalizability to the broader population and the measurement errors inherent to functional near-infrared spectroscopy .

Keywords

adaptive gait fNIRS cortical activity Parkinson’s disease split-belt treadmill

Introduction

Parkinson’s disease (PD) leads to diminished dopamine transmission in the basal ganglia, significantly impacting both motor and non-motor circuitries.¹ Particularly, the control of automatic movements, such as gait, is disrupted.² As a result, people with PD (PwPD) have difficulty adapting their stepping pattern to challenging and asymmetrical conditions, including turning,^3-5 often provoking gait instability and freezing of gait. These adaptive deficits worsen as the disease progresses, and are associated with an increased risk of falls.^6,7 In this study, we used a split-belt (SB) treadmill as a paradigm for studying rapid adaptive gait difficulties in PD for 3 reasons. First, a SB treadmill can induce highly standardized asymmetrical gait perturbations by controlling the walking speed of both legs separately, previously utilized to study prolonged motor learning processes.⁸ Second, our previous work showed that a 4-week SB training program resulted in substantial improvements of SB-walking in PD.⁹ Third, Hinton et al¹⁰ used positron emission tomography (PET) to highlight that the supplementary motor area (SMA) and the posterior parietal cortex (PPC) are central hubs in the so-called adaptive gait network studied during SB-walking in healthy subjects. However, the neural mechanisms underpinning problems with rapid gait adjustments in PD are still unclear.

Functional near-infrared spectroscopy (fNIRS) measures cortical brain activity by detecting oxygenated (HbO₂) and deoxygenated (HHb) hemoglobin levels, and has been used to study upright walking in PwPD and healthy controls (HCs).^11-16 Most previous studies have focused exclusively on the prefrontal cortex (PFC) and found inconsistent results, likely due to protocol heterogeneity and variations in the tasks used.^11-15,17 Recently, Pelicioni et al¹⁶ investigated the premotor cortex (PMC), the SMA, and the PFC in PwPD and HC using fNIRS during a motor planning task. They compared overground walking with more complex tasks, such as stepping onto or avoiding a target and found that PwPD increased PMC activity during both simple and complex walking tasks compared to HC.¹⁶ Moreover, HC exhibited increased PFC activity as the task became more complex, whereas no such difference was seen in PwPD. The authors suggested that the PFC in PwPD might have already been maximally engaged in compensatory executive control during simple walking.¹⁶ Until now, it is unknown whether PwPD modulate brain activity differently than HC when adapting to asymmetric gait on a SB treadmill. This question is relevant as SB-perturbations resemble the asymmetrical gait adjustments during turning, which are closely related to freezing and falling.^6,7

The present study aimed to reveal the brain activity changes during adaptation to asymmetric walking in PwPD and HC in 4 regions of interest (ROIs), earlier shown to be active during adaptive gait in healthy subjects, including the PFC, SMA, PMC, and PPC.^10,16 We expected:

An overall increase in cortical activity in all ROIs during SB- compared to TB-walking (condition effect) given the increased adaptive challenge imposed by SB-walking.

General increased activity in all ROIs in PwPD versus HC during TB-walking (group effect), reflecting compensation for the loss of gait automaticity in PD,¹⁶ except for the SMA due to the hypoactivity often observed in this region.¹⁸

That PwPD would increase their cortical activity during SB-walking in the PMC and PPC compared to TB-walking, more so than HC (group × condition interaction), as these regions were previously associated with motor compensation.¹⁶ Less involvement of the SMA and PFC in coping with SB-walking was also expected due to alterations in SMA-function in PD¹⁸ and the implicit nature of SB-perturbations.¹⁹

Methods

Participants

We included 84 participants (42 PwPD and 42 HC) using the following inclusion criteria: (1) self-reported ability to walk 5 minutes without assistance or assistive device and (2) no probable dementia (Mini-Mental State Examination ≥24). For PwPD, additional inclusion criteria were: (3) clinical diagnosis of idiopathic PD by a neurologist based on established criteria,²⁰ (4) Hoehn & Yahr stage II or III in the ON medication state,²¹ (5) steady antiparkinsonian medication schedule the month before enrollment, and (6) stable medication state for at least 4 hours. Exclusion criteria were: (1) participation in another clinical study, (2) acute orthopedic injuries or musculoskeletal problems affecting balance and gait, (3) other neurological or significant psychiatric impairments, (4) self-reported cardiovascular risk factors, (5) previous SB-walking exposure, and (6) presence of a deep brain stimulator. This study was approved by the Ethical Committee Research UZ/KU Leuven (study ID: S64280). All participants gave written informed consent before participation. The trial was registered at the Open Science Framework (https://osf.io/gkeq3/).

Assessment of fNIRS and Gait on the Treadmill

Participants visited our laboratory for a single 3.5-hour assessment. PwPD were assessed in their “ON” medication state, subjectively confirmed by the participants. Data was collected during 3 blocks, including 1 TB- and 2 SB-walking blocks offered in a fixed order, each consisting of 7 trials (see Figure 1). TB-walking was always performed first to avoid carry-over effects. The 2 SB-blocks (SB1 and SB2) involved 50% reduced belt speed for either the right or left leg, the order of which was randomized for leg dominance based on the question: “With which leg would you kick a ball?”.²²

Figure 1.

Protocol and within-trial timing of the fNIRS assessment. (A) Displays the overall protocol. Each of the 3 blocks (TB, SB1, and SB2) consisted of 7 trials. (B) Each trial consisted of 15 seconds of quiet stance followed by an acceleration phase (7.5 seconds) to reach comfortable walking speed. Next, participants underwent 25 seconds of TB- or SB-walking followed by a deceleration phase.

Figure 1 displays the protocol (A) and within-trial timing (B). A trial started with 15 seconds of quiet stance to acquire baseline cortical activity. During quiet stance participants were instructed to stand still while holding the handrail and look forward to a fixation point on the wall without talking. Next, the treadmill (Motek Medical, Amsterdam, The Netherlands) accelerated to the designated speed, personalized for each subject after which the belts would run in TB or SB for 25 seconds, followed by a deceleration to 0 m/s. As such, the calculation of the hemodynamic change between baseline (quiet stance) and gait could be calculated.

Gait adaptation was quantified as the mean step length asymmetry, standard deviation (SD), and coefficient of variation (CV) of asymmetry over the total trial and the first and last 5 steps of each trial. These outcomes were captured with VICON (Vicon Motion Systems Ltd, Oxford, UK) with 6 passive reflective markers placed on the feet (lateral malleolus, heel, and tip of the shoe on each foot) and calculated as described in previous work.²³

Treadmill speed was based on the participant’s comfortable overground gait speed (see Supplemental Material for details). Before the protocol commenced, participants had a 2-minute TB familiarization. If the comfortable overground gait speed was perceived as too fast on the treadmill, this was reduced in a standardized fashion using 5% steps until the participant confirmed that the speed was comfortable. Then, participants were familiarized with the above-mentioned blocked design, where they experienced 1 trial of each condition (TB, SB1, and SB2). Participants were asked to let go of the handrails as soon as they reached the desired speed. However, if participants were not able to walk safely without holding the handrails, they were allowed to hold on during the entire assessment for consistency (N = 8, in PD only). Participants wore a non-weight-bearing safety harness and were offered regular breaks of ±1 minute between the 3 walking blocks to minimize fatigue.

Clinical and Cognitive Assessment

The following balance and cognitive assessments were obtained from all participants: (1) Mini-Balance Evaluations System Test (MiniBEST), (2) Falls Efficacy Scale, (3) Trail-Making-Test Part A and B, (4) Montreal Cognitive Assessment, and (5) Frontal Assessment Battery. From the PwPD, we also collected: (1) the Movement Disorders Society—Unified Parkinson’s Disease Rating Scale Part III for motor symptoms, (2) Levodopa Equivalent Daily Dosage, and (3) the New Freezing of Gait Questionnaire to determine the patients’ self-reported freezing status.

fNIRS Data Acquisition and Analysis

Brain oxygenated hemoglobin (HbO₂) and deoxygenated hemoglobin (HHb) levels were recorded with the Aurora software (version 2021.4.1.1) using a continuous wave fNIRS system (NIRSport2, NIRx, Berlin, Germany) including 16 LED sources and 16 detectors. Wavelengths were preset at 760 and 850 nm with a sampling frequency of 7.81 Hz. The 32 optodes were placed approximately 3 cm apart on a cap, corresponding with the participant’s head circumference. The cap included 16 short channels, 1 for each source (8 mm distance) to correct for possible extracerebral physiological noise²⁴ and 2 accelerometers on the back of the cap to correct for head movements.²⁵ An opaque cap covered the optodes to reduce possible external light interference. The reference for cap placement was Cz, determined by the locations of the nasion, inion, and preauricular points. Before testing, signal quality was checked and optodes were adjusted if needed.

Brain ROIs and corresponding optode locations were determined according to Zimeo-Morais et al²⁶ using the fOLD toolbox and Brodmann areas (see Supplemental (S) Figure S1). ROIs were selected based on their known involvement in adaptive gait.¹⁰ The specificity of the ROI was set at ≥50% as per previous work.^16,27,28 ROIs included the prefrontal cortex (PFC, BA9, BA10, and BA46; SMA, BA8, or frontal eye fields and BA6 medial; PMC and BA6 lateral; and the PPC, BA5, and BA7).²⁸ Previous work showed that PFC, PPC, and PMC had excellent test-retest reliability in healthy older adults when performing a balance task (intra-class correlations ≥ 0.78 without cap removal). However, this value was moderate for the SMA (intraclass correlation coefficient = .48).²⁸ The midline channels were excluded from the analysis (see Supplemental Figure S1) to avoid noise due to cerebral spinal fluid in the superior sagittal sinus.^28,29

The fNIRS data were exported in .nirs format and preprocessed in NIRStoolbox in MATLAB version R2019a (MathWorks, Natick, USA).³⁰ First, the data was resampled to 5 Hz and converted to optical density data using the Beer–Lambert Law with a partial path length correction factor of 0.1. Second, the change in hemodynamic response between rest and walking (TB or SB) was estimated utilizing a General Linear Model (GLM) with autoregressive pre-whitening, using the iteratively reweighted least squares method. Boxcar stimuli for the duration of the acceleration, steady-walking, and deceleration phases were created and convolved with the canonical hemodynamic response function before entering them into the model. The GLM model also included the short channels and accelerometry data as noise regressors, which is recommended for accommodating physiological noise and movement artefacts.^30-32 The beta for the steady-state walking within each channel was averaged according to the specified ROIs. Mean relative HbO₂ (μmol/L) in the PFC, SMA, PMC, and PPC was chosen as the primary fNIRS outcome. Secondary fNIRS outcomes included the mean relative HHb (μmol/L) values due to the lower signal-to-noise ratio in this signal.^17,33,34

Statistical Analysis

The preregistered analysis plan can be found at https://osf.io/gkeq3/. The group (PwPD versus HC) × condition (SB vs TB) interaction was the main interest as we hypothesized that PwPD would show greater increases in PMC and PPC activity from TB to SB than HC. We calculated the sample size with G*Power (v3.1), using a previous fNIRS study comparing PD and HC during simple and more complex overground walking. This showed an effect size (Cohen’s D) of 0.31 in the PFC,¹⁵ resulting in 42 participants per group when a repeated-measures analysis of variance with a within-between factorial design was taken into account (α = .05 and β = .20).

To analyze between-group differences in the continuous demographic and overground gait data, Independent Student’s t-tests were used or Mann–Whitney-U-Tests in case of an abnormal data distribution. For categorical data, Chi-squared tests were performed. Linear Mixed Models with the factors group (HC and PwPD) and condition (TB and SB) and age and treadmill speed as covariates were used to compare treadmill hemodynamic responses (mean relative HbO₂ and mean relative HHb) and step length asymmetry outcomes (mean, CV, and SD). We intended to use the CV as an outcome of variability (preregistered),^9,23,35 but instead focused on the SD for ease of interpretability (not preregistered). For completeness, we report the CV values in the appendix (Supplemental Tables S1-S3). We also explored the possible differences between SB1 and SB2 exposures to check the validity of pooling both SB-conditions. In addition, we examined the effect of outliers (based on the 1.5 × inter quartile range [IQR]) and the influence of handrail use on our main model (TB vs SB1). The influence of leg and disease side dominance were explored in separate models. Post hoc tests were Bonferroni corrected. To perform an anatomically fine-graded analysis, we also conducted a channel-to-channel analysis with NIRStoolbox.³⁰ All other statistical analyses were performed in SPSS v.28 (IBM Corp., Armonk, NY). Statistical significance was set at P < .05.

To investigate whether the hemodynamic changes between TB- and SB-walking (∆HbO₂) were related to adaptation performance and clinical characteristics, we performed an exploratory correlation analysis including the ROIs and the separate channels which were statistically significant after Bonferroni correction, and with significant Group (not preregistered) or Group × Condition (preregistered) effects. Greater changes of asymmetry (∆asymmetry) and variability (∆SD) between TB- and SB-walking were interpreted as worse adaptation performance. The correlation analyses were Holm–Bonferroni corrected.

Results

Participants’ Characteristics

Participant characteristics are displayed in Table 1. Overall, the age and sex distributions between groups were similar. Comfortable treadmill walking speed was lower in PwPD compared to HC (P < .001). PwPD also showed worse balance, more impaired cognitive performance, higher fear of falling, and slower and fewer turns during the turning-in-place task.

Table 1.

Participant Characteristics.

Characteristic	HC (N = 42)	PD (N = 42)	P-value
Sex (M/F)^a	22/20	23/19	.661
Age (y)	65.62 (5.8)	63.95 (6.9)	.236
Cognitive scales
Education (y)	15.26 (2.69)	14.48 (2.38)	.160
MoCA (0-30)	26.74 (1.73)	25.88 (2.59)	.078
FAB^b (0-18)	16.95 (1.25)	15.69 (2.08)	.002
TMT part A^b (s)	32.84 (10.67)	43.16 (13.40)	<.001
TMT part B^b (s)	72.84 (22.72)	96.91 (49.02)	.005
TMT B-A^b (s)	40.00 (20.37)	53.74 (45.37)	.142
Clinical scales
FES-I^b (16-64)	18.83 (2.54)	26.36 (8.24)	<.001
MiniBEST^b (0-28)	25.19 (4.34)	22.28 (5.51)	<.001
PD-specific scales
Duration since diagnosis (years)	—	8.24 (5.02)	—
LEDD (mg)	—	681.89 (401.47)	—
HY-stage (2/3)	—	36/6	—
MDS-UPDRS-III (0-132)	—	30.52 (11.73)	—
Participants with/without FOG^c	—	10/32	—
Gait outcomes
Personalized treadmill speed (m/s)	1.13 (0.11)	1.02 (0.16)	<.001
Turning speed^b (°/s)	145.26 (31.63)	100.52 (33.47)	<.001
Total 360° turns in 60 s^b	22.5 (4.74)	15.33 (5.25)	<.001

Abbreviations: HC, healthy controls; PD, people with Parkinson’s disease; M, Male; F, Female; MoCA, Montreal Cognitive Assessment; FAB, Frontal Assessment Battery; TMT, Trail-making-test; FES-I, Falls Efficacy Scale International; MiniBEST, Mini Balance Evaluations Systems Test; LEDD, Levodopa Equivalent Daily Dose; mg, milligrams; HY-stage, Hoehn and Yahr stage; MDS-UPDRS-III, Movement Disorders Society-Unified Parkinson’s Disease Rating Scale Part III; FOG, Freezing of Gait.

Significant values (P < .05) are marked in bold. Values represent mean (standard deviation), except for HY-stage and FOG-status. All comparisons made with independent student’s t test except: ^aChi-squared test; ^bMann–Whitney U test. ^cDetermined with question 1 of the New Freezing of Gait Questionnaire.

Gait Outcomes

When comparing the first block of SB (SB1) with the second block (SB2) in the 2 groups, we found no difference in mean step length asymmetry between SB1 and SB2 (Condition effect: P = .135, Supplemental (S) Table S1). Therefore, behavioral results were pooled for SB1 and SB2. Both groups showed more asymmetry (all P < .001) and higher variability of asymmetry (all P < .001) during the SB-walking compared to TB (Supplemental Table S2). Additionally, we explored the initial and prolonged effects of the SB-perturbation by comparing the first and last 5 steps of each trial. This analysis showed a group × condition interaction effect for the last 5 steps of each trial (F_(1,80) = 4.80, P = .031) indicating more step length asymmetry in PwPD than HC (F_(1,80) = 5.21, P = .025). For the first 5 steps, no such interaction was present. The separate results of SB1 versus SB2 and TB versus SB1 are presented in Supplemental Tables S1 and S3.

Cortical Brain Activity Changes

When comparing the fNIRS results of SB1 versus SB2 (Supplemental Table S4), no differences were found in hemodynamic responses in both HC and PwPD. Therefore, we pooled the 2 blocks. We also found that disease side, leg dominance, outliers and handrail use did not substantially affect these results (Supplemental Tables S4-S6) and thus were not further considered in the models. The distribution of outliers did not show a different pattern for PD or HC or condition (median exclusion per ROI 3.36%). When comparing TB versus SB-pooled with TB versus SB1 (Table 2 and Supplemental Table S7) we found no meaningful differences between the model outcomes, justifying the approach taken.

Table 2.

Relative Oxygenated and Deoxygenated Hemoglobin Concentrations (HbO2/HHb) TB Versus SB Pooled in the Total ROI.

		HC (N = 42)		PD (N = 42)		Group	Condition	Interaction
Chromophore	ROI	TB	SB	TB	SB	P-value	P-value	P-value
HbO2	Total
(μmol/L)	PFC	3.75 (1.50)	3.72 (1.12)	3.35 (1.29)	4.33 (1.28)	.953	.336	.310
	SMA	8.08 (1.13)	7.78 (1.07)	8.08 (1.33)	9.94 (1.48)	.532	.150	.048
	PMC	3.93 (1.02)	4.33 (1.02)	8.71 (1.43)	9.11 (1.32)	.005	.451	.999
	PPC	6.45 (0.940	7.24 (1.11)	10.80 (1.16)	11.76 (1.14)	.004	.076	.861
HHb	Total
(μmol/L)	PFC	0.20 (0.46)	−0.02 (0.40)	−1.60 (0.63)	−1.35 (0.47)	.025	.920	.215
	SMA	1.33 (0.54)	0.59 (0.36)	−0.14 (0.76)	−0.67 (0.56)	.075	.036	.712
	PMC	0.11 (0.46)	−0.19 (0.45)	−1.49 (0.80)	−2.35 (0.63)	.023	.045	.324
	PPC	0.97 (0.42)	0.40 (0.35)	0.54 (0.46)	0.03 (0.36)	.436	.036	.899

Abbreviations: ROI, Region of Interest; HC, healthy controls; PD, people with Parkinson’s Disease; TB, Tied-belt; SB, Split-belt; HbO2, oxygenated hemoglobin; HHb, deoxygenated hemoglobin.

Significant values (P < .05) are marked in bold. Values are presented as mean (standard deviation).

Table 2 and Figure 2 show the main and interaction effects for the final models on HbO₂ and HHb for the ROIs in both hemispheres. The lateralized activity changes are reported in Supplemental Table S8. As for HbO₂, a significant group × condition effect was found in the SMA (total, F_(1,79) = 4.05, P = .048), signifying increased activity during SB-walking in PD only (F_(1,52) = 4.84, P = .032). In addition, main group effects were present in the PMC (total, F_(1,74) = 8.58, P = .005) and PPC (total, F_(1,80) = 9.04, P = .004), whereby PwPD had higher HbO₂ levels compared to HC. No significant differences in HbO₂ were observed for the PFC. Channel-to-channel analysis of HbO₂ showed similar results regarding the group differences (Figure 3, left panel). PwPD showed elevated HbO₂ concentrations compared to HC in the PMC and PPC areas. However, in contrast to the ROI analysis, the channel-to-channel analysis did not show a significant interaction effect in the right SMA (Figure 3, right panel).

Figure 2.

Mean relative activations in total PFC, SMA, PMC, and PPC.

Figure 3.

Channel to channel analysis for HbO2.

The HHb results corroborated the group effects found for HbO₂ (Table 2) in the PMC (total, F_(1,73) = 5.38, P = .023); however not for the PPC (total, F_(1,81) = 0.61, P = .436). HHb did not confirm the group × condition interaction effect found in the SMA for HbO₂ (total, F_(1,67) = 0.14, P = .712). However, a condition effect appeared for HHb in the SMA (total, F_(1,67) = 4.59, P = .036) and also for the PMC (total, F_(1,69) = 4.18, P = .045) and PPC (total, F_(1,80) = 4.56, P = .036), indicating that both HC and PwPD presented lower HHb values (ie, increased activity) during SB- compared to TB-walking. Additionally, a significant group effect was present in the PFC (total, F_(1,78) = 5.24, P = .025), suggesting that HHb was lower (ie, increased activity) in PwPD compared to HC, a result which was not mirrored by the HbO₂ results. Of note, the HbO₂ SMA interaction and the HHb results did not surpass Bonferroni corrections.

Correlations Between Behavioral, Clinical, and fNIRS Outcomes

We performed a correlation analysis between HbO2 changes in the significantly involved regions and the gait adjustments from TB to SB within each group. Figure 4A shows a significant positive correlation in PwPD indicating that a larger increase in activity in channel S15-D14 (higher ∆HbO₂), representing a superiorly located channel of the right PPC, was associated with higher variability in asymmetry (worse adaptation, ∆SD, r = .425, P = .005). A similar result was apparent when considering the total PPC-ROI (∆HbO₂, r = .35, P = .025), not shown in Figure 4. Figure 4B displays that larger increases in activity in the SMA-ROI (higher ∆HbO₂) was also associated with higher variability in asymmetry, but this time in the HC group (∆SD, r = .32, P = .042). Finally, in contrast to the other regions, Figure 4C shows that larger increases in activity in channel S13-D12, representing a more inferiorly located channel of the right PPC, was associated with lower asymmetry in PwPD but not in HC (better adaptation, Δasymmetry, r = −.324, P = .036). See Supplemental Table S9 for detailed results.

Figure 4.

Spearman correlations between ΔHbO2 (SB-TB HbO2) and adaptation performance (Δasymmetry and ΔSD asymmetry). Panel A and B show the correlation between ΔHbO2 and ΔSD asymmetry, Panel C shows the correlation between ΔHbO2 and Δasymmetry. Panel A and C include the ΔHbO2 of individual channels and panel B shows the ΔHbO2 of the averaged ROI. Correlations were performed to ascertain the within-group relationships between the respective outcomes. Only the correlation in panel C showed a significant between-group difference between slopes (interaction effect group × ΔHbO2, F(1) = 5.00, P = .028).

Associations between turning performance, clinical scales, and the HbO₂ results showed that in HC increased PPC-ROI activity was associated with worse turning performance (number of turns, r = −.34, P = .030), more impaired executive functioning (TMT part-A, r = .36, P = .019), and set-shifting skills (TMT part-B, r = .39, P = .010), as well as worse balance (MiniBEST, r = −.39, P = .012). No associations were found in the PwPD. As none of these correlations passed the Holm–Bonferroni correction no multiple regression analysis was undertaken. See Supplemental Table S10 for detailed outcomes.

Discussion

This is the first study investigating the effect of SB-walking on brain activity in HC and PwPD. Behavioral results showed no difference in adaptive treadmill walking between PwPD and HC, except for the last 5 SB-steps in which PwPD remained more asymmetrical. Unexpectedly, we found no changes in HbO₂ levels during SB walking relative to TB, although HHb did suggest increased activity in the PMC and PPC. Irrespective of walking condition, however, PwPD did show increased HbO₂ in the PMC and the PPC. In contrast to our hypothesis, during SB-walking only, PwPD specifically increased activity in the SMA, a change not observed in the HC.

The overall hyperactivation found in the PMC in PD relative to HC is in line with previous fNIRS work involving complex gait tasks.¹⁶ As gait automaticity becomes impaired, PwPD may rely more on secondary cortical motor areas such as the PMC even during normal treadmill walking.^36,37 The PMC is essential for selecting appropriate motor programs, a process that depends on the sensory input transmitted from the PPC.³⁸ Consistently, we also observed increased activity in the PPC in PD (see interpretation below). Increased brain activity during motor tasks has been interpreted as either a compensatory mechanism or unhelpful neural dedifferentiation³⁹ in aging or in individuals with neuropathology. Indeed, the HHb values suggested an increased PMC involvement during SB-walking in both groups, although not surpassing Bonferroni correction. In this study, we found no correlations between changes in PMC-related channels and gait performance, suggesting that the hyperactivity signifies dedifferentiation rather than compensation.^40,41

Our results also revealed hyperactivity in the PPC in PD compared to HC, irrespective of SB-perturbations (no interaction effect). Similar to the PMC, the HbO₂ values did not show a condition effect, whereas the HHb values did. This suggests that both groups increased their activity in response to the higher task demands of SB- versus TB-walking. Of note, the group and condition effects were not congruent in both chromophores, nor did the HHb values withstand multiple comparison correction. A recent neurostimulation study in young adults showed that tDCS suppression in the PPC during SB-walking resulted in more steps needed for successful adaptation and longer-lasting after-effects, corroborating that the PPC plays an important role in fine-tuning gait.^10,42,43 We used the Brodmann map in the fOLD toolbox to determine the anatomical location of our ROIs, and did not make an a-priori distinction between superior and inferior segments of the PPC. Exploratory correlation analyses between activity changes in the PPC and behavioral changes revealed a differentiated pattern for more superiorly and inferiorly located channels, only observed in PD. Increased PPC HbO₂ changes (TB versus SB) were associated with poorer adaptive gait (>variability) in the full ROI as well as in the more superiorly located channel. In contrast, such changes in the more inferior channel were associated with improved adaptive gait (<asymmetry). Several fMRI studies also observed different functions for various portions of the PPC, indicating that the superior PPC might be more involved in acute feedback-based movement planning, whereas the inferior PPC might utilize feedforward predictions beyond planning the situation at hand.^44-46 The fact that the abnormally high recruitment of the superior PPC was correlated with worse adaptive gait is in line with the Compensation-Related Utilization of Neural Circuits (CRUNCH) hypothesis,⁴⁷ suggesting that at the higher levels of task load, the compensatory mechanism may no longer be effective. This idea of “compensatory overload” was also previously found in PD, with respect to greater PFC activity during a toe-tapping task,⁴⁸ In HC, however, lower baseline scores in cognition, balance, and turning were associated with higher PPC–ROI activation, suggesting compensatory inefficiency in individuals with reduced compensatory reserve,⁴⁹ For the inferior PPC, a different pattern emerged: a positive brain–behavior association, which aligns with beneficial compensatory recruitment. Here, PwPD who engaged this brain area more effectively coped with perturbations, potentially through anticipatory motor planning. However, these correlations should be interpreted with caution as they were only observed in 2 channels and were not significant after correction for multiple comparisons.

Previous fMRI studies reported an overall hypoactivity of the SMA during various motor tasks in PD versus HC,^18,50,51 Contrary to our hypothesis, we found that PwPD increased HbO₂ in the SMA during SB-walking, more so than HC (significant group × condition interaction). As the SMA is important for bilateral movement control and gait initiation,^10,52 it is as such unsurprising that this area is involved during asymmetrical walking. However, 2 recent studies on gait initiation and adaptive walking did not show substantial increases in SMA activity in PD, unlike in HC,^16,27 This lack of increase was attributed to the close link between the SMA and the PD-related degradation of the basal ganglia, which hinders the modulation of SMA activity leading to a subsequent decline in behavioral performance,^16,27 In contrast, PwPD in our study were able to react to SB-perturbations to some extent but failed to completely correct their asymmetry in the last 5 steps of the SB-condition. Failure to fully correct asymmetry in response to SB-perturbations was found previously in PwPD with freezing.⁴ The abnormal SMA recruitment in this study probably signifies a higher engagement of compensatory reserve in PD⁵³ to manage the higher demands of SB-walking, although this result did not remain after multiple comparison correction. Additionally, no association was found between the change in SMA activity and the quality of adaptive behavior. Correlation analysis in HC did show that a SMA activity increase was associated with worse adaptive performance (more variability) in line with the concept of “compensational overload.”

We did not find changes in HbO₂ levels in the PFC, albeit that HHb values in PwPD suggested increased PFC activity compared to HC. However, this was not influenced by the treadmill conditions nor mirrored by the more robust HbO₂ outcomes. Most fNIRS studies to date have reported PFC-modulations during complex overground walking in older adults and PwPD.^11-15,54 Additionally, traditional treadmill walking has been shown to reduce PFC activity compared to overground walking, likely due to the implicit prompting provided by the belts.¹¹ A previous PET study examining SB-walking in young adults reported decreased PFC activity, which was interpreted as a possible shift of resources from the PFC to motor planning regions.¹⁰ Finally, a randomized controlled study from our group, investigating the long-term effects of a 4-week SB-training program in PwPD, showed that adaptive gait was improved and that these gains were unaffected by dual-tasking.⁹ Together, these results suggest that SB-walking relies more on reactive rather than cognitive control, possibly driven by cerebellar input^10,19 rather than by engaging the PFC. These results are corroborated by another recent study combining fNIRS with resting state fMRI showing that after 6 weeks of TB-training PFC activity in PwPD was decreased and gait performance improved. This was accompanied by an increase of Rs-MRI activity in the ROIs involved in the gait network including the cerebellum.⁵⁵

This study included an a priori powered sample size, and analyses were pre-registered. Nevertheless, some comparisons might have been underpowered since estimations were based on dual-task changes in the PFC.¹⁵ We preregistered several of the secondary analyses, including a channel-to-channel analysis, as averaging channels into ROIs could have concealed focal activations. Predominantly, the correlation analyses showed associations with specific channels and not with the full ROIs. Furthermore, none of the HHb results nor the exploratory analyses surpassed correction for multiple comparisons and should thus be interpreted with caution. However, reporting channel by channel changes allowed interpretation of the results in a more data-driven and anatomically specific way than is common in the fNIRS-field.^17,56 Although fNIRS has a greater spatial resolution compared to EEG, its temporal resolution and the maximal depth of measurement are limited.⁵⁷ As a result, we were unable to analyze sub-sections of the data in an event-related manner, such as focusing on the initial response to the perturbation. Even though fNIRS is quite resilient to movement artefacts, we included accelerometry and short separation channels as covariates. However, some residual noise could have contaminated the signals. For instance, HbO₂-results were not always mirrored in the HHb outcomes and vice versa. Therefore, a cautious interpretation remains indicated. Additionally, as we always started with the TB-walking condition to prevent carry-over effects, order effects cannot be discounted. Since the use of fNIRS is relatively new in the field of neurorehabilitation, studies on the validity, repeatability, responsiveness and minimal clinical important differences of fNIRS in PwPD and adults in general are scarce. Currently, these possible sources of error limit the interpretation of the clinical significance of our results and the causation between fNIRS and behavior should therefore be further explored. Future studies should consider a fully randomized protocol with sufficient wash-out periods, although this would imply longer experimental protocols compromising feasibility for patients. Finally, although the PD participants presented with gait problems, the sample was limited in size and was mostly drawn from the early to the mid disease stages (mostly HY 2), limiting the generalizability of the results to the PD population at large.

Implications and Conclusions

This study showed that PwPD have difficulty adapting to SB-perturbations even while ON-medication, justifying the utilization of SB-training to improve the adaptive capacity of gait in PD.^9,23,35 The present findings showed no consistent increases in PFC activity in response to SB-perturbations, suggesting that SB-walking does not impose additional cognitive load but rather facilitates a feedforward mode of control. Future studies need to determine whether SB-training is also effective and safe for more cognitively impaired PwPD and whether prolonged SB-training can reverse the abnormal cortical recruitment found.

In conclusion, PwPD showed more cortical activity than HC during treadmill walking in the sensorimotor regions previously associated with adaptive gait,¹⁰ and specific SB-related changes in SMA activity. These increments in cortical activity seemed largely non-functional as they did not normalize asymmetry. Therefore, greater difficulty with reaching a symmetrical gait during SB-walking may be explained by abnormal recruitment of these adaptive gait areas.

Supplemental Material

sj-docx-1-nnr-10.1177_15459683251329882 – Supplemental material for Cortical Activation During Split-Belt Treadmill Walking in People With Parkinson’s Disease and Healthy Controls

Supplemental material, sj-docx-1-nnr-10.1177_15459683251329882 for Cortical Activation During Split-Belt Treadmill Walking in People With Parkinson’s Disease and Healthy Controls by Femke Hulzinga, Paulo Henrique Silva Pelicioni, Nicholas D’Cruz, Veerle de Rond, Christopher McCrum, Pieter Ginis, Moran Gilat and Alice Nieuwboer in Neurorehabilitation and Neural Repair

Footnotes

Acknowledgements

We would like to thank all participants for their motivated and generous engagement in this study. Additionally, we would like to express our sincere gratitude to Peter Smolders for his technical lab support and Lander Boeckx for his help with data collection.

Author Contributions

Femke Hulzinga: Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Software, Validation; Visualization; and Writing—original draft. Paulo Pelicioni: Conceptualization; Data curation; Formal analysis; Methodology; Software; Supervision; Validation; and Writing—review & editing. Nicholas D’Cruz: Conceptualization; Data curation; Formal analysis; Methodology; Software; Validation; and Writing—review & editing. Veerle de Rond: Conceptualization; Data curation; Investigation; Methodology; Software; and Writing—review & editing. Christopher McCrum: Investigation and Writing—review & editing. Pieter Ginis: Conceptualization; Methodology; and Writing—review & editing. Moran Gilat: Conceptualization; Methodology; and Writing—review & editing. Alice Nieuwboer: Conceptualization; Data curation; Formal analysis; Funding acquisition; Methodology; Project administration; Resources; Supervision; Validation; and Writing—review & editing.

Data Availability

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by Research Foundation Flanders (FWO, project: 11B5421N and V440722N). FWO played no role in study design, data collection, analysis and interpretation of data, or the writing of this manuscript. Christopher McCrum was supported by a grant from NWO/ZonMw (Rubicon; 452020220).

ORCID iDs

Christopher McCrum

Femke Hulzinga

Paulo Henrique Silva Pelicioni

Nicholas D’Cruz

Supplementary material for this article is available on the Neurorehabilitation & Neural Repair website along with the online version of this article.

References

Klein

Bloem

BR.

Research in movement disorders in 2022: a new era of biomarker and treatment development. Lancet Neurol. 2023;22(1):17-19. doi:10.1016/S1474-4422(22)00494-X

Zhang

Gao

Tan

Chen

S Di

. Prevalence of freezing of gait in Parkinson’s disease: a systematic review and meta-analysis. J Neurol. 2021;268(11):4138-4150. doi:10.1007/s00415-021-10685-5

Bouça-Machado

Jalles

Guerreiro

, et al. Gait kinematic parameters in Parkinson’s disease: a systematic review. J Parkinsons Dis. 2020;10(3):843. doi:10.3233/JPD-201969

Mohammadi

Bruijn

Vervoort

, et al. Motor switching and motor adaptation deficits contribute to freezing of gait in Parkinson’s disease. Neurorehabil Neural Repair. 2015;29(2):132-142. doi:10.1177/1545968314545175

Spildooren

Vinken

Van Baekel

Nieuwboer

Turning problems and freezing of gait in Parkinson’s disease: a systematic review and meta-analysis. Disabil Rehabil. 2019;41(25):2994-3004. doi:10.1080/09638288.2018.1483429

Pelicioni

PHS

Menant

Latt

Lord

. Falls in Parkinson’s disease subtypes: risk factors, locations and circumstances. Int J Environ Res Public Health. 2019;16(12):2216.

Walton

Shine

Hall

, et al. The major impact of freezing of gait on quality of life in Parkinson’s disease. J Neurol. 2015;262(1):108-115. doi:10.1007/s00415-014-7524-3

Reisman

Block

Bastian

AJ.

Interlimb coordination during locomotion: what can be adapted and stored?

J Neurophysiol. 2005;94:2403-2415. doi:10.1152/jn.00089.2005

Hulzinga

Seuthe

D’Cruz

Ginis

Nieuwboer

Schlenstedt

Split-belt treadmill training to improve gait adaptation in Parkinson’s disease. Mov Disord. 2023;38(1):92-103. doi:10.1002/mds.29238

10.

Hinton

Thiel

Soucy

Bouyer

Paquette

Adjusting gait step-by-step: brain activation during split-belt treadmill walking. Neuroimage. 2019;202:116095.

11.

Thumm

Maidan

Brozgol

, et al. Treadmill walking reduces pre-frontal activation in patients with Parkinson’s disease. Gait Posture. 2018;62:384-387. doi:10.1016/J.GAITPOST.2018.03.041

12.

Stuart

Belluscio

Quinn

Mancini

Pre-frontal cortical activity during walking and turning is reliable and differentiates across young, older adults and people with Parkinson’s disease. Front Neurol. 2019;10:536.

13.

Maidan

Shustak

Sharon

, et al. Prefrontal cortex activation during obstacle negotiation: what’s the effect size and timing? Brain Cogn. 2018;122:45-51. doi:10.1016/j.bandc.2018.02.006

14.

Maidan

Nieuwhof

Bernad-Elazari

, et al. The role of the frontal lobe in complex walking among patients with Parkinson’s disease and healthy older adults: an fNIRS study. Neurorehabil Neural Repair. 2016;30(10):963-971. doi:10.1177/1545968316650426

15.

Nieuwhof

Reelick

Maidan

, et al. Measuring prefrontal cortical activity during dual task walking in patients with Parkinson’ s disease: feasibility of using a new portable fNIRS device. Pilot Feasibility Stud. 2016;2:59.

16.

Pelicioni

PHS

Lord

Okubo

Menant

. Cortical activation during gait adaptability in people with Parkinson’s disease. Gait Posture. 2022;91:247-253. doi:10.1016/J.GAITPOST.2021.10.038

17.

Menant

Maidan

Alcock

, et al. A consensus guide to using functional near-infrared spectroscopy in posture and gait research. Gait Posture. 2020;82:254-265. doi:10.1016/j.gaitpost.2020.09.012

18.

Burciu

Ofori

Shukla

, et al. Distinct patterns of brain activity in progressive supranuclear palsy and Parkinson’s disease. Mov Disord. 2015;30(9):1248-1258. doi:10.1002/mds.26294

19.

Morton

Bastian

AJ.

Behavioral/systems/cognitive cerebellar contributions to locomotor adaptations during splitbelt treadmill walking. J Neurosci. 2006;26(36):9107-9116. doi:10.1523/JNEUROSCI.2622-06.2006

20.

Hughes

Daniel

Kilford

Lees

AJ.

Accuracy of clinical diagnosis of idiopathic Parkinson’s disease: a clinico-pathological study of 100 cases. J Neurol Neurosurg Psychiatry. 1992;55(3):181-184. doi:10.1136/jnnp.55.3.181

21.

Hoehn

Yahr

MD.

Parkinsonism: onset, progression, and mortality. Neurology. 1967;17(5):427-442.

22.

van Melick

Meddeler

Hoogeboom

Nijhuis-van

der

Sanden

MWG

van Cingel

REH.

How to determine leg dominance: the agreement between self-reported and observed performance in healthy adults. PLoS ONE. 2017;12(12):e0189876. doi:10.1371/journal.pone.0189876

23.

Seuthe

D’Cruz

Ginis

, et al. The effect of one session split-belt treadmill training on gait adaptation in people with Parkinson’s disease and freezing of gait. Neurorehabil Neural Repair. 2020;34(10):954-963.

24.

Santosa

Zhai

Fishburn

Sparto

Huppert

TJ.

Quantitative comparison of correction techniques for removing systemic physiological signal in functional near-infrared spectroscopy studies. Neurophotonics. 2020;7(3):035009. doi:10.1117/1.nph.7.3.035009

25.

von Lühmann

Müller

Boas

Yücel

MA.

Improved physiological noise regression in fNIRS: a multimodal extension of the General Linear Model using temporally embedded Canonical Correlation Analysis. Neuroimage. 2020;208:116472.

26.

Zimeo Morais

Balardin

Sato

JR.

FNIRS Optodes’ Location Decider (fOLD): a toolbox for probe arrangement guided by brain regions-of-interest. Sci Rep. 2018;8:3341. doi:10.1038/s41598-018-21716-z

27.

Pelicioni

PHS

Lord

Okubo

Sturnieks

Menant

. People with Parkinson’s disease exhibit reduced cognitive and motor cortical activity when undertaking complex stepping tasks requiring inhibitory control. Neurorehabil Neural Repair. 2020;34(12):1088-1098. doi:10.1177/1545968320969943

28.

de Rond

Gilat

D’Cruz

Hulzinga

Orban de Xivry

J-J

Nieuwboer

. Test-retest reliability of functional near-infrared spectroscopy during a finger-tapping and postural task in healthy older adults. Neurophotonics. 2023;10(2):025010. doi:10.1117/1.nph.10.2.025010

29.

Letchuman

Donohoe

Neuroanatomy, Superior Sagittal Sinus. StatPearls Publishing; 2023. Accessed May 19, 2023. https://www.ncbi.nlm.nih.gov/books/NBK546615/

30.

Santosa

Zhai

Fishburn

Huppert

The NIRS brain AnalyzIR toolbox. Algorithms. 2018;11(73):a11050073. doi:10.3390/A11050073

31.

Barker

Aarabi

Huppert

TJ.

Autoregressive model based algorithm for correcting motion and serially correlated errors in fNIRS. Biomed Opt Express. 2013;4(8):1366-1379. doi:10.1364/boe.4.001366

32.

Dravida

Noah

Zhang

Hirsch

Comparison of oxyhemoglobin and deoxyhemoglobin signal reliability with and without global mean removal for digit manipulation motor tasks. Neurophotonics. 2017;5(1):011006. doi:10.1117/1.nph.5.1.011006

33.

Niu

Chen

Zhang

ZJ.

Reduced frontal activation during a working memory task in mild cognitive impairment: a non-invasive near-infrared spectroscopy study. CNS Neurosci Ther. 2013;19(2):125-131. doi:10.1111/CNS.12046

34.

Schaeffer

Yennu

Gandy

Tian

Liu

Park

An fNIRS investigation of associative recognition in the prefrontal cortex with a rapid event-related design. J Neurosci Methods. 2014;235:308-315. doi:10.1016/J.JNEUMETH.2014.07.011

35.

D’Cruz

Seuthe

Ginis

Hulzinga

Schlenstedt

Nieuwboer

Short-term effects of single-session split-belt treadmill training on dual-task performance in Parkinson’ s disease and healthy elderly. Front Neurol. 2020;11:560084. doi:10.3389/fneur.2020.560084

36.

Gilat

Bell

Ehgoetz Martens

, et al. Dopamine depletion impairs gait automaticity by altering cortico-striatal and cerebellar processing in Parkinson’s disease. Neuroimage. 2017;152:207-220. doi:10.1016/j.neuroimage.2017.02.073

37.

Hallett

Chan

Motor automaticity in Parkinson’s disease. Neurol Dis. 2015;82:226-234. doi:10.1016/j.nbd.2015.06.014

38.

Shields

Park

Srivanitchapoom

, et al. Probing the interaction of the ipsilateral posterior parietal cortex with the premotor cortex using a novel transcranial magnetic stimulation technique. Clin Neurophysiol. 2016;127(2):1475-1480. doi:10.1016/j.clinph.2015.06.031

39.

Barulli

Stern

Emerging concepts in cognitive reserve. Trends Cogn Sci. 2013;17(10):1-17. doi:10.1016/j.tics.2013.08.012.Efficiency

40.

Cassady

Ruitenberg

MFL

Reuter-Lorenz

Tommerdahl

Seidler

RD.

Neural dedifferentiation across the lifespan in the motor and somatosensory systems. Cereb Cortex. 2020;30(6):3704-3716. doi:10.1093/cercor/bhz336

41.

Kim

Criaud

Cho

, et al. Abnormal intrinsic brain functional network dynamics in Parkinson’s disease. Brain. 2017;140(11):2955-2967. doi:10.1093/brain/awx233

42.

Young

Parikh

Layne

CS.

The posterior parietal cortex is involved in gait adaptation: a bilateral transcranial direct current stimulation study. Front Hum Neurosci. 2020;14:581026. doi:10.3389/fnhum.2020.581026

43.

Drew

Marigold

DS.

Taking the next step: cortical contributions to the control of locomotion. Curr Opin Neurobiol. 2015;33:25-33.

44.

Spiers

Maguire

EA.

Thoughts, behaviour, and brain dynamics during navigation in the real world. Neuroimage. 2006;31:1826-1840. doi:10.1016/j.neuroimage.2006.01.037

45.

Vingerhoets

Contribution of the posterior parietal cortex in reaching, grasping, and using objects and tools. Front Psychol. 2014;5:151. doi:10.3389/fpsyg.2014.00151

46.

Macuga

Frey

SH.

Differential contributions of the superior and inferior parietal cortex to feedback versus feedforward control of tools. Neuroimage. 2014;92:36-45. doi:10.1016/j.neuroimage.2014.01.024.Differential

47.

Reuter-Lorenz

Cappell

KA.

Neurocognitive aging and the compensation hypothesis. Curr Dir Psychol Sci. 2008;17(3):177-182. doi:10.1111/j.1467-8721.2008.00570.x

48.

Liu

Tang

, et al. Greater prefrontal activation during sitting toe tapping predicts severer freezing of gait in Parkinson’s disease: an fNIRS study. Cereb Cortex. 2023;33(4):959-968. doi:10.1093/cercor/bhac114

49.

Chen

Pillemer

England

Izzetoglu

Mahoney

Holtzer

Neural correlates of obstacle negotiation in older adults: an fNIRS study. Gait Posture. 2017;58:130-135. doi:10.1016/j.gaitpost.2017.07.043

50.

Prodoehl

Spraker

Corcos

Comella

Vaillancourt

. Blood oxygenation level dependent activation in basal ganglia nuclei relates to specific symptoms in De Novo Parkinson’s disease. Mov Disord. 2010;25(13):2035-2043. doi:10.1002/mds.23360.Blood

51.

Heremans

Nieuwboer

Vercruysse

Freezing of gait in Parkinson’s disease: where are we now? Topical collection on movement disorders. Curr Neurol Neurosci Rep. 2013;13:350. doi:10.1007/s11910-013-0350-7

52.

Brugger

Wegener

Walch

, et al. Altered activation and connectivity of the supplementary motor cortex at motor initiation in Parkinson’s disease patients with freezing. Clin Neurophysiol. 2020;131:2171-2180. doi:10.1016/j.clinph.2020.05.023

53.

Wang

Hallett

Chan

Neural correlates of bimanual anti-phase and in-phase movements in Parkinson’s disease. Brain. 2010;133(8):2394-2409. doi:10.1093/brain/awq151

54.

Pelicioni

PHS

Tijsma

Lord

Menant

. Prefrontal cortical activation measured by fNIRS during walking: effects of age, disease and secondary task. PeerJ. 2019;7:e6833. doi:10.7717/PEERJ.6833

55.

Ding

Droby

Anwar

, et al. Treadmill training in Parkinson’s disease is underpinned by the interregional connectivity in cortical-subcortical network. NPJ Park Dis. 2022;8(1). doi:10.1038/s41531-022-00427-3

56.

Yücel

von Lühmann

Scholkmann

, et al. Best Practices for fNIRS publications. Neurophotonics. 2020;8(1):012101.

57.

Quaresima

Ferrari

Functional near-infrared spectroscopy (fNIRS) for assessing cerebral cortex function during human behavior in natural/social situations: a concise review. Organ Res Methods. 2019;22(1):46-68. doi:10.1177/1094428116658959

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.22 MB