https://orcid.org/0000-0002-1976-898X
Figures
Abstract
Cognitive impairments and gait disturbances are often concurrent in Parkinson disease (PD), but the extent to which specific cognitive deficits relate to gait abnormalities remains unclear, especially in early-stage PD. To address this gap, we conducted an observational correlational study to determine if cognitive performance in distinct domains is associated with gait kinematics in PD. This study included a cohort of 19 individuals with early-stage PD who underwent gait analysis with a three-dimensional marker-less motion capture system (Theia Markerless Inc., Kingston, ON, CA) during three conditions: single-task walking, cognitive dual-task walking, and the Timed Up and Go (TUG) test. Cognitive abilities were assessed using computerized tests (Creyos, Toronto, ON, CA) of working memory (WM), response inhibition (RI), and mental rotation (MR). We found that higher WM scores significantly correlated with better gait performance under dual-task conditions, including faster gait velocity (r = 0.6524, p = 0.0025), longer stride length (r = 0.6758, p = 0.0015), higher toe clearance (r = 0.5234, p = 0.0215), and greater hip range of motion (ROM) (r = 0.6803, p = 0.0014). Similarly, better MR ability was associated with longer stride length (r = 0.5178, p = 0.0231) and greater hip ROM (r = 0.4886, p = 0.0338) during dual-task walking. Furthermore, superior WM and MR performance were linked to faster TUG completion times, particularly during the second walking phase of the TUG (WM: r = −0.682, p = 0.0013; MR: r = −0.6755, p = 0.002). These results indicate that WM and MR may be linked to gait performance in PD, especially during cognitively challenging walking tasks. These findings may have clinical, rehabilitative, and neuroscientific utility for those involved in the study and management of PD.
Citation: Lander JJ, Moran MF, Alexanian HR (2025) Correlations of gait kinematics and cognitive skills in Parkinson disease. PLoS One 20(6): e0317389. https://doi.org/10.1371/journal.pone.0317389
Editor: Miray Budak, Rutgers University - Newark, UNITED STATES OF AMERICA
Received: January 17, 2025; Accepted: May 23, 2025; Published: June 13, 2025
Copyright: © 2025 Lander et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data supporting the findings in this study are openly available at the Harvard Dataverse, V1, Persistent ID: doi:10.7910/DVN/ATCQVI. https://doi.org/10.7910/DVN/ATCQVI.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Parkinson disease (PD) is the second most common neurodegenerative disease worldwide [1] with projected increased prevalence through the next two decades [2]. Current evidence suggests PD results from a confluence of genetic and environmental factors that cause abnormal α-synuclein protein accumulation, neuroinflammation, and cellular dysfunction [3,4]. The evidence of neuropathology, both centrally and peripherally, has generated diverging theories regarding the localization of disease onset [5,6] Early loss of dopaminergic neurons in the nigrostriatal pathway underlies the hallmark motor symptom of bradykinesia [7,8]. Neurodegenerative processes begin decades before diagnosis [9] with subtle symptoms developing in the prodromal stage and overt signs presenting in the manifestation stage. Parkinson disease is primarily characterized by motor dysfunction, but non-motor symptoms are equally present and cause considerable morbidity and disability. Autonomic dysregulation, sleep disorders, depression, and cognitive impairment are significant contributors to disease burden [4].
Gait dysfunction is a cardinal biomarker of PD. Bradykinesia, reduced joint range of motion (ROM), and postural instability are characteristic signs of PD. In comparison to healthy controls, individuals with PD display multiple spatiotemporal and kinematic differences such as slower gait velocity, shorter step and stride lengths, alterations of cadence, and asymmetrical arm movement [10,11]. Gait changes may become more pronounced in later stages of PD (i.e., festinating or freezing of gait) reflecting degeneration in central locomotor networks [12–15]. Healthy gait relies on motor automaticity, requiring minimal attentional output from frontal cortical areas. Motor automaticity is impaired in PD (review by Wu et al. [16]) as feedforward gait programs become deficient secondary to pathological nigrostriatal-thalamo-cortico loops. Individuals with PD compensate for these neural deficiencies by increasing activation of attentional fronto-striatal networks during locomotion [17]. However, the additional cognitive load, particularly during dual-task (DT) gait, can fatigue fronto-striatal and forebrain cholinergic regions [17,18]. These neurodegenerative processes begin early in the silent stage of PD and are progressive. Although non-specific, gait is a practical clinical biomarker for detection and monitoring of PD [19–21].
The Timed Up and Go (TUG) task is a widely used tool for musculoskeletal and neurological examination, including for individuals with PD [22,23]. The components of vertical transitions (i.e., sit-to stand, stand-to-sit), walking, turning, and spatial planning involve both motor and cognitive efforts. Phase analysis of the TUG task has revealed locomotor deficits that are not evident in routine walking [24]. For example, freezing-of-gait may be expressed during the turning sub-phase of TUG but not in other sub-phases [25].
The level of cognitive impairment (CI) is heterogenous in individuals with PD. In their frequently cited study, Buter et al. [26] reported 80% of patients with PD will develop dementia over the course of disease progression. However, recent authors [27] have argued the incidence of dementia in PD is significantly lower (~50% over 15 years). Nevertheless, CI is recognized as a prevalent non-motor concomitant that may occur in early stages of PD [28] and progress to dementia in susceptible individuals. Cognitive impairments of attention, language, and memory are associated with a reduction in daily activities independent of motor disability [29].
Working memory (WM) is a cognitive task involving multiple neural substrates and circuits. WM localizes most notably to the dorsolateral prefrontal cortex, frontal eye fields, and inferior parietal lobe with direct connections to the caudate [30]. Introductory work by Owen et al. [31] and more recently Hattori et al. [30] have explored the cognitive demand of WM in PD, revealing that patients without overt cognitive impairment require increased neural activation to accomplish simple WM tasks. Furthermore, patients with more advanced cognitive deficits often cannot meet the neural demand as task difficulty increases.
Response inhibition (RI) is a cognitive-motor task often measured with Stroop tasks or go-no-go paradigms. Neural correlates of RI are predominantly frontal and include the prefrontal cortex, anterior cingulate cortex, and the subthalamic nucleus of basal ganglia [32,33]. These structures are inherent to the fronto-striatal-thalamo-cortical circuits susceptible to neurodegeneration and denervation present in PD. Deficiencies of the noradrenergic locus coeruleus is associated with CI and correlates with poor RI performance by individuals with PD [34,35]. Response inhibition has been used as an evaluation tool in the PD population as a surrogate biomarker of psychological and/or motor inhibition (see review by MacDonald et al. [36]).
Mental rotation (MR) is a visuospatial cognitive task localizing to posterior cortical regions with parallel anterior feedforward contribution. Superior and inferior parietal lobes, visual cortices, ventral and dorsal visual pathways are the principal neural substrates [37,38]. Importantly, there is also pre-motor and cerebellar involvement as MR requires motor planning [37]. Visuospatial dysfunction is evident in PD [39] and MR is a unique challenge due to its pre-motor planning and imagery requirements [40]. Furthermore, MR deficits in PD may indicate neurodegeneration beyond dopaminergic networks [41].
Gait and cognition share neural substrates particularly in nigro-striatal-thalamo-cortical circuits. In PD, motor-cognitive skills create demands that may overwhelm susceptible neural circuitry resulting in motor faults, cognitive dysfunction, or both. Although gait disturbances during cognitive DT efforts are well reported in the PD population, very few studies have reported the relationship between specific cognitive skills and gait performance in PD. Penko, et al. [42,43] found that individuals with PD walked slower with decreased stride length and joint ROM during cognitive DT, particularly during a WM challenge. Harrie et al. [44] reported TUG phase times under DT conditions correlated strongly with visuospatial and processing skills, more so than executive functions. Amboni, et al. [45] performed a correlation analysis of neuropsychological tests and gait, revealing strong correlation of global cognition with multiple gait parameters (i.e., cadence, velocity, support time). However, visuospatial orientation was the only cognitive domain significantly correlated with a specific gait parameter (velocity).
In this study, we explore correlations between distinct cognitive domains and detailed gait kinematics in individuals with PD. Using marker-less motion capture technology and validated neuropsychological tasks, we aim to reveal the relationships of WM, RI, and MR with spatiotemporal gait parameters during ST walking, DT walking, and TUG phase times. Clarifying these relationships can improve our understanding of specific motor-cognitive interactions in PD and guide more precise clinical and rehabilitative management strategies. Furthermore, these findings may be useful for neuroscience researchers investigating PD circuit pathophysiology.
Materials and methods
All study procedures were approved by the University’s Institutional Review Board and conducted in accordance with the Declaration of Helsinki. All participants were required to have a diagnosis of PD and be able to ambulate without assistance. Additionally, all participants met the following criteria: (i) Free of any history of ancillary orthopedic conditions that precluded walking (e.g., acute injury, recent surgery) (ii) Free of any other neurological disorders with concomitant CI (e.g., stroke), (iii) Maintenance of normal medication schedule. There were no exclusion criteria related to sex or baseline cognitive status. Twenty-one participants (15 females, 6 males; 65.7 ± 5.7 yo), were enrolled from a referral sample of local PD support group members and a convenience sample from a private medical practice. All participants met the above criteria and volunteered for this investigation.
Initial screening and assessment
After participants volunteered for study, a focused health history was taken to obtain demographic information and confirm candidacy. Motor function was evaluated using part three (motor) of the Movement Disorder Society-Unified Parkinson’s Disease Rating Scale (MDS-UPDRS) [46]; Patient Health Questionnaire (PHQ-9) [47] and the Montreal Cognitive Assessment (MoCA) were completed, respectively. Participants maintained their normally prescribed medication schedule. As standard, levodopa equivalent daily dose (LEDD) was calculated for each participant [48]. For all cognitive assessments, individuals used corrective lenses as needed. No participants reported color visual deficits. All assessments were conducted by one of the study’s primary investigators (JJL).
Computer neuropsychological assessment (CNA)
Participants completed three CNA domain tasks from Creyos web-based software (Toronto, ON, CA) in a controlled environment. Participants were seated in a quiet setting, utilized a tablet computer (Apple iPad, Cupertino, CA) and were provided verbal instructions. Data was collected via an online portal (www.creyos.com). Creyos software offers easy to use gamified modes that are engaging to the user. Along with verbal instructions, Creyos software provided individualized instructions prior to each assessment with a series of practice trials. No instructions or assistance were provided during the assessments. Domain tasks of working memory (‘Token Search’), mental rotation/visuospatial processing (‘Rotations’), and attention/response inhibition Stroop task (‘Double Trouble’) were assessed. The WM task involved remembering the location of a hidden token in a group of boxes. Participants were challenged to find the token by tapping individual boxes and remembering which boxes were already selected. Choosing the same box twice in a round resulted in a penalty. Each successful round increased the difficulty in the following round. Three total penalties ended the game. To assess visuospatial processing, a timed task (90 seconds) of MR was used. Participants were charged with mentally rotating a grid of shapes to discern if it was identical to a corresponding grid. Participants selected if the grids were a “match” or “no match”. The game progressed in difficulty with each successful round. Finally, attention/response inhibition was assessed via a timed Stroop task (90 seconds). The Stroop task included word-color pairs of varied congruencies (congruent-congruent, incongruent-congruent, incongruent-incongruent). Participants were charged with selecting the appropriate word-color match. These cognitive tasks are commonly used with various demographics and have previously assessed cognition in PD patients [31,49,50]. Creyos software has norm-based age/sex scoring, has adequate concurrent validity when compared to in-person neuropsychological assessment, marginal-to-low test-retest reliability depending on specific test item, and good usability among older adults with cognitive decline [51,52].
Experimental protocol
Participants performed two TUG assessments and four 15-meter overground walking trials with a single task (ST) and dual task (DT) condition. Participants wore their preferred closed toe walking footwear and self-selected attire. For the TUG assessment, participants were instructed to stand from a seated position, walk three meters before turning 180o around a mark taped to the floor, walk back to chair, turn, and sit down. The TUG assessment has been previously reported as having high test-retest reliability in individuals with PD [23]. Participants then performed two ST walking trials where they were instructed to walk at a self-selected pace for 15 meters. Sport cones marked the beginning and end of the 15-meter walkway. To conclude the protocol, participants performed two DT walking trials, when they were given a word to spell backward while they walked at their self-selected pace. Backward spelling was used in lieu of the common DT of serial 7 subtraction to avoid potential learning effects from previous exposure as the subtraction task is often used in the clinical setting. Hassan, et al. [53] have reported backward spelling as an effective dual-task cognitive load for older adults. One five-letter word for each trial (1 -“Radio”, 2 -“Baker”) was provided to participants within the first two steps that began the trial. No other verbal instructions were provided, a practice trial was afforded for each ST test condition (TUG, overground walking), and adequate recovery (~60 seconds) was granted between all trials.
Biomechanical assessment
An eight-camera video system (Sony RXO II; Sony Corporation; Minato, Japan) recorded trials at 120 frames per second. Synchronized videos were calibrated to estimate a 19-segment kinematic model with a commercial marker-less (ML) motion capture software package (v2023.1.03161 patch 11; Theia Markerless Inc., Kingston, ON, CA). The ML motion capture system utilizes deep learning neural networks to estimate three-dimensional pose estimates (Fig 1) with a more detailed explanation of pose determination provided at Kanko et al. [54]. Previous investigations have reported good-to-excellent agreement for the determination of spatiotemporal metrics of walking gait (e.g., velocity, stride length) using this methodology when compared to a pressure-sensitive mat [54,55]. Additionally, sagittal plane joint kinematics were reported to have excellent agreement (RMS < 4o degrees) with marker-based motion capture [56].
Kinematic data (4x4 segment rotation matrices) and locations of virtual bilateral heel markers were exported and subsequently processed with Visual3D Professional (C-Motion Inc.; Germantown, MD, USA; v2023.12.1). Data was filtered with a GCVSPL filter with 12 Hz cutoff frequency GCVSPL filter [57]. A Visual3D segmental model with physical properties (segmental mass, center of mass, center of mass location) derived from previous work [58–60] was used to estimate and track full body center of mass (COM) location.
Data processing
For both TUG trials, the COM’s x-coordinate (direction of travel) and z-coordinate (superior-inferior) respective positions (Table 1) were utilized to determine the relevant phase times for each trial. Component TUG times were measured for the following phases: (1) sit-to-stand, (2) gait initiation, (3) walk 1, (4) turn, (5) walk 2, (6) slow down & turn, and (7) sit down. These component tasks are similar to a past report that partitioned the TUG into requisite phases [61]. Similarly, COM location was also tracked during a gait initiation task within a group of individuals with PD, with the same ML mocap system used in the current study and deemed to efficiently detect motor performance [62].
For all overground walking trials, a 3.3-meter zone within the center of the calibrated three-dimensional volume was used to demarcate the beginning and end of trial respectively. Location of COM was used to determine these respective endpoints with walking velocity computed for this interval. All participants had > 5m walking prior to this collection zone to avoid gait initiation phase and allowing for at least two strides within capture period. Stance phase event markers (foot strike, foot off) were determined via a coordinate-based algorithm of the heel and toe marker relative to the body’s COM. This is a similar method to a previous gait investigation using the same ML mocap system [55]. Based on these gait events, stride length (SL) and double support (DS) time were computed and averaged for each trial. Additionally, minimum toe clearance (TC) during swing phase was determined bilaterally for each trial. Hip and shoulder flexion-extension joint angles were computed in Visual3D utilizing an XYZ Cardan rotation sequence consistent with a standardized joint coordinate system [63]. Bilateral hip and shoulder ROM during the gait cycle were determined bilaterally for each trial. All TUG phase times, and biomechanical walking gait metrics (SL, DS, TC, Hip ROM, and shoulder ROM) were exported to a custom Python (ver. 3.12.4) script for further data analysis and statistical processing. As arm swing asymmetry during gait has been reported to be increased in PD [19], the symmetry angle [64] was computed between bilateral shoulder ROM metrics. For statistical processing, only the toe clearance and hip ROM from the impaired side was used for analysis. For any participant with bilateral impairments (n = 1) or no impairments (n = 1), the average across sides was used for analysis.
Statistical analysis
All dependent variables were first assessed for normality with the Shapiro-Wilk test to determine the appropriate correlation test. Pairwise correlations were conducted between (i) all cognitive task metrics and TUG metrics, and (ii) all cognitive task metrics and walking metrics using either Pearson or Spearman correlation depending on the results of normality test. Both the normality and correlation tests were computed utilizing the SciPy library [65] within Python. Finally, an ordinary least squares (OLS) regression and a multivariate analysis of variance (MANOVA) were performed to evaluate the overall effect of cognitive task scores on the TUG and walking metrics, respectively. These analyses were conducted using the statsmodels library [66] in Python.
Results
Data from nineteen participants was used for subsequent analysis with two participants excluded for either not completing all sessions of protocol or methodological issues related to video file storage. The mean age was 65.7 ± 5.7 years. Participants had mild motor symptoms on average with a mean UPDRS Part III score 11.4 ± 8.9, indicating early-stage PD. Education level was high with a median of 16 years completed and 43% of participants achieving graduate degrees. The median MoCA score was 27 with one participant meeting the criteria suggestive of mild cognitive impairment. All participants maintained their PD medication schedule (two participants were non-medicated). The mean LEDD 446 ± 279 mg (range 0–1080 mg). Complete descriptive statistics of participant demographics (Table 2) and dependent variables were computed with accompanying normality checks (Table 3). All variables were normally distributed (p > 0.05) except for RI (p = 0.031), TUG sit down phase time (p = 0.003), shoulder asymmetry ST (p = 0.005), double support DT (p = 0.026) and shoulder asymmetry DT (p = 0.004). Additionally, the variation inflation factor (VIF) for all cognitive skill task scores were < 2.7, indicating that multicollinearity of variables would not influence regression findings.
Timed up & go test (TUG)
The sit-to-stand phase time was negatively correlated with RI (r = −0.4704, p = 0.042), WM (r = −0.613, p = 0.005) and MR (r = −0.5401, p = 0.017) cognitive skills respectively (Table 4, Fig 2). The Walk 1 phase time was negatively correlated with
RI (r = −0.4861, p = 0.035), WM (r = −0.5186, p = 0.023), and MR (r = −0.6458, p = 0.003) cognitive skills respectively. The turning phase time was negatively correlated with RI (r = −0.4366, p = 0.062), WM (r = −0.5592, p = 0.013), and MR (r = −0.6377, p = 0.003) cognitive skills respectively. The Walk 2 phase time was negatively correlated with RI (r = −0.5114, p = 0.025), WM (r = −0.6820, p = 0.001), and MR (r = −0.6755, p = 0.002) cognitive skills respectively. The slow down & turn phase time was negatively correlated with WM (r = −0.4619, p = 0.047) and MR (r = −0.4989, p = 0.030), but showed no significant correlation with RI (r = −0.2148, p = 0.377). Both the gait initiation and sit down phases had no significant correlation with any of the cognitive skills.
The OLS regression models explained significant portions of variance for each TUG phase: Sit to Stand (44.4%), Gait Initiation (7.1%), Walk 1 (45.8%), Turn Around (48.8%), Walk 2 (61.9%), Slow Down & Turn (41.1%), and Sit Down (19.8%) (Table 5). All TUG phases exhibited significant intercepts (p < 0.05) in the OLS regression models, indicating substantial baseline phase times independent of cognitive skill metrics. Mental rotation performance was significantly associated with Walk 1 phase time (B = −0.005, p = 0.048), Turn phase time (B = −0.012, p = 0.039), and Walk 2 phase time (B = −0.006, p = 0.027) respectively. Additionally, WM was significantly associated with Walk 2 phase time (B = −0.184, p = 0.029).
The MANOVA results indicated that cognitive scores were collectively associated with the TUG performance (Table 6). The overall model was significant (Wilks’ lambda = 0.0725, F(7, 9) = 16.4564, p = 0.0002). Specifically, WM was significantly associated with the TUG phase times (Wilks’ lambda = 0.2485, F(7, 9) = 3.8886, p = 0.0313), whereas RI and MR did not reach significance (p > 0.05).
Single task (ST) walking
During the ST walking condition, WM demonstrated a significant positive correlation with velocity (r = 0.4619, p = 0.0465), stride length (r = 0.5707, p = 0.0107), toe clearance (r = 0.5517, p = 0.0143), and hip ROM (r = 0.5715, p = 0.0106). The only other significant correlation during the ST walking condition was between MR and stride length (r = 0.4610, p = 0.0470) (Table 7, Fig 3).
The OLS regression models explained portions of variance for each walking metric under single-task conditions: Velocity (25.4%), Stride Length (43.9%), Double Support (3.2%), Toe Clearance (34.2%), Hip ROM (34.2%), and Shoulder Asymmetry (23.6%) (Table 8). Most walking metrics exhibited significant intercepts (p < 0.05) in the OLS regression models, indicating substantial baseline phase times independent of cognitive skill metrics. However, the intercepts for toe clearance (p = 0.684) and shoulder asymmetry (p = 0.166) were not significant. Working memory performance was significantly associated with stride length (B = 0.1119, p = 0.027) and toe clearance (B = 0.0064, p = 0.028). Additionally, WM was associated with Hip ROM (B = 3.0743, p = 0.063).
The MANOVA results indicated that cognitive scores were collectively associated with the walking performance in the single-task condition (Table 9). The overall model was significant (Wilks’ lambda = 0.0339, F(6, 10) = 47.5500, p < 0.0001). Specifically, WM was significantly associated with the walking metrics (Wilks’ lambda = 0.3420, F(6, 10) = 3.2070, p = 0.0504), whereas RI and MR did not reach significance (p > 0.05).
Dual task (DT) walking condition
During the DT walking condition, RI demonstrated a significant positive correlation with hip ROM (r = 0.5632, p = 0.0120). Working memory demonstrated a significant positive correlation with velocity (r = 0.6524, p = 0.0025), stride length (r = 0.6758, p = 0.0015), toe clearance (r = 0.5234, p = 0.0215), and hip ROM (r = 0.6803, p = 0.0014). Additionally, MR showed significant positive correlations with stride length (r = 0.5178, p = 0.0231), hip ROM (r = 0.4886, p = 0.0338) and shoulder asymmetry (r = 0.5894, p = 0.0079) (Table 10, Fig 3
The OLS regression models explained significant portions of variance for each walking metric under dual-task conditions: velocity (43.9%), stride length (58.9%), double support (11.6%), toe clearance (28.7%), hip ROM (48.4%), and shoulder asymmetry (44.4%) (Table 11). Most walking metrics exhibited significant intercepts (p < 0.05) in the OLS regression models, indicating substantial baseline phase times independent of cognitive skill metrics. However, the intercepts for velocity (p = 0.285), toe clearance (p = 0.63), and shoulder asymmetry (p = 0.535) were not significant. Working memory performance was significantly associated with velocity (B = 0.1769, p = 0.021) and stride length (B = 0.1582, p = 0.004). Mental rotations were significantly associated with shoulder asymmetry (B = 0.3349, p = 0.006).
The MANOVA results indicated that cognitive scores were collectively associated with the walking performance in the dual-task condition (Table 12). The overall model was significant (Wilks’ lambda = 0.0310, F(6, 10) = 52.0835, p < 0.0001). However, none of the individual cognitive scores (RI, WM, MR) reached significance (p > 0.05) in predicting the walking metrics under dual-task conditions.
Discussion
This study investigated the relationships between specific cognitive skills and gait performance in individuals with early-stage PD. Although global cognitive deficiency is observed in CI and dementia, certain cognitive skills may be more susceptible to decline depending on environmental factors and individual phenotype. In non-PD populations, Jayakody et al. [67] and Toots et al. [68] both reported gait velocity to have the strongest relationship with executive and visuospatial functions in comparison to other cognitive domains. In individuals with PD and mild CI, Amboni et al. [45] found visuospatial orientation to be strongly correlated to gait velocity. Recently, Piet et al. [69] found gait speed and step variability during DT walking to predict performance on executive function in older patients with PD. Consistent with this prior research, we found that motor-cognitive interference is evident even in high-functioning PD patients. Our results extend previous findings identifying which cognitive domains are most strongly associated with gait impairments. We report WM and MR abilities have strong correlations with locomotor behavior, especially under DT conditions and during TUG. Our observations reveal that in this PD population, specific cognitive domains are highly correlated to fine kinematic details such as velocity, stride length, hip and ankle ROM, and turning performance. All kinematic variables during ST gait were positively associated with WM. During normal walking, lower WM was associated with slower velocity, shorter stride length, reduced hip ROM, and poor toe clearance. These relationships were pronounced during DT, suggesting reduced neural compensation to the increased cognitive load. Participants with stronger WM are likely to have more intact or compensatory fronto-striatal function, enabling them to better maintain feedforward gait processes or devote additional attention to gait when needed. Phase times of TUG were negatively correlated with WM. Specifically, Sit-to-Stand, Turn Around, and Walk 2 phases were the strongest associations. WM is a complex task requiring activation of several cortical regions including frontal-associative and temporal regions with neostriatal reciprocal circuitry. WM also involves attention and goal-oriented behavior partially mediated by frontal eye-fields and subcortical visuomotor substrates [70]. The engagement of visuomotor neural circuits enables a mental image of WM problem solving tasks. Visuomotor dysfunction is common in PD and associated with cognitive [71] and gait [72,73] performance. The parallel processing necessary for healthy WM increases metabolic demand on fronto-striatal neural networks. Hence, the neural activation patterns in PD during WM tasks are heterogenous. Hyperactivation is observed with difficult WM tasks in cognitively normal PD patients while there is network hypoactivity in those with CI [30]. As such, individuals with PD that have poor WM may also struggle with locomotion, particularly under DT, because they cannot marshal the neural resources necessary to compensate for motor automaticity deficit. According to our data, WM has a strong correlation to locomotion in PD and may be a prime indicator of overall neural integrity and health.
The correlations of MR and gait kinematics were weak-to-moderate for ST and DT conditions. Specifically, SL and hip ROM were positively associated and approached clinical significance for fall risk [74,75]. Mental rotations and TUG were highly correlated with Turn Around and Walk 2 phases revealing strong negative associations. Turning around an object while walking challenges balance, limb placement, visuospatial awareness, and motor planning. Both visuospatial awareness and motor planning are components of MR and involve frontal, posterior, and subcortical structures [37]. Hence, in addition to motor circuitry dysfunction [25,76], visuospatial substrates of MR may be factors in the etiology of turning difficulties observed in PD.
In this study, RI showed weak correlation to gait kinematics and TUG performance. Although RI was not as strongly correlated to locomotion as MR or WM, the findings may be a subtle expression of global cognitive deficiency often observed with neurodegenerative disease and aging. Various RI tasks may provide different neural substrate challenges based on task context such as impulse control, error correction, and conflict resolution [32,77]. The Stroop color-word task was selected to challenge frontal attentional and conflict resolution as opposed to a “go-no-go” task which has an inherently stronger motor response component. Using a RI task with minimal motor demand decreased the risk of individual motor dysfunction influencing RI performance; thus, improving the specificity of cognitive assessment for gait correlation analysis. Manza et al. [78] reported RI deficiency in older individuals with more severe PD burden compared to those newly diagnosed. Collectively, our study participants were in the early diagnostic stage, highly functional with low UPDRS motor scores, and healthy MoCA screenings. As such, it is possible that our cohort did not reach a clinically deficient threshold of RI to reveal a stronger relationship.
The interaction of motor and cognitive deficiency, referred to as the motoric-cognitive risk syndrome [79], begins with subtle signs and progresses during aging and/or neurodegenerative processes. The initial sign of motoric-cognitive risk syndrome is bradykinetic gait often with subjective cognitive complaints. Purks et al. [80] recently reported that cognition is a common concern in PD with >30% of patients reporting a cognitive symptom as the most bothersome. These subjective symptoms are often missed by cognitive screening tools, while more sensitive neuropsychological batteries are not routinely administered in early stages of PD. Typically, PD has been a clinical diagnosis based on overt signs of rigidity, tremor, and bradykinesia along with physiological symptoms (i.e., hyposmia, constipation). Patients presenting with uncertain or atypical clinical signs may receive special imaging (i.e., DaTscan) or, more recently, the newly developed alpha-synuclein seed analysis [81]. These advanced procedures improve the specificity and sensitivity of diagnosis but do not provide means to measure physical function nor predict future impairments. To this end, gait assessment has been proposed as a vital biomarker for detection and measure of progression in PD [20,21]. Moreover, early gait analysis may be a predictor of cognitive impairments during aging and/or disease advancement [69,82]. Thus, given the consistent data supporting the use of gait as an assessment tool in PD, we suggest gait is a vital component of the neurological evaluation in all stages of disease. New gait analysis technologies offer a high yield, low risk tool feasible for clinical use in the neurological setting [55,83]. Further, global and specific cognitive assessments in the early diagnostic stage may help reveal impairment risk. Due to the inherent neuropathological processes, PD has a strong predictive incidence of motor-cognitive risk syndrome [79] suggesting detection of both gait and cognitive deficiencies are critical for comprehensive patient care during all phases of disease management. Our findings suggest these deficiencies may have specific relationships that may influence clinician strategies for management and rehabilitation.
Limitations
We acknowledge several limitations of this study. The small sample (68% female) limits statistical power and may not be generalizable. A larger sample may have revealed significant and/or stronger correlations between cognitive domains (i.e., RI) and gait parameters. We also acknowledge the lack of a control group which may limit internal validity. However, the evidence is consistent that motor and cognitive deficiencies are accelerated in PD compared to healthy aging controls. Our findings suggest there may be specific cognitive domain-gait parameter interactions that likely follow this same pattern. Participants were enrolled from support groups and a private practice, and as such, may have been highly motivated, altruistic, and interested in learning about their condition [84]. Despite these limitations, our results are consistent with previous research and provide additional considerations on the presence of heterogenous cognitive domains with correlation to gait dysfunction in early PD. Our study participants were (a) highly educated, (b) relatively newly diagnosed, in early Hoehn and Yahr stage, with low UPDRS scores, (c) did not meet screening criteria for cognitive impairment, (d) did not report depression, and (e) maintained their medication schedule. Nevertheless, we report distinct correlations between gait and cognitive function in these individuals. Still, given the limitations stated above, these findings are preliminary and should be interpreted with caution. Future researchers may wish to compare different types of DT (e.g., cognitive vs. motor dual tasks, different cognitive domains) to observe which best reveal gait vulnerabilities. Additionally, intervention studies could explore whether targeted therapies such as DT training and/or customized cognitive rehabilitation (i.e., WM tasks) can enhance both cognition and gait, thereby potentially slowing the progression of the motoric-cognitive risk syndrome in PD. Future investigations may also wish to measure motor-cognitive correlations longitudinally in a more diverse, controlled sample.
Conclusion
Our findings provide insights into the motor-cognitive interactions in a cohort of individuals with PD. Motor and cognitive deficits may progress concurrently over the course of disease progression. Specifically, our data suggests WM and MR are strong cognitive correlates to locomotor function in PD, particularly to velocity, stride length, and hip ROM. Understanding these relationships can potentially help researchers and clinicians develop effective strategies focused on maintaining mobility and cognitive function in individuals with PD. For example, specific DT interventions using WM tasks and MR/visuospatial challenges may enhance gait training and rehabilitative sessions. Further, these findings may encourage new study of neural substrate pathology, mapping, and PD topology to help further our understanding of PD pathophysiology.
References
- 1. Ben-Shlomo Y, Darweesh S, Llibre-Guerra J, Marras C, San Luciano M, Tanner C. The epidemiology of Parkinson’s disease. Lancet. 2024;403(10423):283–92. pmid:38245248
- 2. Dorsey ER, Sherer T, Okun MS, Bloem BR. The Emerging Evidence of the Parkinson Pandemic. J Parkinsons Dis. 2018;8(s1):S3–8. pmid:30584159
- 3. Hatano T, Okuzumi A, Matsumoto G, Tsunemi T, Hattori N. α-Synuclein: A Promising Biomarker for Parkinson’s Disease and Related Disorders. J Mov Disord. 2024;17(2):127–37. pmid:38589016
- 4. Blesa J, Foffani G, Dehay B, Bezard E, Obeso JA. Motor and non-motor circuit disturbances in early Parkinson disease: which happens first?. Nat Rev Neurosci. 2022;23(2):115–28. pmid:34907352
- 5. Braak H, Del Tredici K, Rüb U, de Vos RAI, Jansen Steur ENH, Braak E. Staging of brain pathology related to sporadic Parkinson’s disease. Neurobiol Aging. 2003;24(2):197–211. pmid:12498954
- 6. Horsager J, Knudsen K, Sommerauer M. Clinical and imaging evidence of brain-first and body-first Parkinson’s disease. Neurobiol Dis. 2022;164:105626. pmid:35031485
- 7. Schwarz ST, Xing Y, Tomar P, Bajaj N, Auer DP. In Vivo Assessment of Brainstem Depigmentation in Parkinson Disease: Potential as a Severity Marker for Multicenter Studies. Radiology. 2017;283(3):789–98. pmid:27820685
- 8. Bloem BR, Okun MS, Klein C. Parkinson’s disease. Lancet. 2021;397(10291):2284–303. pmid:33848468
- 9. Bezard E, Gross CE, Brotchie JM. Presymptomatic compensation in Parkinson’s disease is not dopamine-mediated. Trends Neurosci. 2003;26(4):215–21. pmid:12689773
- 10. Bonacina D, Tosatto D, Ugolini A, Rossi F, Corno I, Pellicciari L, et al. Spatiotemporal, kinematic and kinetic gait characteristics in Parkinson’s disease compared to healthy individuals: A systematic review with meta-analysis. Clin Biomech (Bristol). 2024;120:106359. pmid:39426345
- 11. Bouça-Machado R, Jalles C, Guerreiro D, Pona-Ferreira F, Branco D, Guerreiro T, et al. Gait Kinematic Parameters in Parkinson’s Disease: A Systematic Review. J Parkinsons Dis. 2020;10(3):843–53. pmid:32417796
- 12. Takakusaki K. Functional Neuroanatomy for Posture and Gait Control. J Mov Disord. 2017;10(1):1–17. pmid:28122432
- 13. Karachi C, Grabli D, Bernard FA, Tandé D, Wattiez N, Belaid H, et al. Cholinergic mesencephalic neurons are involved in gait and postural disorders in Parkinson disease. J Clin Invest. 2010;120(8):2745–54. pmid:20628197
- 14. Takakusaki K, Habaguchi T, Ohtinata-Sugimoto J, Saitoh K, Sakamoto T. Basal ganglia efferents to the brainstem centers controlling postural muscle tone and locomotion: a new concept for understanding motor disorders in basal ganglia dysfunction. Neuroscience. 2003;119(1):293–308. pmid:12763089
- 15. Joza S, Camicioli R, Martin WRW, Wieler M, Gee M, Ba F. Pedunculopontine Nucleus Dysconnectivity Correlates With Gait Impairment in Parkinson’s Disease: An Exploratory Study. Front Aging Neurosci. 2022;14:874692. pmid:35875799
- 16. Wu T, Hallett M, Chan P. Motor automaticity in Parkinson’s disease. Neurobiol Dis. 2015;82:226–34. pmid:26102020
- 17. Nishida A, Shima A, Kambe D, Furukawa K, Sakamaki-Tsukita H, Yoshimura K, et al. Frontoparietal-Striatal Network and Nucleus Basalis Modulation in Patients With Parkinson Disease and Gait Disturbance. Neurology. 2024;103(3):e209606. pmid:38976821
- 18. Pereira JB, Hall S, Jalakas M, Grothe MJ, Strandberg O, Stomrud E, et al. Longitudinal degeneration of the basal forebrain predicts subsequent dementia in Parkinson’s disease. Neurobiol Dis. 2020;139:104831. pmid:32145376
- 19. Koh S-B, Park Y-M, Kim M-J, Kim W-S. Influences of elbow, shoulder, trunk motion and temporospatial parameters on arm swing asymmetry of Parkinson’s disease during walking. Hum Mov Sci. 2019;68:102527. pmid:31689575
- 20. Guo Y, Yang J, Liu Y, Chen X, Yang G-Z. Detection and assessment of Parkinson’s disease based on gait analysis: A survey. Front Aging Neurosci. 2022;14:916971. pmid:35992585
- 21. Pistacchi M, Gioulis M, Sanson F, De Giovannini E, Filippi G, Rossetto F, et al. Gait analysis and clinical correlations in early Parkinson’s disease. Funct Neurol. 2017;32(1):28–34. pmid:28380321
- 22. Son M, Youm C, Cheon S, Kim J, Lee M, Kim Y, et al. Evaluation of the turning characteristics according to the severity of Parkinson disease during the timed up and go test. Aging Clin Exp Res. 2017;29(6):1191–9. pmid:28220396
- 23. Morris S, Morris ME, Iansek R. Reliability of measurements obtained with the Timed “Up & Go” test in people with Parkinson disease. Phys Ther. 2001;81(2):810–8. pmid:11175678
- 24. Van Uem JMT, Walgaard S, Ainsworth E, Hasmann SE, Heger T, Nussbaum S, et al. Quantitative Timed-Up-and-Go Parameters in Relation to Cognitive Parameters and Health-Related Quality of Life in Mild-to-Moderate Parkinson’s Disease. PLoS One. 2016;11(4):e0151997. pmid:27055262
- 25. Son M, Cheon S-M, Youm C, Kim JW. Turning reveals the characteristics of gait freezing better than walking forward and backward in Parkinson’s disease. Gait Posture. 2022;94:131–7. pmid:35306381
- 26. Buter TC, van den Hout A, Matthews FE, Larsen JP, Brayne C, Aarsland D. Dementia and survival in Parkinson disease: a 12-year population study. Neurology. 2008;70(13):1017–22. pmid:18362281
- 27. Gallagher J, Gochanour C, Caspell-Garcia C, Dobkin RD, Aarsland D, Alcalay RN, et al. Long-Term Dementia Risk in Parkinson Disease. Neurology. 2024;103(5):e209699. pmid:39110916
- 28. Weintraub D, Simuni T, Caspell-Garcia C, Coffey C, Lasch S, Siderowf A, et al. Cognitive performance and neuropsychiatric symptoms in early, untreated Parkinson’s disease. Mov Disord. 2015;30(7):919–27. pmid:25737166
- 29. Becker S, Bode M, Brockmann K, Gasser T, Michaelis K, Solbrig S, et al. Cognitive-Driven Activities of Daily Living Impairment as a Predictor for Dementia in Parkinson Disease: A Longitudinal Cohort Study. Neurology. 2022;99(23):e2548–60. pmid:36240089
- 30. Hattori T, Reynolds R, Wiggs E, Horovitz SG, Lungu C, Chen G, et al. Neural correlates of working memory and compensation at different stages of cognitive impairment in Parkinson’s disease. Neuroimage Clin. 2022;35:103100. pmid:35780660
- 31. Owen AM, Iddon JL, Hodges JR, Summers BA, Robbins TW. Spatial and non-spatial working memory at different stages of Parkinson’s disease. Neuropsychologia. 1997;35(4):519–32. pmid:9106280
- 32. Sharp DJ, Bonnelle V, De Boissezon X, Beckmann CF, James SG, Patel MC, et al. Distinct frontal systems for response inhibition, attentional capture, and error processing. Proc Natl Acad Sci U S A. 2010;107(13):6106–11. pmid:20220100
- 33. Chambers CD, Garavan H, Bellgrove MA. Insights into the neural basis of response inhibition from cognitive and clinical neuroscience. Neurosci Biobehav Rev. 2009;33(5):631–46. pmid:18835296
- 34. Prasuhn J, Prasuhn M, Fellbrich A, Strautz R, Lemmer F, Dreischmeier S, et al. Association of Locus Coeruleus and Substantia Nigra Pathology With Cognitive and Motor Functions in Patients With Parkinson Disease. Neurology. 2021;97(10):e1007–16. pmid:34187859
- 35. Ye R, Hezemans FH, O’Callaghan C, Tsvetanov KA, Rua C, Jones PS, et al. Locus Coeruleus Integrity Is Linked to Response Inhibition Deficits in Parkinson’s Disease and Progressive Supranuclear Palsy. J Neurosci. 2023;43(42):7028–40. pmid:37669861
- 36. MacDonald HJ, Byblow WD. Does response inhibition have pre- and postdiagnostic utility in Parkinson’s disease?. J Mot Behav. 2015;47(1):29–45. pmid:25575221
- 37. Zacks JM. Neuroimaging studies of mental rotation: a meta-analysis and review. J Cogn Neurosci. 2008;20(1):1–19. pmid:17919082
- 38. Koshino H, Carpenter PA, Keller TA, Just MA. Interactions between the dorsal and the ventral pathways in mental rotation: an fMRI study. Cogn Affect Behav Neurosci. 2005;5(1):54–66. pmid:15913008
- 39. Kawashima S, Shimizu Y, Ueki Y, Matsukawa N. Impairment of the visuospatial working memory in the patients with Parkinson’s Disease: an fMRI study. BMC Neurol. 2021;21(1):335. pmid:34479502
- 40. Bek J, Humphries S, Poliakoff E, Brady N. Mental rotation of hands and objects in ageing and Parkinson’s disease: differentiating motor imagery and visuospatial ability. Exp Brain Res. 2022;240(7–8):1991–2004. pmid:35680657
- 41. Crucian GP, Armaghani S, Armaghani A, Foster PS, Burks DW, Skoblar B, et al. Dopamine does not appear to affect mental rotation in Parkinson’s disease. J Mov Disord. 2014;7(2):77–83. pmid:25360231
- 42. Penko AL, Streicher MC, Dey T, Rosenfeldt AB, Alberts JL. Parkinson’s gait kinematics deteriorates across multiple cognitive domains under dual-task paradigms. Clin Neurol Neurosurg. 2020;197:106083. pmid:32683199
- 43. Penko AL, Streicher MC, Koop MM, Dey T, Rosenfeldt AB, Bazyk AS, et al. Dual-task Interference Disrupts Parkinson’s Gait Across Multiple Cognitive Domains. Neuroscience. 2018;379:375–82. pmid:29577998
- 44. Harrie A, Hampstead BM, Lewis C, Herreshoff E, Kotagal V. Cognitive correlates of dual tasking costs on the timed up and go test in Parkinson disease. Clin Park Relat Disord. 2022;7:100158. pmid:35957864
- 45. Amboni M, Ricciardi C, Cuoco S, Donisi L, Volzone A, Ricciardelli G, et al. Mild Cognitive Impairment Subtypes Are Associated With Peculiar Gait Patterns in Parkinson’s Disease. Front Aging Neurosci. 2022;14:781480. pmid:35299943
- 46. Martínez-Martín P, Rodríguez-Blázquez C, Alvarez M, Arakaki T, Arillo VC, Chaná P, et al. Parkinson’s disease severity levels and MDS-Unified Parkinson’s Disease Rating Scale. Parkinsonism Relat Disord. 2015;21(1):50–4. pmid:25466406
- 47. Chagas MHN, Tumas V, Rodrigues GR, Machado-de-Sousa JP, Filho AS, Hallak JEC, et al. Validation and internal consistency of Patient Health Questionnaire-9 for major depression in Parkinson’s disease. Age Ageing. 2013;42(5):645–9. pmid:23761457
- 48. Jost ST, Kaldenbach M-A, Antonini A, Martinez-Martin P, Timmermann L, Odin P, et al. Levodopa Dose Equivalency in Parkinson’s Disease: Updated Systematic Review and Proposals. Mov Disord. 2023;38(7):1236–52. pmid:37147135
- 49. Owen AM, Beksinska M, James M, Leigh PN, Summers BA, Marsden CD, et al. Visuospatial memory deficits at different stages of Parkinson’s disease. Neuropsychologia. 1993;31(7):627–44. pmid:8371837
- 50. Hosseini M, Borhani-Haghighi A, Petramfar P, Foroughi AA, Ostovan VR, Nami M. Evaluating cognitive impairment in the early stages of Parkinson’s disease using the Cambridge brain sciences-cognitive platform. Clin Neurol Neurosurg. 2023;232:107866. pmid:37413872
- 51. Kochan NA, Croot K, Crawford JD, Allison KC, Rossie M, Brodaty H, et al. Computer‐administered neuropsychological assessment batteries: Validity, reliability, and user experience in an Australian sample of community‐living older adults in the CogSCAN Study. Alzheimer’s & Dementia. 2022;18(S7).
- 52. Kochan NA, Heffernan M, Valenzuela M, Sachdev PS, Lam BCP, Fiatarone Singh M, et al. Reliability, Validity, and User-Experience of Remote Unsupervised Computerized Neuropsychological Assessments in Community-Living 55- to 75-Year-Olds. J Alzheimers Dis. 2022;90(4):1629–45. pmid:36314208
- 53. Hassan SA, Bonetti LV, Kasawara KT, Beal DS, Rozenberg D, Reid WD. Decreased automaticity contributes to dual task decrements in older compared to younger adults. Eur J Appl Physiol. 2022;122(4):965–74. pmid:35084541
- 54. Kanko RM, Laende EK, Strutzenberger G, Brown M, Selbie WS, DePaul V, et al. Assessment of spatiotemporal gait parameters using a deep learning algorithm-based markerless motion capture system. J Biomech. 2021;122:110414. pmid:33915475
- 55. McGuirk TE, Perry ES, Sihanath WB, Riazati S, Patten C. Feasibility of Markerless Motion Capture for Three-Dimensional Gait Assessment in Community Settings. Front Hum Neurosci. 2022;16:867485. pmid:35754772
- 56. Riazati S, McGuirk TE, Perry ES, Sihanath WB, Patten C. Absolute Reliability of Gait Parameters Acquired With Markerless Motion Capture in Living Domains. Front Hum Neurosci. 2022;16:867474. pmid:35782037
- 57. Woltring HJ. A Fortran package for generalized, cross-validatory spline smoothing and differentiation. Advances in Engineering Software (1978). 1986;8(2):104–13.
- 58.
Dempster WT. Space requirements of the seated operator: geometrical, kinematic, and mechanical aspects of the body, with special reference to the limbs. 1955.
- 59. Hanavan E. A Mathematical Model of the Human Body (Technical report AMRL-TDR-64–102) Wright-Patterson Air Force Base. OJ Wright Air Dev Cent. 1964.
- 60. Plagenhoef S, Evans FG, Abdelnour T. Anatomical Data for Analyzing Human Motion. Research Quarterly for Exercise and Sport. 1983;54(2):169–78.
- 61. Wall JC, Bell C, Campbell S, Davis J. The Timed Get-up-and-Go test revisited: measurement of the component tasks. J Rehabil Res Dev. 2000;37(1):109–13. pmid:10847578
- 62. Simonet A, Fourcade P, Loete F, Delafontaine A, Yiou E. Evaluation of the Margin of Stability during Gait Initiation in Young Healthy Adults, Elderly Healthy Adults and Patients with Parkinson’s Disease: A Comparison of Force Plate and Markerless Motion Capture Systems. Sensors (Basel). 2024;24(11):3322. pmid:38894112
- 63. Grood ES, Suntay WJ. A joint coordinate system for the clinical description of three-dimensional motions: application to the knee. J Biomech Eng. 1983;105(2):136–44. pmid:6865355
- 64. Zifchock RA, Davis I, Higginson J, Royer T. The symmetry angle: a novel, robust method of quantifying asymmetry. Gait Posture. 2008;27(4):622–7. pmid:17913499
- 65. Virtanen P, Gommers R, Oliphant TE, Haberland M, Reddy T, Cournapeau D, et al. SciPy 1.0: fundamental algorithms for scientific computing in Python. Nat Methods. 2020;17(3):261–72. pmid:32015543
- 66.
Seabold S, Perktold J. Statsmodels: Econometric and Statistical Modeling with Python. In: Proceedings of the Python in Science Conference, 2010. 92–6. https://doi.org/10.25080/majora-92bf1922-011
- 67. Jayakody O, Breslin M, Srikanth VK, Callisaya ML. Gait Characteristics and Cognitive Decline: A Longitudinal Population-Based Study. J Alzheimers Dis. 2019;71(s1):S5–14. pmid:30958358
- 68. Toots ATM, Taylor ME, Lord SR, Close JCT. Associations Between Gait Speed and Cognitive Domains in Older People with Cognitive Impairment. J Alzheimers Dis. 2019;71(s1):S15–21. pmid:31256121
- 69. Piet A, Geritz J, Garcia P, Irsfeld M, Li F, Huang X, et al. Predicting executive functioning from walking features in Parkinson’s disease using machine learning. Sci Rep. 2024;14(1):29522. pmid:39604483
- 70. Peysakhovich B, Zhu O, Tetrick SM, Shirhatti V, Silva AA, Li S, et al. Primate superior colliculus is causally engaged in abstract higher-order cognition. Nat Neurosci. 2024;27(10):1999–2008. pmid:39300307
- 71. Armstrong RA. Oculo-Visual Dysfunction in Parkinson’s Disease. J Parkinsons Dis. 2015;5(4):715–26. pmid:26599301
- 72. Nemanich ST, Earhart GM. Freezing of gait is associated with increased saccade latency and variability in Parkinson’s disease. Clin Neurophysiol. 2016;127(6):2394–401. pmid:27178858
- 73. Li H, Ma W, Li C, He Q, Zhou Y, Xie A. Combined diagnosis for Parkinson’s disease via gait and eye movement disorders. Parkinsonism Relat Disord. 2024;123:106979. pmid:38669851
- 74. Guzik A, Drużbicki M, Perenc L, Wolan-Nieroda A, Turolla A, Kiper P. Establishing the Minimal Clinically Important Differences for Sagittal Hip Range of Motion in Chronic Stroke Patients. Front Neurol. 2021;12:700190. pmid:34539552
- 75. Alcock L, Galna B, Lord S, Rochester L. Characterisation of foot clearance during gait in people with early Parkinson׳s disease: Deficits associated with a dual task. J Biomech. 2016;49(13):2763–9. pmid:27363617
- 76. Spildooren J, Vercruysse S, Heremans E, Galna B, Vandenbossche J, Desloovere K, et al. Head-pelvis coupling is increased during turning in patients with Parkinson’s disease and freezing of gait. Mov Disord. 2013;28(5):619–25. pmid:23408374
- 77. Djamshidian A, O’Sullivan SS, Lees A, Averbeck BB. Stroop test performance in impulsive and non impulsive patients with Parkinson’s disease. Parkinsonism Relat Disord. 2011;17(3):212–4. pmid:21247790
- 78. Manza P, Amandola M, Tatineni V, Li C-SR, Leung H-C. Response inhibition in Parkinson’s disease: a meta-analysis of dopaminergic medication and disease duration effects. NPJ Parkinsons Dis. 2017;3:23. pmid:28702504
- 79. Verghese J, Ayers E, Barzilai N, Bennett DA, Buchman AS, Holtzer R, et al. Motoric cognitive risk syndrome: Multicenter incidence study. Neurology. 2014;83(24):2278–84. pmid:25361778
- 80. Purks JL, Arbatti L, Hosamath A, Amara AW, Anderson KE, Chahine L, et al. Cognitive Symptoms in Cross-Sectional Parkinson Disease Cohort Evaluated by Human-in-the-Loop Machine Learning and Natural Language Processing. Neurol Clin Pract. 2024;14(5):e200334. pmid:38962394
- 81. Siderowf A, Concha-Marambio L, Lafontant D-E, Farris CM, Ma Y, Urenia PA, et al. Assessment of heterogeneity among participants in the Parkinson’s Progression Markers Initiative cohort using α-synuclein seed amplification: a cross-sectional study. Lancet Neurol. 2023;22(5):407–17. pmid:37059509
- 82. Guo Y, Liu F-T, Hou X-H, Li J-Q, Cao X-P, Tan L, et al. Predictors of cognitive impairment in Parkinson’s disease: a systematic review and meta-analysis of prospective cohort studies. J Neurol. 2021;268(8):2713–22. pmid:32162063
- 83. Nocera JR, Arsik I, Keskinocak P, Lepley-Flood A, Lah JJ, Levey AI, et al. The Feasibility of Measuring Gait in an Outpatient Cognitive Neurology Clinical Setting. J Alzheimers Dis. 2019;71(s1):S51–5. pmid:31322564
- 84. Weemering DN, Beelen A, Kliest T, van Leeuwen LAG, van den Berg LH, van Eijk RPA. Trial Participation in Neurodegenerative Diseases: Barriers and Facilitators: A Systematic Review and Meta-Analysis. Neurology. 2024;103(1):e209503. pmid:38830181