GRAPHICAL ABSTRACT

INTRODUCTION

Parkinson’s disease (PD) is characterized by motor symptoms such as bradykinesia, rigidity, tremor, and postural instability. Among these symptoms, gait disturbances are particularly debilitating, as they reduce mobility, impair independence, and increase the risk of falls. Considerable research has investigated gait abnormalities in PD patients during usual walking and under dual-task conditions [1,2]. However, more complex tasks, such as obstacle walking, which demand motor control and cognitive engagement, have received less attention and are crucial for daily life. Obstacle walking requires precise motor execution and adaptive planning in response to environmental changes [3]. Compared with healthy adults, people with PD exhibit reduced step length and slower gait speed during obstacle walking [4,5]. Furthermore, compared with usual walking, obstacle walking leads to more pronounced gait impairments among individuals with PD [4]. Everyday walking often requires the integration of the cognitive and motor systems to adapt to environmental changes. This integration is essential for safe movement. Many factors affect obstacle walking performance, including attention, executive function and visuospatial ability [6-8]. Impairments in these domains can reduce gait stability and increase fall risk during obstacle walking in patients with PD [9].

Functional near-infrared spectroscopy (fNIRS) studies have revealed increased prefrontal cortex (PFC) activation during obstacle walking in PD patients, indicating greater cognitive load and reduced neural efficiency [10]. Motor-related cortical regions, including the premotor cortex (PMC) and supplementary motor area (SMA), also contribute to gait modulation under cognitively demanding conditions, such as dual-task walking [11]. The PMC integrates visuospatial information for movement adaptation [12]. The SMA contributes to internally guided sequences and anticipatory postural adjustments—both of which are critical for stepping over obstacles [13]. Collectively, the PMC and SMA enable adaptive gait and complex motor adjustments.

Most previous works have examined single-obstacle events, which do not capture the continuous demands of real-world walking. In everyday contexts, obstacle negotiation involves repeated adjustments and sustained attention engagement, executive control, and motor planning—imposing a prolonged cognitive load that may differentially affect gait and brain activation over time.

Approximately 40% of individuals with PD develop mild cognitive impairment (PD-MCI) [14]. Compared with their cognitively intact counterparts (PD-non-MCI), PD-MCI patients demonstrate poor gait performance even during usual walking [15]. Because gait modulation under complex conditions relies on both motor and cognitive resources (the PFC, PMC, and SMA), PD-MCI patients may face greater difficulty adapting dynamically. However, how cognitive impairment influences obstacle walking and cortical activation patterns remains largely unexplored. No studies have investigated phase-specific brain activation during prolonged obstacle walking [5]. Therefore, this study examined temporal cortical activation and gait dynamics across phases of prolonged obstacle walking in PD-MCI patients and PD-non-MCI patients. Using fNIRS and continuous gait analysis, our findings revealed how cognitive impairment affects motor behavior and alters the temporal trajectory of cortical adaptation. These findings advance our understanding of compensatory mechanisms in PD and highlight distinct neural strategies employed by PD-MCI and PD-non-MCI populations while overcoming real-world walking challenges.

MATERIALS & METHODS

Study design

We recruited PD-MCI and PD non-MCI patients and collected clinical data such as age, sex, educational level, disease duration, the more affected side, motor function assessed through part III of the Unified Parkinson’s Disease Rating Scale [16], and PD medication usage, which were converted to levodopa equivalent daily doses [17]. Cognition was assessed via the Mini-Mental State Examination (MMSE) [18] and Montreal Cognitive Assessment (MoCA). Executive function was measured with the Semantic Verbal Fluency Test [19], working memory and attention with the Digit Span Test [20], and visuospatial ability with the Rey–Osterrieth Complex Figure Test [21].

The primary outcomes included obstacle walking performance, brain activation during obstacle walking, and obstacle crossing performance. All the participants were assessed during the on-phase of the medication. The protocol was approved by the Institutional Review Board of National Yang Ming Chiao Tung University (YM109001F) and is registered in the Thai Clinical Trial Registry (TCTR20210420001). Written informed consent was obtained from all participants prior to participation.

Participants

Two groups were included: PD-MCI (MoCA <26) and PD-non-MCI (MoCA ≥26). MCI was defined by the Movement Disorder Society (MDS) Task Force Level I criteria [22]. To increase diagnostic precision within the local context, we utilized the Taiwanese version of the MoCA, with scores adjusted for age and education level on the basis of recent localized normative data [23]. The inclusion criteria for both groups were 1) idiopathic PD, 2) aged 50–80 years, 3) Hoehn–Yahr stage 1 to 3, 4) MMSE score >24, 5) ability to walk for one minute independently, and 6) stable PD treatment for ≥3 months. The exclusion criteria were 1) an unstable medical condition; 2) the presence of any neurological or psychiatric disorder (e.g., stroke or depression) that could affect participation in the study; 3) dementia, defined as a clinical diagnosis consistent with the DSM-5 criteria and/or a neuropsychological assessment (MMSE score ≤24 or Clinical Dementia Rating [24] ≥1); and 4) atypical or secondary parkinsonism, including vascular parkinsonism, drug-induced parkinsonism, normal-pressure hydrocephalus, progressive supranuclear palsy, multiple system atrophy, and corticobasal degeneration. All the participants provided written consent before the assessment.

Measurements

Obstacle walking performance

Obstacle walking was recorded by two wearable Physilog sensors (GaitUp system) attached to both shoes. Participants walked continuously for one minute at their comfortable speed through 10 obstacles placed on a 30 m long walkway (interval 3 m; obstacle height=20% of leg length) (Figure 1) [25]. The test was repeated four times. GaitUp software was used to calculate gait speed, cadence, stride length and stride time, and averages were used for analysis.

Brain activation during obstacle walking

A wearable 8-source, 8-detector fNIRS system (NIRSport2; NIRx Medical Technologies) recorded hemodynamic changes across 14 channels covering the bilateral PFC (Ch.1 and 2), PMC (Ch.3, 4, 5, and 6 in the left PMC; Ch.11, 12, 13, and 14 in the right PMC), and SMA (Ch.7, 8, 9, and 10) (Figure 2). For regional analyses, left- and right-hemisphere channels were averaged to yield a single value per region of interest (ROI). Activation for each channel was quantified as the mean change in oxygenated hemoglobin (ΔHbO) relative to a 5-second prewalk baseline. During the 60-second walking task, analyses focused on the 5–40 seconds interval to account for the hemodynamic delay and minimize the influence of gait initiation. This interval was divided into early (5–20 s) and late (20–40 s) phases to differentiate initial adaptation from sustained cortical engagement during obstacle walking, which is consistent with prior PD fNIRS studies [11]. The final 40–60 seconds segment was excluded to reduce potential confounding effects related to fatigue or anticipatory adjustments prior to task termination. ΔHbO was averaged for early and late phases per repetition. The average values of four repetitions were considered to derive a single mean ΔHbO per ROI for each phase for statistical analysis.

For signal-to-noise quality, the relative coefficient of variance (CV, %) of each channel and session was calculated by the following formula: CV=standard deviation [SD]/mean×100%. The raw signal was visually inspected, and the channels with a CV greater than 15% and the sessions with a CV greater than 10% were manually rejected. Motion artifacts (SD >10%, amplitude >5%) were corrected with a wavelet algorithm. Data were filtered (0.005 Hz–0.3 Hz) [26] and converted to oxygenated (HbO) and deoxygenated hemoglobin (HbR) concentrations using the modified Beer–Lambert law [27,28]. The signals were processed with the HOMER2 fNIRS package (https://www.nitrc.org/projects/homer2/), and HbO values were computed via custom MATLAB scripts (MathWorks).

Obstacle crossing performance

Obstacle crossing performance was measured using an 8-m GAITRite walkway (CIR Systems Inc.) with a plastic obstacle (height=20% of leg length) positioned midwalk (Figure 2). Participants completed five trials, stepping over the obstacle at a comfortable pace. The first crossing foot was the leading leg, followed by the trailing leg. Gait parameters—crossing speed, stride length, stride time, and step width—were extracted for both legs. The crossing speed was defined as the time from heeloff to heel contact; the stride length and time were measured between consecutive steps of the same foot, and the step width was defined as the lateral distance between both feet after crossing. The mean values from five trials were analyzed.

Sample size calculation

An a priori power analysis was performed using G*Power (v3.1.9.7; https://www.psychologie.hhu.de/arbeitsgruppen/allgemeine-psychologie-und-arbeitspsychologie/gpower). Because the primary hypothesis concerned the between-group difference in phase-related changes, sample size estimation was informed by a prior study reporting a large between-group difference in late-early PFC activity (Cohen’s d=1.81) [29]. To avoid overestimation, we conservatively assumed a smaller effect size (Cohen’s d=1.0) for the anticipated group-by-phase effect. On the basis of this assumption, with α=0.05 and 80% power, a total sample size of 34 participants was needed.

Statistical analyses

All analyses were performed using SPSS 25.0 (IBM Corp.). Data normality was tested with the Shapiro–Wilk test. Continuous variables are expressed as the mean±SD and were compared between groups using independent t-tests; categorical variables were compared using chi-square tests.

Generalized estimating equations (GEEs) were employed to examine the effects of group, phase, and their interaction. GEEs were chosen to account for the dependency of repeated measures among participants. To determine the optimal working correlation structure, independent, exchangeable, and unstructured structures were compared using the quasilikelihood under the independence model criterion (QIC), with the structure yielding the lowest QIC. The exchangeable correlation structure showed the best fit across outcomes. The significance of the main effects and the group × phase interaction was evaluated using Wald chi-square tests. Significant effects were followed by post hoc comparisons to examine between-group differences within each phase and within-group changes across phases. Bonferroni correction controlled for type I error. Spearman’s rank correlations were used to examine associations between brain activation and gait performance for exploratory purposes and were not adjusted for multiple comparisons. Significance was set at two-tailed p<0.05.

RESULTS

A total of 19 individuals with PD-non-MCI (MMSE 29.0±1.3, range 28–30; MoCA 27.1±1.3, range 26–30) and 15 individuals with PD-MCI (MMSE 26.4±1.8, range 24–29; MoCA 22.0±2.0, range 18–25) were included. There were no significant differences between the two groups in terms of most baseline characteristics, except for cognitive performance (Table 1). As expected, the PD-MCI group demonstrated significantly lower performance regarding attention/working memory, executive function, and visuospatial abilities. These multidomain deficits support the validity of this Level I classification and our adjusted MoCA-based grouping in identifying a population with clinically relevant cognitive impairment.

Obstacle walking performance

For gait speed, a significant main effect of group was observed (β=-0.18; p=0.02), indicating that the PD-MCI group demonstrated slower gait speed than the PD-non-MCI group across phases. The main effects of phase and the group × phase interaction were not significant, suggesting that the difference in gait speed between the groups was consistent across the early and late phases (Table 2).

For cadence, no significant main effects of group or phase or a significant group × phase interaction were detected. For stride length, the main effect of group tended to be significant (p=0.05), whereas the phase effect and the group × phase interaction were not significant. For stride time, a significant main effect of phase was observed (β=0.02, p=0.01), indicating that stride time increased from the early phase to the late phase across groups. However, the main effect of group and the group × phase interaction were not significant.

Descriptive statistics and pairwise comparisons, specifically comparing the PD-non-MCI and PD-MCI groups across the early and late phases, were conducted to further explore the data. In both phases, the PD-MCI group demonstrated significantly slower gait speed than the PD-non-MCI group (early phase: t=2.65, p=0.02; late phase: t=2.54, p=0.02). With respect to stride length, PD-MCI patients exhibited shorter strides in both phases (early phase: t=2.33, p=0.03; late phase: t=2.38, p=0.03). However, no significant differences were observed in cadence or stride time in either phase among the groups. Additionally, no within-group changes were observed across the early and late phases for either group.

Brain activation

Across all three regions, significant main effects of group were observed, indicating higher HbO levels in the PD-MCI group than in the PD-non-MCI group (PFC: β=0.22, p<0.01; SMA: β=0.13, p=0.03; PMC: β=0.19, p<0.01) (Table 3). Significant main effects of phase were also detected in all the regions (PFC: β=-0.07, p=0.01; SMA: β=-0.09, p<0.01; PMC: β=-0.08, p<0.01), suggesting that HbO levels decreased from the early phase to the late phase. With respect to the group× phase interaction, significant interactions were detected between the SMA and the PMC (SMA: β=-0.12, p=0.02; PMC: β=-0.12, p=0.02), whereas the interaction effect in the PFC did not reach significance (p=0.09). These results indicate that the decrease in HbO levels from the early phase to the late phase was greater in the PD-MCI group than in the PD-non-MCI group, particularly in the SMA and PMC.

In addition to the GEE analysis, comparisons were performed between the PD-non-MCI and PD-MCI groups for each region across the early and late phases. Within-group early-late comparisons revealed significant changes across all areas in both groups. In the early phase, PD-MCI patients demonstrated greater levels of activation in all the regions, in which the difference between the groups was significant (PFC: t=-3.53, p=0.002; SMA: t=-3.56, p=0.002; PMC: t=-2.97, p=0.009) (Table 3). However, in the late phase, no significant differences were observed between the two groups. A significant interaction between the group and phase was found in both the SMA and the PMC (SMA: coefficient=-0.12, p=0.02; PMC: coefficient=-0.12, p=0.02), indicating that the changes in the SMA and the PMC from the early phase to the late phase differed between the groups. To further explore this interaction, we calculated the activation change (late phase - early phase) for each group. The results revealed significant between-group differences in both the SMA (p=0.009) and the PMC (p=0.012).

Obstacle crossing

In terms of obstacle crossing performance, participants with PD-non-MCI exhibited significantly longer stride lengths (p=0.007 and 0.018) and faster crossing speeds (p=0.023 and 0.039) for both the leading and trailing foot than those with PD-MCI did. However, no significant group differences were found in stride duration or step width (Table 4).

Correlations between cortical activation and gait in the PD-MCI and PD-non-MCI groups

Correlation analyses were conducted to explore the associations between gait parameters and cortical hemodynamic responses across different brain regions. In the PD-MCI group, a significant negative correlation was observed between HbO levels in the PFC and gait speed during the late phase (ρ=-0.56; p=0.0389). Additionally, significant negative associations were detected between HbO levels in the PMC and gait speed during the early phase (ρ=-0.28; p=0.0442), suggesting that distinct brain regions may contribute to gait regulation at different stages of task execution in individuals with cognitive impairment. In contrast, the PD-non-MCI group did not exhibit significant correlations between HbO levels and gait speed or stride length in any region (Table 5).

DISCUSSION

This study provides novel insights into the temporal dynamics of motor and neural responses during the early and late phases of obstacle walking in PD-MCI and PD-non-MCI patients, highlighting potential decreases in sustained motor-cognitive engagement over time. The main findings are as follows: 1) PD-MCI patients exhibited poorer gait performance in both the early and late phases but maintained stable performance over time; 2) cortical activation was greater in PD-MCI patients than in PD-non-MCI patients, with both groups showing greater activation in the early phase than in the late phase. Additionally, compared with the PD-non-MCI group, the PD-MCI group exhibited a greater decline in performance across phases; and 3) in the PD-MCI group, early-phase gait performance was negatively associated with PMC activity, while late-phase performance was negatively associated with PFC activity, suggesting that phase-specific neural engagement may reflect an inefficient compensation mechanism.

Obstacle gait alterations and preserved stability over time in PD-MCI patients

Our results revealed that participants with PD-MCI exhibited reduced speed and stride length compared with those with PD-non-MCI during both phases, suggesting that cognitive deficits limit dynamic gait adaptation [30], which is consistent with evidence that locomotor stability depends on physical and cognitive resources [31,32]. Impairments in working memory and executive function may further compromise gait adjustment [32]. Despite these deficits, PD-MCI participants maintained stable gait performance over time, providing evidence that dynamic stability can be preserved while performing cognitively demanding walking tasks [33]. Optimal feedback control theory posits that motor adaptation relies on the formation of internal models of the environment and the selection of movement plans that minimize cost and maximize success [34,35]. In the context of our task, both PD-MCI and PD-non-MCI patients may have rapidly acquired such internal models and adopted control strategies, enabling the maintenance of gait performance despite ongoing cognitive demands. These findings suggest that the capacity for motor adaptation, as a mechanism supporting dynamic stability, may remain relatively intact in PD-MCI patients, particularly under consistent task requirements.

Greater cortical activation and phase-related decline during obstacle walking in PD-MCI patients

PD-MCI participants exhibited higher PFC, SMA, and PMC activation but slower gait speed. Similar findings have been reported in studies of older adults, which revealed that increased prefrontal activity was associated with slower obstacle walking speed [36,37]. This finding supports the concept of neural inefficiency, suggesting that the elevated cortical recruitment observed in PD-MCI patients reflects a less efficient compensatory strategy, where increased activation fails to mitigate motor decline [36,38].

We observed a reduction in both cognitive and motor-related cortical activity from the early phase to the late phase, without a corresponding decline in obstacle gait performance. This pattern may reflect neural adaptation, whereby initial cortical engagement facilitates task initiation, and subsequent attenuation of activity with task repetition indicates a shift from effortful control to a more automatic processing [39]. Similar patterns have been reported in motor learning paradigms, where early increases in activation within cognitive and motor regions during skill acquisition are followed by reduced cortical engagement during later stages or retention, reflecting adaptive changes in neural processing as the task becomes more familiar [40,41]. Thus, attenuation of cortical activity, without a decline in walking performance, may be interpreted as evidence of adaptive neural optimization.

Both groups exhibited reduced PMC and SMA activation over time, with a greater decrease observed in PD-MCI patients, indicating that distinct neural adaptation strategies were employed by the two groups. PD-MCI patients showed a more pronounced modulation of cortical activity across the task, suggesting a strategy characterized by initial upregulation of cortical engagement followed by a more rapid attenuation as the task progressed. In contrast, PD-non-MCI patients demonstrated a smaller but significant reduction in activation, indicating more gradual and sustained modulation of cortical resources. Although the PFC also tended to decrease over time in both groups, the absence of a significant group-by-phase interaction suggests that its engagement remained relatively consistent across groups, highlighting the role of the PFC in cognitive control throughout the task. In contrast, the PMC and SMA, which are directly involved in motor planning and execution, demonstrated group-specific modulation patterns, suggesting that these regions more distinctly reflect the neural adaptations associated with obstacle walking performance. Collectively, these findings suggest that gait performance is maintained through distinct cortical adaptation profiles: accelerated neural adjustment in PD-MCI patients and more stable adaptation in PD-non-MCI patients.

Phase-specific brain–gait performance associations in the PD-MCI group

Our results demonstrated that slower gait speed was linked to greater PMC activation in the early phase of obstacle walking, while this association shifted to the PFC in the later phase, possibly indicating a shift from motor to cognitive control mechanisms in PD-MCI patients but not in PD-non-MCI patients. The negative correlations suggest inefficient compensation, where increased neural activation does not translate into better gait performance [37]. Although the association between early-phase PMC activation and gait speed was significant (Spearman’s ρ=-0.28; p=0.044), the small effect size indicates limited shared variance. Accordingly, these findings should be interpreted conservatively. Given the exploratory nature of these analyses and the absence of multiplicity correction, the associations should be considered preliminary rather than confirmatory. This pattern implies that individuals with PD-MCI may initially engage motor-related regions, such as the PMC, to support gait control by integrating sensory inputs into motor planning. However, as the walking task becomes more familiar, they appear to shift their cognitive resources toward precision stepping, an effort that requires spatial processes and typically involves prefrontal regions. This view is supported by that of Koenraadt et al. [42], who highlighted the importance of the PFC during spatially precise stepping, as measured by fNIRS.

Temporal associations between brain activation and gait performance were observed in the PD-MCI group but not in the PD-non-MCI group. One possible explanation is that the obstacle walking speed of the PD-non-MCI group (1.13±0.19 m/sec for the early phase and 1.12±0.18 m/sec for the late phase) was close to the typical walking speed reported for individuals with PD (approximately 1.2 m/sec) [43], indicating sufficient motor capacity to navigate obstacles without additional cortical recruitment. Thus, their gait performance remained stable and comparable to their usual walking speed, thereby diminishing the observable coupling between cortical activation and gait performance. These findings suggest that the relationship between brain activity and gait might be modulated not only by cognitive status but also by the difficulty level of the task. Future studies with larger samples are warranted to better characterize the role of cortical activation in obstacle walking performance in PD-MCI patients.

Effects of cognitive impairment on obstacle crossing in PD-MCI patients

PD-MCI participants demonstrated slower step speed and shorter stride length for both the leading and trailing foot during obstacle crossing. Such decreases observed in the PD-MCI group may reflect a cautious gait strategy as a result of impaired cognitive–motor integration, possibly adopted to minimize the risk of collision or loss of balance. Compared with healthy controls, individuals with PD tend to adopt more conservative gait strategies, such as reduced step length and slower stepping speed during obstacle crossing [44]. Our study further demonstrated that cognitive impairment exacerbated difficulties in obstacle crossing among individuals with PD. Previous studies have shown that obstacle crossing performance is associated with reduced cognitive performance [45,46], indicating that individuals with cognitive decline may have difficulty integrating environmental information and adapting motor output accordingly. These findings further emphasize that obstacle crossing is not merely a motor task but also involves higher-level cognitive processing, highlighting the importance of cognitive-motor integration in navigating walking environments that include obstacles. Evidence from Pieruccini-Faria et al.’s [47] study revealed that individuals with MCI and reduced frontotemporal gray matter volumes have difficulty adjusting their gait when they are approaching obstacles, suggesting that structural atrophy may contribute to impaired cognitive–motor integration. Although brain volumes were not measured in this study, cautious gait patterns likely reflect these integration challenges, warranting further investigation.

Limitations

Our study has several limitations. First, although our sample included individuals with a Hoehn–Yahr stage up to 3, the majority of participants were in stages 1 to 2 of PD. As such, the findings primarily reflect milder PD and may not be generalizable to patients with more advanced PD. Future studies including a broader range of disease severity are needed to determine whether similar neural activation patterns are observed in individuals with more advanced PD. Second, short-separation channels were not included in our fNIRS setup, limiting the ability to separate cortical signals from superficial physiological noise and potentially affecting activation estimates. Future studies should incorporate short-separation optodes to enable more precise removal of extracerebral signals and improve the validity of fNIRS results. Third, PD-MCI classification was based on a Level I (screening-based) assessment using the MoCA. Although the groups differed in terms of additional neuropsychological measures, the battery did not include the domain-specific testing required for a formal Level II diagnosis. Therefore, the risk of misclassification cannot be excluded. Fourth, although the PD motor phenotype (tremor-dominant/postural instability and gait difficulty/indeterminate) was comparable between groups, emerging pathophysiological subtypes (e.g., body-first and brain-first) were not considered, which may differentially influence gait control and cortical activation. Finally, PD-MCI subtypes were not distinguished, although prior studies suggest distinct neural activation patterns across subtypes [48]. Future studies should consider cognitive profiles to better elucidate their influence on gait-related brain activity.

Conclusions

Compared with PD-non-MCI individuals, PD-MCI individuals exhibited poorer gait performance and greater cortical activation, indicating that more neural resources are needed under challenging walking conditions. Notably, only PD-MCI patients demonstrated a temporal shift in brain–gait associations during obstacle walking, suggesting altered or less efficient neural strategies. These findings highlight the critical role of cognitive–motor integration in complex walking tasks and suggest the need for future research to examine whether targeted intervention can increase neural efficiency in individuals with PD-MCI.

Notes

Conflicts of Interest

The authors have no financial conflicts of interest.

Funding Statement

This work was supported by grants from the National Science and Technology Council (MOST109-2314-B-010-035-MY3) and National Health Research Institutes (NHRI-EX-111-10913PI).

Acknowledgments

We would like to thank all the participants for their time.

Author Contributions

Conceptualization: Pei-Ling Wong, Ray-Yau Wang. Data curation: Pei-Ling Wong, Chen-Wei Hung, Nai-Chen Yeh. Formal analysis: Pei-Ling Wong, Nai-Chen Yeh. Funding acquisition: Ray-Yau Wang. Investigation: Pei-Ling Wong, Yea-Ru Yang, Ray-Yau Wang. Project administration: Yea-Ru Yang, Ray-Yau Wang. Resources: Ray-Yau Wang, Shih-Jung Cheng. Software: PeiLing Wong. Supervision: Ray-Yau Wang. Validation: Nai-Chen Yeh. Visualization: Pei-Ling Wong. Writing—original draft: Pei-Ling Wong. Writing—review & editing: all authors.

Figure 1.

Setup for obstacle walking assessment with fNIRS source–detector placement and gait parameter measurement. PFC, prefrontal cortex; PMC, premotor cortex; SMA, supplementary motor area; fNIRS, functional near-infrared spectroscopy.

Figure 2.

Schematic representation of gait parameters measured during the obstacle crossing task.

REFERENCES

• 1. Raffegeau TE, Krehbiel LM, Kang N, Thijs FJ, Altmann LJP, Cauraugh JH, et al. A meta-analysis: Parkinson’s disease and dual-task walking. Parkinsonism Relat Disord 2019;62:28–35.ArticlePubMedPMC
• 2. Zanardi APJ, da Silva ES, Costa RR, Passos-Monteiro E, Dos Santos IO, Kruel LFM, et al. Gait parameters of Parkinson’s disease compared with healthy controls: a systematic review and meta-analysis. Sci Rep 2021;11:752.ArticlePubMedPMCPDF
• 3. Lu C, Twedell E, Elbasher R, McCabe M, MacKinnon CD, Cooper SE. Avoiding virtual obstacles during treadmill gait in Parkinson’s disease. Front Aging Neurosci 2019;11:76.ArticlePubMedPMC
• 4. Orcioli-Silva D, Barbieri FA, Simieli L, Vitorio R, Santos PCRD, Beretta VS, et al. Walking behavior over multiple obstacles in people with Parkinson’s disease. Gait Posture 2017;58:510–515.ArticlePubMed
• 5. Cursiol JA, Dos Santos PCR, Beretta VS, Orcioli-Silva D, Simieli L, Schlenstedt C, et al. Rhythmic auditory cues improve gait asymmetry during unobstructed walking in people with Parkinson’s disease but have no effect on obstacle avoidance - AsymmGait-Parkinson study. Front Aging Neurosci 2025;17:1455432.ArticlePubMedPMC
• 6. Chen HC, Schultz AB, Ashton-Miller JA, Giordani B, Alexander NB, Guire KE. Stepping over obstacles: dividing attention impairs performance of old more than young adults. J Gerontol A Biol Sci Med Sci 1996;51:M116–M122.ArticlePubMed
• 7. Yogev-Seligmann G, Hausdorff JM, Giladi N. The role of executive function and attention in gait. Mov Disord 2008;23:329–342.quiz 472. ArticlePubMedPMCPDF
• 8. Chu NCW, Sturnieks DL, Lord SR, Menant JC. Visuospatial working memory and obstacle crossing in young and older people. Exp Brain Res 2022;240:2871–2883.ArticlePubMedPMCPDF
• 9. Anis S, Zimmerman E, Jansen AE, Kaya RD, Fernandez HH, Lopez-Lennon C, et al. Cognitive measures predict falls in Parkinson’s disease: insights from the CYCLE-II cohort. Parkinsonism Relat Disord 2025;133:107328.ArticlePubMedPMC
• 10. Maidan I, Nieuwhof F, Bernad-Elazari H, Reelick MF, Bloem BR, Giladi N, 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;30:963–971.ArticlePubMedPDF
• 11. Liu YC, Yang YR, Yeh NC, Ku PH, Lu CF, Wang RY. Multiarea brain activation and gait deterioration during a cognitive and motor dual task in individuals with Parkinson disease. J Neurol Phys Ther 2022;46:260–269.ArticlePubMed
• 12. Li N, Daie K, Svoboda K, Druckmann S. Robust neuronal dynamics in premotor cortex during motor planning. Nature 2016;532:459–464.ArticlePubMedPMCPDF
• 13. Nordin AD, Hairston WD, Ferris DP. Human electrocortical dynamics while stepping over obstacles. Sci Rep 2019;9:4693.ArticlePubMedPMCPDF
• 14. Nicoletti A, Luca A, Baschi R, Cicero CE, Mostile G, Davì M, et al. Incidence of mild cognitive impairment and dementia in Parkinson’s disease: the Parkinson’s disease cognitive impairment study. Front Aging Neurosci 2019;11:21.ArticlePubMedPMC
• 15. Kim SM, Kim DH, Yang Y, Ha SW, Han JH. Gait patterns in Parkinson’s disease with or without cognitive impairment. Dement Neurocogn Disord 2018;17:57–65.ArticlePubMedPMCPDF
• 16. Goetz CG, Fahn S, Martinez-Martin P, Poewe W, Sampaio C, Stebbins GT, et al. Movement Disorder Society-sponsored revision of the Unified Parkinson’s Disease Rating Scale (MDS-UPDRS): process, format, and clinimetric testing plan. Mov Disord 2007;22:41–47.PubMed
• 17. Jost ST, Kaldenbach MA, 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;38:1236–1252.ArticlePubMed
• 18. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975;12:189–198.PubMed
• 19. Patterson J. Multilingual aphasia examination. In: Kreutzer JS, DeLuca J, Caplan B. Encyclopedia of clinical neuropsychology. New York: Springer, New York. 2011;1674–1676.
• 20. Wechsler D. WAIS-III: Wechsler Adult Intelligence Scale. 3rd ed. San Antonio, TX: The Psychological Corporation. 1997.
• 21. Meyers JE, Meyers KR. Rey complex figure test and recognition trial: professional manual. Odessa, FL: Psychological Assessment Resources. 1995.
• 22. Litvan I, Goldman JG, Tröster AI, Schmand BA, Weintraub D, Petersen RC, et al. Diagnostic criteria for mild cognitive impairment in Parkinson’s disease: Movement Disorder Society Task Force guidelines. Mov Disord 2012;27:349–356.ArticlePubMedPMCPDF
• 23. Cheng YH, Lee SC, Chen YC, Chen JH, Yu RL, Lee WJ, et al. Normative study of the Taiwanese version of the Montreal Cognitive Assessment (MoCA) in community-dwelling individuals in Taiwan. J Chin Med Assoc 2025;88:617–623.ArticlePubMedPMC
• 24. Morris JC. The clinical dementia rating (CDR): current version and scoring rules. Neurology 1993;43:2412–2414.ArticlePubMedPMC
• 25. Wong PL, Hung CW, Yang YR, Yeh NC, Cheng SJ, Liao YY, et al. Effects of motor and cognitive complex training on obstacle walking and brain activity in people with Parkinson’s disease: a randomized controlled trial. Eur J Phys Rehabil Med 2024;60:611–620.ArticlePubMedPMC
• 26. Molavi B, Dumont GA. Wavelet-based motion artifact removal for functional near-infrared spectroscopy. Physiol Meas 2012;33:259–270.ArticlePubMed
• 27. Cope M, Delpy DT. System for long-term measurement of cerebral blood and tissue oxygenation on newborn infants by near infra-red transillumination. Med Biol Eng Comput 1988;26:289–294.ArticlePubMedPDF
• 28. Kocsis L, Herman P, Eke A. The modified Beer-Lambert law revisited. Phys Med Biol 2006;51:N91–N98.ArticlePubMed
• 29. Silva-Batista C, Liu W, Vitorio R, Stuart S, Quinn JF, Mancini M. The time course of changes in prefrontal cortex activity during walking in people with Parkinson’s disease. Neurorehabil Neural Repair 2024;38:635–645.ArticlePubMedPDF
• 30. Pottorf TS, Nocera JR, Eicholtz SP, Kesar TM. Locomotor adaptation deficits in older individuals with cognitive impairments: a pilot study. Front Neurol 2022;13:800338.ArticlePubMedPMC
• 31. Hiew S, Eibeck L, Nguemeni C, Zeller D. The influence of age and physical activity on locomotor adaptation. Brain Sci 2023;13:1266.ArticlePubMedPMC
• 32. Sasikumar S, Sorrento G, Lang AE, Strafella AP, Fasano A. Cognition affects gait adaptation after split-belt treadmill training in Parkinson’s disease. Neurobiol Dis 2023;181:106109.ArticlePubMed
• 33. Simpkins C, Brown D, Ahn J, Khalili SM, Yang F. Examination of dynamic gait stability during overground walking in people with mild cognitive impairment. J Biomech 2025;189:112836.ArticlePubMedPMC
• 34. Todorov E, Jordan MI. Optimal feedback control as a theory of motor coordination. Nat Neurosci 2002;5:1226–1235.ArticlePubMedPDF
• 35. Izawa J, Rane T, Donchin O, Shadmehr R. Motor adaptation as a process of reoptimization. J Neurosci 2008;28:2883–2891.ArticlePubMedPMC
• 36. Chen M, Pillemer S, England S, Izzetoglu M, Mahoney JR, Holtzer R. Neural correlates of obstacle negotiation in older adults: an fNIRS study. Gait Posture 2017;58:130–135.ArticlePubMedPMC
• 37. Hawkins KA, Fox EJ, Daly JJ, Rose DK, Christou EA, McGuirk TE, et al. Prefrontal over-activation during walking in people with mobility deficits: interpretation and functional implications. Hum Mov Sci 2018;59:46–55.ArticlePubMedPMC
• 38. Chatterjee SA, Seidler RD, Skinner JW, Lysne PE, Sumonthee C, Wu SS, et al. Obstacle negotiation in older adults: prefrontal activation interpreted through conceptual models of brain aging. Innov Aging 2020;4:igaa034.ArticlePubMedPMCPDF
• 39. Baek CY, Kim HD, Yoo DY, Kang KY, Lee JW. Change in activity patterns in the prefrontal cortex in different phases during the dual-task walking in older adults. J Neuroeng Rehabil 2023;20:86.ArticlePubMedPMCPDF
• 40. Lage GM, Ugrinowitsch H, Apolinário-Souza T, Vieira MM, Albuquerque MR, Benda RN. Repetition and variation in motor practice: a review of neural correlates. Neurosci Biobehav Rev 2015;57:132–141.ArticlePubMed
• 41. Carius D, Kaminski E, Clauß M, Ragert P. Dynamic alterations in cortical activation during motor adaptation in table tennis using whole‑brain fNIRS. Sci Rep 2025;15:10399.ArticlePubMedPMCPDF
• 42. Koenraadt KL, Roelofsen EG, Duysens J, Keijsers NL. Cortical control of normal gait and precision stepping: an fNIRS study. Neuroimage 2014;85(Pt 1):415–422.ArticlePubMed
• 43. Frenkel-Toledo S, Giladi N, Peretz C, Herman T, Gruendlinger L, Hausdorff JM. Effect of gait speed on gait rhythmicity in Parkinson’s disease: variability of stride time and swing time respond differently. J Neuroeng Rehabil 2005;2:23.ArticlePubMedPMCPDF
• 44. Caparrós-Manosalva C, Espinoza J, Caballero PM, da Cunha MJ, Yang F, Galen S, et al. Movement strategies during obstacle crossing in people with Parkinson disease: a systematic review with meta-analysis. PM R 2024;16:1114–1127.ArticlePubMed
• 45. Lee SM, Lee HS. Correlation between executive function and walk while crossing over an obstacle under different gait phases. Dement Neurocogn Disord 2023;22:139–147.ArticlePubMedPMCPDF
• 46. Sakurai R, Kodama K, Ozawa Y, Pieruccini-Faria F, Kobayashi-Cuya KE, Ogawa S. Association of age-related cognitive and obstacle avoidance performances. Sci Rep 2021;11:12552.ArticlePubMedPMCPDF
• 47. Pieruccini-Faria F, Hassan Haddad SM, Bray NW, Sarquis-Adamson Y, Bartha R, Montero-Odasso M. Brain structural correlates of obstacle negotiation in mild cognitive impairment: results from the gait and brain study. Gerontology 2023;69:1115–1127.ArticlePubMedPDF
• 48. Wang HY, Ren L, Yan Z, Zhou T, Liang Z. Neural basis of dysexecutive and visuospatial impairments in Parkinson’s disease with MCI: a task-based fNIRS study. NPJ Parkinsons Dis 2025;11:163.ArticlePubMedPMCPDF

Figure & Data

References

Citations

Citations to this article as recorded by  

Comments on this article