Inter-segmental coordination patterns in Parkinson’s disease are particularly disturbed during preferred walking speed: a data-driven network approach

Inter-segmental coordination patterns in Parkinson’s disease are particularly disturbed during preferred walking speed: a data-driven network approach | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Inter-segmental coordination patterns in Parkinson’s disease are particularly disturbed during preferred walking speed: a data-driven network approach Karolina Saegner, Robbin Romijnders, Inga Ruff, Julius Welzel, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7415461/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 12 Dec, 2025 Read the published version in Journal of NeuroEngineering and Rehabilitation → Version 1 posted 9 You are reading this latest preprint version Abstract Background: Human gait involves complex coordination between musculoskeletal segments. This coordination is disturbed in Parkinson's disease (PD) and likely influenced by different walking speeds. Objectives: To investigate inter-segmental coordination during different walking speeds in people with PD (pwPD) using an unconstrained and data-driven network theory approach. Methods: Twenty-nine pwPD and 29 controls walked at preferred, fast and slow speeds. Data was collected using optical motion capture. Body segment accelerations were correlated pairwise to build kinectomes for each speed and movement direction. Anatomical body segments were defined as nodes and their co-accelerations as edges to build network graphs. The kinectomes and maximum-weighted graph patterns were compared between groups. Results: Permutation testing revealed no significant kinectome differences between groups across speeds or directions. Coordination deficits in the PD group were observed predominantly at preferred walking speed (162 significantly different graph patterns) in anteroposterior and mediolateral directions. At fast walking speed, 4 significantly different graph patterns were found in anteroposterior and vertical directions. Slow walking speed showed 1 significantly different pattern in mediolateral direction. Conclusions: PD affects inter-segmental coordination, becoming most apparent at preferred walking speed. This is surprising and highly relevant, as it is the most common gait condition in real life. 'Non-preferred' walking speeds in PD exhibit more control-like patterns, which could inform future treatment studies. The direction-specific coordination deficits could provide novel insights into patho- and compensatory mechanisms in PD gait. Trial registration: The study is registered in the German Clinical Trials Register (DRKS00022998, registered on 04 Sep 2020). Neurogeriatrics gait analysis kinectome network graphs Figures Figure 1 Figure 2 Figure 3 Figure 4 1. Introduction Parkinson’s disease (PD) is the second most common ( 1 ) and the fastest growing ( 2 ) neurodegenerative disease worldwide. Typical motor symptoms include bradykinesia, resting tremor, rigidity, freezing of gait, and unstable posture ( 3 , 4 ). Furthermore, people with PD (pwPD) often have gait disturbances, such as shuffling gait, increased difficulty lifting feet from the ground, decreased stride length and velocity ( 4 , 5 ), impaired foot placement ( 6 ), and reduced range of walking speed ( 7 ). Importantly, these disturbances are associated with activity limitations during daily life ( 8 ). Typically gait of pwPD is evaluated with parameters that can be extracted from the lower extremities and the lower back, including velocity, stride and step length and width, cadence, gait phase duration, and variability and asymmetry of these metrics, as well as the range of motion of the joints of the lower extremities ( 9 , 10 ). However, for an overall view of the entire human body as an integrated system it is important to capture many simultaneous interactions arising from coordinated work of different musculoskeletal segments. This can be analysed with information coming from more than only one area of the body ( 11 ) and, for complex analyses, using the network theory ( 12 ). The latter offers a framework for investigating complex systems, consisting of many interconnected elements, and how they interact with each other. This approach allows insights into the overall behaviour of the whole system ( 12 , 13 ). Troisi Lopez et al. ( 14 ) have used principles from network theory to explore human locomotion. For that purpose, they have established a kinectome , where human gait is represented as a matrix of acceleration correlations between the body segments. The kinectome was interpreted as a network, where the nodes corresponded to body segments, and edges were defined by the correlation of acceleration signals between the segments ( 14 , 15 ). Using a variety of analysis methods, such as modularity and various topological measures, they were able to distinguish between pwPD and controls and to predict clinical symptoms in PD ( 14 ). More specifically, they showed a reduced joint coordination in pwPD compared to controls ( 16 ), and were able to demonstrate an effect of dopaminergic treatment on this coordination ( 17 ). However, to our best knowledge, how the full body is involved in walking speed-related deficits has not been adequately investigated to date. It has already been shown that rigidity contributes to reduced forward limb propulsion in PD, thus negatively affecting walking speed and step length ( 18 ). Furthermore, pwPD have difficulties in modulating their walking speed ( 7 ) and adapting to changes in walking speed ( 19 ). To address the questions about the effect of different walking speed on inter-segmental coordination, which is defined as a relationship between different body segments during movement, we used the aforementioned network theory ( 14 ) and applied it to a dataset produced in our lab ( 11 ). We hypothesized that the inter-segmental coordination during walking differs between pwPD and healthy controls in lower and upper parts of the body, with the differences depending on different walking speeds. 2. Methods 2.1 Participants, in-and exclusion criteria Participants aged 18 years or older who were able to walk independently without the use of walking aids were eligible for inclusion. Individuals were excluded if they scored below 15 on the Montreal Cognitive Assessment or if they had other movement disorders known to impact mobility, based on evaluation by a movement disorder specialist (WM). For full details regarding inclusion and exclusion criteria, see Warmerdam et al. (11). A total of 58 participants were included in the present analysis. This comprised 29 pwPD (11 females), i.e., the PD group, who performed walking tasks during their self-reported optimal "ON" medication state, assessed between 30 and 120 minutes after oral intake of levodopa. The control group consisted of 29 age-matched healthy adults (14 females), with no comorbidities affecting mobility, as confirmed by the same movement disorder specialist. Demographic and clinical characteristics are summarized in Table 1. Table 1. Demographic and clinical parameters of the participating groups Group Controls PD Total N (males/females) 29 (15/14) 29 (18/11) Age (years) 67 ± 12 67 ± 10 Height [cm] 176 ± 10 174 ± 9 Weight [kg] 79 ± 16 82 ± 18 BMI [kg/m 2 ] 26 ± 5 27 ± 5 Disease duration (years) - 9 ± 6 Hoehn & Yahr (1-5) - 3 ± 1 Medication dose (LEDD) - 713 ± 301 MDS-UPDRS III 4 ± 4 29 ± 21 The values are displayed as mean ± standard deviation. PD – Parkinson’s disease; BMI – Body Mass Index; LEDD – Levodopa equivalent daily dose; MDS-UPDRS III – Motor part of the Movement Disorder Society Unified Parkinson’s Disease Rating Scale 2.2 Determining the most and least affected sides The most and least affected upper extremities and lower extremities were determined based on the motor part of the Movement Disorder Society Unified Parkinson’s Disease Rating Scale (MDS-UPDRS III) (20) scores, namely the sum of items 3.3 (rigidity), 3.4 (finger tapping), 3.5 (hand movements), and 3.6 (pronation and supination of the hand) for left and right upper extremity, and the sum of 3.3 (rigidity), 3.7 (toe tapping), and 3.8 (leg agility) for left and right lower extremity. A difference of ≥1 point was considered as having lateralization (21) . If the MDS-UPDRS III scores between the sides were even, as well as for the control group, the non-dominant hand was taken as the most affected side due to lower motor performance on the non-dominant side (22). For participants missing the handedness data (9 PD and 6 controls), the left hand was taken as non-dominant due to handedness distribution in the general population, where around 90% of all people are right-handed (23). 2.3 Equipment, data acquisition and data dimensionality reduction All the motor tasks were recorded using a twelve-camera optical motion capture system recording with 200 Hz (Qualisys AB, Göteborg, Sweden). A total of 47 reflective markers were placed on the body (Fig. 1a). Full details on the experimental protocol can be found in Warmerdam et al. (11). For the present study, the four head markers were calculated into one midpoint, the 3-marker cluster on the sternum was calculated into one midpoint, the 4-marker clusters on the thighs and shanks were calculated into one midpoint per segment and per side, and the upper arm, forearm, and heel marker locations were excluded from further data analysis. The resulting 22 marker data (Fig. 1b), representing body segments, were taken for further analysis. 2.4 Motor assessment All study participants were instructed to walk a straight path along a 5 m long and 1 m wide walkway. There were three walking trials, namely at preferred (“Please walk at your normal walking speed”), fast (“Please walk as fast as possible, without running, falling or feeling unsafe”), and slow (“Please walk half of your normal walking speed”) speeds. The walking speed and the differences between the trials can be found Supplementary Figure 1. The beginning and end of the walkway were marked by cones with reflective markers. The participants were asked to start walking approximately 2 m before the start cones and finish walking approximately 2 m behind the end cones to make sure that steady state walking was measured within the 5 m gait trial and any variation in gait appearing due to gait initiation and termination (7,24) is excluded from the data acquisition. 2.5 Marker data pre-processing Marker data were prepared in a format following the Brain Imaging Data Structure (BIDS) (25,26) and were pre-processed using Python (Python Software Foundation. Python Language Reference, version 3.11.2). The walking data were trimmed to only contain the samples collected between the start and end cones. These data were then cut so each gait acquisition contains one full left and one full right gait cycle (Fig. 1c). Missing values were linearly interpolated using the fill_missing_samples function from Kinetics Toolkit (27) with the threshold of 271 samples (1.35 s). The remaining data were filtered using a 2 nd order Butterworth filter with the cut-off frequency of 6 Hz. Next, the data were rotated using Principal Component Analysis to align the x direction of the markers with the anteroposterior (AP) movement direction. Lastly, the data, originally containing positions, was differentiated twice to get the acceleration of the markers. 2.6 Building the kinectome Time series containing acceleration data of each marker during each gait acquisition in the AP, mediolateral (ML), and vertical (V) directions were first normalised to 500 samples, and then were used to build person, trial and gait acquisition specific kinectomes. The kinectomes were built in each movement direction separately by correlating the data from each marker with that from every other marker, resulting in a mirrored 22✕22✕3 correlation matrix (where 22 is the number of body segments, and 3 is the number of movement directions, Fig. 1d). This matrix will be referred to as the kinectome from now on (14). In this paper we present the results obtained using Pearson's correlation. Then for each person an average kinectome was built by averaging the kinectomes from multiple gait kinectomes obtained from the respective trial in a respective movement direction (e.g., averaging kinectomes one, two, and three in Fig. 1c). For each person standard deviation of the kinectomes was obtained by calculating the standard deviation between all gait kinectomes (Fig. 1c) for each trial and movement direction separately. This resulted in each person having one person-specific average and one person-specific standard deviation kinectome per trial and direction (total of 9 average and 9 standard deviation kinectomes per person). 2.7 Group kinectome analysis The individual average and standard deviation kinectomes were used to build group-specific average and standard deviation kinectomes, respectively, for each walking speed and each movement direction separately for the control and PD groups, by averaging all individual kinectomes between the people within the groups. The latter kinectomes were used for permutation analysis. 2.8 Representing the kinectome as a graph Each person-specific average kinectome was represented as a graph (Fig. 1e) for further analysis. The body segments were represented as nodes, and the acceleration correlations between them as edges between the nodes. These graphs were non-directional, i.e., there was no particular direction of the edges between the nodes; and complete, i.e., each node was connected with each other node, based on the assumption that motor coordination involves relationships between anatomically distant joints (e.g. contralateral limb swing during gait) (28). Network analysis was done using NetworkX package (Version 3.4.2) (29) for Python. 2.9 Patterns A pattern of a graph refers to a specific arrangement of its nodes and edges (Fig. 1f). Here we analysed path patterns, i.e., sequences of connected nodes and edges. In this paper, a path refers to a sequence of nodes connected by edges, and a pattern is a path represented in the kinematic data. We have looked at the maximum-weighted patterns in the respective person-specific average kinectome that started from a given node and had a fixed length. This approach provided an unconstrained and data-driven view on respective comparisons. These comparisons were done on three arbitrarily defined levels, namely local (2-6 body segments in the pattern), referring to simple coordination networks, multi-segmental (7-14 body segments) – more complex networks, and whole-body (15+ body segments) – global coordination networks. The maximum-weighted patterns were chosen for the analysis because they provide the baseline for understanding coordination differences between the groups. The algorithm src.graph_utils.kinectome2pattern.strongest_pattern_subgraph (available on (30)) starts at a given node and constructs a path by iteratively selecting edges with the highest absolute weight while avoiding node repetition. This ensures that each subsequent step in the path represents the strongest local connection available from the current position, creating a path that follows the most prominent signal through the network structure. Note that while the starting node is fixed, there is no predetermined target node—only the path length is constrained. Since there are 22 bone segments taken for the analysis, the algorithm looped over each body segment and took it as a starting node. Pattern lengths ranging from 2 to 20 nodes were considered. First, the strongest pattern for each individual was found. Then, within each group, patterns of the same length and starting node were averaged to determine the strongest group-specific pattern. Finally, the values of this pattern were then taken from each individual in both groups and compared statistically. This strategy enabled an exploratory analysis of emerging patterns; rather than specifying a particular pattern, it allowed the algorithm to first identify the strongest pattern and then assess the differences between these patterns across groups. 2.10 Statistics Statistical analysis was done using SciPy (31) for Python and JASP (JASP Team (2023). JASP (Version 0.18)). The characteristics of the groups are presented as mean ± standard deviation. Normality distribution was checked using the Shapiro-Wilk test. Outliers were removed using the Z-score method (removing values three standard deviations above or below the group mean) for normally distributed data, or by calculating the interquartile range and removing values 1.5 times greater or less than the interquartile range for non-normally distributed data. Chi-square test was used to compare the gender distribution between the groups. To compare all other variables between the groups, t-test and Mann-Whitney U tests were used for normally and non-normally distributed data, respectively. The group-specific average and standard deviation kinectomes were compared through permutation testing, by randomly shuffling the control group kinectomes 5000 times. Spearman’s rho was used to generate the distribution of the differences that are to be expected by chance alone (32,33). To assess the stability of the kinectomes across sample sizes, we performed bootstrap resampling (n=5000) at varying percentages (10-90%) of the original dataset. For each bootstrap iteration, kinectomes for each walking speed and movement direction were computed using the subset of the full sample, and permutation testing was applied to these subsets to build a bootstrap distribution of correlations. There were two pattern analyses done, namely strongest patterns found in the control group were compared to the same patterns in the PD group, and strongest patterns found in the PD group were compared to those in the control group. The patterns derived from the control group are presented in the result section, and the results comparing PD group’s patterns can be found in Supplementary Table 1. Patterns with an overlap of 80% as determined by Jaccard index (34) were seen as the same pattern. Higher values of the pattern indicate stronger correlation between the body segments of that particular pattern. The patterns between the groups were compared using t-test or Mann-Whitney U test for normally and non-normally distributed data, respectively. The effect size was calculated using Cohen’s d. Effect size measure was interpreted as small, medium, or large if the effect size was 0.2<x<0.5, 0.5<x<0.8, and x≥0.8, respectively (35). To adjust for multiple comparisons when comparing the patterns of the same length, but different starting nodes, a Bonferroni correction was used (n=22). All comparisons were made between the groups within the same walking speed and same movement direction. 3. Results There were no significant differences between the groups in gender distribution, age, height, weight, or body mass index ( p > .05). The MDS-UPDRS III scores differed between the control and PD groups, with the latter group having higher scores. Details are shown in Table 1. 3.1 Comparing the group-specific kinectomes Average and standard deviation kinectomes’ comparisons between the PD and control groups using permutation testing did not reveal any significant differences. The correlations between the average and between the standard deviation kinectomes (correlation values are displayed in Table 2) are unlikely to have occurred by chance, and the observed correlation is stronger than what could be expected by random chance. Table 2. Spearman’s rho values between the group-specific kinectomes for preferred, fast, and slow walking speeds Spearman’s rho Preferred Fast Slow AP ML V AP ML V AP ML V Average kinectomes .99 .93 .97 .98 .91 .92 .98 .94 .98 Standard deviation kinectomes .73 .63 .83 .88 .61 .66 .88 .74 .80 Spearman’s rho values between control and PD standard deviation kinectomes were consistently lower than those for the average kinectomes across all walking speeds and directions. A visual representation of a standard deviation kinectomes can be found in Figure 2. Moreover, the standard deviation values were consistently higher in the PD group than in the control group. These findings indicate that while the overall structure of the inter-segmental coordination is similar between the groups, the variability in this inter-segmental coordination is higher in PD group than in controls. 3.1.1 Bootstrapping Bootstrap analysis demonstrated high correlation stability across all conditions. The correlations approached asymptotic values by 40-60% sample size for average and 80-90% for standard deviation kinectomes. Using 80% of the original sample for the subset size, the observed correlations exceeded bootstrap means (Supplementary Fig. 3 and 4). 3.2 Graph patterns Pattern analysis revealed distinct differences in movement coordination patterns between PD and control groups across movements in AP, ML, and V directions mainly during preferred and fast walking, while slow walking had only one significantly different pattern. Full patterns can be found in Supplementary Table 2. The results consistently show altered inter-segmental coordination in the PD group across various organizational levels (local, multi-segmental, and whole-body). Importantly, the manifestation and characteristics of these deficits varied with walking speed. Preferred walking speed At preferred walking speed, the inter-segmental coordination deficits of the PD group were most apparent (Fig. 3). The reduction in coordination was consistent and multi-level in both the AP and ML directions. In the following, the main findings are summarized: AP direction: the PD group showed weaker coordination than controls across all organizational levels, with the most significantly different patterns observed in this speed and direction (n=146): On a local level (n=26), patterns mostly involving axial segments (head, trunk, pelvis), both upper limbs, and the less affected lower limb, were weaker in the PD group (effect sizes: 0.53-1.86, mean=0.8). This suggests deficits in basic postural coordination. On a multi-segmental level (n=80), patterns involving multiple body regions and including contralateral limb movement were weaker in the PD group (effect sizes: 0.53-1.67, mean=0.78). These findings suggest PD-related deficits in simultaneous multi-segmental coordination during walking. On a whole-body level (n=40), patterns involving most body segments showed the most pronounced deficits in the PD group (effect sizes: 0.53-1.35, mean=0.7). This suggests that pwPD cannot effectively coordinate whole-body inter-segmental movements during walking. ML direction: Weaker inter-segmental coordination was observed across all organizational levels in the PD group compared to controls, although notably fewer patterns were significantly different (n=16) compared to the AP direction. On a local level (n=5), patterns involving sternum, the less affected shoulder and thigh, as well as between the most affected upper limb and several segments in both most and least affected lower limbs, were weaker in the PD group (effect sizes: 0.54-1.59, mean=1.33). On a multi-segmental level (n=7), patterns primarily including the pelvis, and upper and lower limbs, were weaker in the PD group (effect sizes: 0.54-1.59, mean=1.33). This suggests impaired synchronization between the more and less affected body sides during walking. On a whole-body level (n=4), patterns involving most body segments showed the largest deficits in the PD group (effect sizes: 1-1.04, mean=1.02). This suggests that whole-body coordination in the ML direction, crucial for stability during walking, is affected by PD. V direction: No significant differences were observed. Fast Walking Speed Fewer and distinct significant differences between the PD group and controls were observed compared to preferred speed, and they were observed exclusively on a multi-segmental level: AP direction: Two patterns on a multi-segmental level were found to be significantly weaker in the PD group (effect sizes: 0.51-1, mean=0.76) (Fig. 4, AP direction, subplot b). The patterns involved core segments, both upper limbs and the less affected lower limb, suggesting an impaired synchronization between the core and limb swing during walking. ML direction: No significant differences. V Direction: Two patterns on a multi-segmental level were found to be significantly different (Fig. 4, V direction, subplot b). The patterns involved sternum, most affected side of the pelvis and both upper and lower limbs. Interestingly, these values were higher in the PD group (effect sizes: 0.96-99, mean=0.98). This suggestsuggesting compensatory activation of muscles and potentially also increased rigidity within complex networks mainly including upper and lower limbs, as well as pelvic segments. Slow Walking Speed At slow walking speed, minimal significant differences in inter-segmental coordination were observed between the groups: AP and V direction: No significant differences were found. ML direction: One pattern was found to be different in the slow walking speed (Supplementary Figure 2, ML direction, subplot a), involving the less affected shoulder and thigh, with an effect size of 1.17. 4. Discussion This study describes the inter-segmental coordination of the whole body during walking in pwPD and age-matched controls, and how this coordination changes at different walking speeds. The coordination between the segments was evaluated by the means of kinectomes, containing pairwise acceleration interactions between different body segments. Analysis methods were based on the network theory to evaluate the inter-segmental coordination differences between pwPD and controls ( 14 ). The high correlations in the average kinectomes suggest that the fundamental inter-segmental coordination structure remains intact in PD. This result reflects our clinical experience: pwPD are able to propel themselves forward during their entire mobile phase of the disease, while maintaining the swing of contralateral limbs. As shown with our collected data ( 11 ) (Supplementary Fig. 1), as well as previous research, pwPD are also able to modulate their walking speed ( 36 ), although to a lesser extent than healthy older adults ( 7 ). However, movement variability during gait, as seen in reduced correlations in the standard deviation kinectomes, as well as their overall higher standard deviation in the PD group, was increased especially during the preferred walking speed. Increased variability in gait of pwPD has been previously shown by multiple studies, however they mostly investigated discrete spatial and temporal gait parameters ( 37 ) like swing time ( 36 , 38 ), stride time ( 36 , 39 ), stride length ( 40 ), and stride-to-stride variability ( 41 ). The kinectome allows us to look at the inter-segmental coordination (and the variability thereof) by analysing simultaneous kinematic relationships across multiple body segments, as well as all three movement directions. This approach helps to dive deeper into pathological mechanisms during gait, but also to evaluate potential adaptability and compensatory mechanisms, providing insights for allied therapeutic health interventions ( 42 , 43 ). The analysis of strongest inter-segmental coordination patterns revealed the predominance of coordination deficits at preferred walking speed, suggesting that pwPD experience impaired motor control during their most comfortable walking speed, which is typically the most often used and most efficient for healthy people ( 44 , 45 ). Our results therefore suggest that in pwPD, the motor coordination system is more challenged during preferred than fast and slow walking speed. This consistent and multi-level (starting from pairs of two body segments and ending at patterns involving all body segments) deficit of inter-segmental coordination strength indicates fundamental impairments in integrating movements, which are crucial for dynamic stability ( 39 ) and overall gait execution. These coordination deficits likely reflect the underlying pathophysiology of PD where the degeneration of dopaminergic neurons affects the whole basal ganglia network ( 46 ), which is responsible for selecting optimal motor strategies ( 47 ). In addition, the reduced coordination between the segments in pwPD, compared to controls, could make gait less efficient and require increased effort, contributing to further symptoms well known to occur in PD: fatigue ( 48 ) and reduced walking endurance ( 49 ). Furthermore, the motor coordination deficits were most pronounced in the AP and ML directions. Deficits in the axial segments, especially in the AP direction, suggest issues with core postural control during ambulation, which is a known difficulty in pwPD ( 50 ). Our data show that, in AP direction, the hip segments are centrally involved in significantly weaker patterns pwPD, suggesting that especially interventions targeting pelvic coordination (and its coordinated “interaction” with distant body parts) could help these people to keep their gait stable during forward movement. Furthermore, the decreased limb coordination in ML direction that we observed in our PD group, point to the presence of lateral stability impairments during gait in this disease. This is of particular relevance, as lateral stability deficits in pwPD have been linked to falls ( 51 , 52 ). Interestingly, the fast walking speed revealed a completely different picture of gait coordination pattern in pwPD. The patterns showed a shift from severely impaired gait coordination, pronounced in the AP and ML direction, as observed in the preferred walking condition, towards generally much “better” gait coordination during the fast walking condition. Previous studies have shown that, in pwPD, the variabilities of stride length and stride time are lower during fast walking than during preferred walking ( 53 ), bringing the gait pattern of pwPD during fast walking closer to that of healthy controls. We have observed similar results. During the fast walking condition, only two patterns in the AP direction and none in the ML direction were significantly weaker in the PD group than in the control group, as opposed to 146 significant patterns in the AP direction and 16 in the ML direction during preferred speed (Supplementary Table 2). Why did we observe most differences of gait coordination between PD and controls in preferred walking speed? One explanation could be that walking at a speed above the “comfort zone” improves the arousal state ( 54 ) and with that resources to coordinate gait. Previously it was shown that a known phenomenon in pwPD called “paradoxical kinesis” occurs in situations when a person is highly influenced by external cues or higher arousal ( 54 , 55 ) and is able to produce movements like healthy people. Another reason could be that higher walking speeds lead to a proportionally reduced energy consumption ( 45 ), potentially making brain reserves free for gait coordination. Moreover, PD is known to be a disease which affects automatic movements – pwPD lose previously acquired automatic skills and have difficulty restoring them ( 56 ). Importantly, these automatic movements are associated with posterior putamen, a region in basal ganglia which is primarily affected by PD due to the loss of dopamine ( 57 ). Preferred walking speed can be seen as automatic protocol for walking, and may therefore be particularly affected by PD. Fast walking speed, as assessed in this study, could have served as the cue to move in a non-automated manner and demonstrate more control-like performance. Why pwPD, based on the pattern analysis, showed an even higher inter-limb coordination than controls in the fast walking condition in the V direction, is not yet clear. We hypothesize that, to maintain stability and forward propulsion at higher walking speed, pwPD may stiffen multiple segments and put them to work together, sacrificing fluid, relaxed movements to increase stability and meet the more demanding walking speed requirements. Future studies should confirm our results and, if this is the case, investigate whether this high interlimb coordination in the V axis indeed reflects better coordination, or whether this adaptation comes with a loss of adaptability and thus an increased risk of negative consequences, such as falls. The differences in gait coordination were least significant in the slow walking condition. We hypothesize that, at slow walking speed, the motor system has sufficient time to plan and execute movements, bringing the movement pattern closer to that observed in healthy adults. It is already known that, in order to compensate for reduced dynamic stability during walking, pwPD adopt strategies with shorter step length ( 58 ), which could have allowed them choosing a control-like inter-segmental integration on a whole-body level. There are a few considerations worth mentioning regarding the pattern analysis. We have chosen to compare the maximum-weighted patterns between the groups as baseline, however future work examining weaker patterns could reveal compensatory mechanisms, fatigue, less optimal motor strategies, or subtle early stage changes during the disease progression. Furthermore, a sub-analysis of the patterns themselves could identify where do the strongest group patterns rank on an individual list, and if they are significantly stronger than the other possible patterns. In addition, future studies should inlcude measures of varability (e.g. by bootstrapping) to understand how generalizable the results from such high-level analysis are. This study has several strengths and limitations. Among the strengths are the evaluation of inter-segmental coordination of the entire body, different walking speeds, and including bootstrapping to assess the stability of the kinectomes. The limitations include the evaluation of the PD group only after intake of medication. Different coordination patterns may be revealed in the off-medication state. Other limitations include high disease variability, the focus of relatively short bouts of straight walking, and a relatively high proportion of pwPD without lateralization. Furthermore, the data was collected in a clinical setting, namely a movement analysis laboratory, with no obstacles and no disturbances, which certainly occur in usual daily lives. 5. Conclusion In conclusion, our results from a data-driven network analysis indicate that inter-segmental coordination of gait at preferred walking speed is severely affected in pwPD. We hypothesise that this alteration is due to deficits in automaticity and arousal networks. Furthermore, the distribution of the patterns that are significantly different from controls provide insights into the pathomechanistic aspects of body control during walking in pwPD, and potentially interesting treatment options. Interestingly, these coordination deficits were much less obvious during fast and slow walking, respectively. These findings support the implementation of different walking speeds in mobility training in pwPD. Our direction-specific analyses suggest that exercises which help in improving coordination in AP and ML directions, including contralateral limb swing and pelvic segmental coordination, have the most potential in addressing inter-limb coordination deficits in PD. Declarations Electronic supplementary material Available. Ethics approval and consent to participate The study was approved by the ethics committee of the Medical Faculty of Kiel University (D438/18) and was conducted in accordance with the principles of the Declaration of Helsinki. The study is registered in the German Clinical Trials Register (DRKS00022998, registered on 04 Sep 2020). All study participants have read and signed an informed consent prior to the study. Consent for publication All authors have read and agreed to the final version of the manuscript. Availability of data and materials Part of the data analysed during the current study are available at (59). The remaining data contains patient information and are available from the corresponding author on reasonable request. The data analysis code is available at (30). It is platform independent, written in Python programming language. Requirements include Python version 3.11 or higher. Competing interests None declared Funding Not applicable. Author‘s contributions Validation (1); formal analysis (2); investigation (3); resources (4); data curation (5); writing—original draft preparation (6); writing—review and editing (7); visualization (8); supervision (9). K.S.: 1, 2, 3, 6, 7, 8; R.R.: 2, 3, 7; I.R.: 7; J.W.: 3, 7; C.H.: 1, 7, 9; E.W.: 5, 7; P.C.: 2, 3, 7; W.M.: 4, 7, 9. All authors have read and agreed to the final version of the manuscript. Acknowledgements Landesprojekt LPW21-E/1.2.2/179. References Zhu J, Cui Y, Zhang J, Yan R, Su D, Zhao D, et al. Temporal trends in the prevalence of Parkinson’s disease from 1980 to 2023: a systematic review and meta-analysis. The Lancet Healthy Longevity. 2024 Jul 1;5(7):e464–79. Feigin VL, Abajobir AA, Abate KH, Abd-Allah F, Abdulle AM, Abera SF, et al. Global, regional, and national burden of neurological disorders during 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015. The Lancet Neurology. 2017 Nov 1;16(11):877–97. Bloem BR, Okun MS, Klein C. Parkinson’s disease. The Lancet. 2021 Jun 12;397(10291):2284–303. Moustafa AA, Chakravarthy S, Phillips JR, Gupta A, Keri S, Polner B, et al. Motor symptoms in Parkinson’s disease: A unified framework. Neurosci Biobehav Rev. 2016 Sep;68:727–40. Stuopelytė A, Šakalienė R. Gait Training Methods and ChangesiIn Gait Parameters in Parkinson`s Disease (Literature Review). Reabilitacijos mokslai: slauga, kineziterapija, ergoterapija [Internet]. 2012 Dec 27 [cited 2025 May 23];2(7). Available from: https://journals.lsu.lt/reabilitacijos-mokslai/article/view/857 Lang C, Hausdorff JM, Bruijn SM, Brodie MA, Okubo Y, Maetzler W, et al. Foot placement coordination is impaired in people with Parkinson’s Disease [Internet]. bioRxiv; 2025 [cited 2025 Aug 6]. p. 2025.06.11.659042. Available from: https://www.biorxiv.org/content/10.1101/2025.06.11.659042v1 von der Recke F, Warmerdam E, Hansen C, Romijnders R, Maetzler W. Reduced Range of Gait Speed: A Parkinson’s Disease-Specific Symptom? Journal of Parkinson’s Disease. 2023 Jan 1;13(2):197–202. Lang C, van Dieen JH, Brodie MA, Welzel J, Maetzler W, Singh NB, et al. Complexities and challenges of translating intervention success to real world gait in people with Parkinson’s disease. Front Neurol [Internet]. 2024 Oct 9 [cited 2025 Aug 6];15. Available from: https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2024.1455692/full 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. Journal of Parkinson’s Disease. 2020 Jul 28;10(3):843–53. di Biase L, Di Santo A, Caminiti ML, De Liso A, Shah SA, Ricci L, et al. Gait Analysis in Parkinson’s Disease: An Overview of the Most Accurate Markers for Diagnosis and Symptoms Monitoring. Sensors (Basel). 2020 Jun 22;20(12):3529. Warmerdam E, Romijnders R, Geritz J, Elshehabi M, Maetzler C, Otto JC, et al. Proposed Mobility Assessments with Simultaneous Full-Body Inertial Measurement Units and Optical Motion Capture in Healthy Adults and Neurological Patients for Future Validation Studies: Study Protocol. Sensors. 2021 Jan;21(17):5833. Barabási AL. Network science. Philos Trans A Math Phys Eng Sci. 2013 Mar 28;371(1987):20120375. Boonstra TW, Danna-Dos-Santos A, Xie HB, Roerdink M, Stins JF, Breakspear M. Muscle networks: Connectivity analysis of EMG activity during postural control. Sci Rep. 2015 Dec 4;5:17830. Troisi Lopez E, Sorrentino P, Liparoti M, Minino R, Polverino A, Romano A, et al. The kinectome: A comprehensive kinematic map of human motion in health and disease. Ann N Y Acad Sci. 2022 Oct;1516(1):247–61. Newman MEJ. Modularity and community structure in networks. Proceedings of the National Academy of Sciences. 2006 Jun 6;103(23):8577–82. Troisi Lopez E, Liparoti M, Minino R, Romano A, Polverino A, Carotenuto A, et al. Kinematic network of joint motion provides insight on gait coordination: An observational study on Parkinson’s disease. Heliyon. 2024 Aug 15;10(15):e35751. Romano A, Liparoti M, Minino R, Polverino A, Cipriano L, Carotenuto A, et al. The effect of dopaminergic treatment on whole body kinematics explored through network theory. Sci Rep. 2024 Jan 22;14(1):1913. 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 Apr 6;32(1):28–34. Saegner K, Romijnders R, Hansen C, Holder J, Warmerdam E, Maetzler W. Inter-joint coordination with and without dopaminergic medication in Parkinson’s disease: a case-control study. J Neuroeng Rehabil. 2024 Jul 13;21:118. Goetz CG, Tilley BC, Shaftman SR, Stebbins GT, Fahn S, Martinez-Martin P, et al. Movement Disorder Society-sponsored revision of the Unified Parkinson’s Disease Rating Scale (MDS-UPDRS): Scale presentation and clinimetric testing results: MDS-UPDRS: Clinimetric Assessment. Mov Disord. 2008 Nov 15;23(15):2129–70. Orcioli-Silva D, Barbieri FA, dos Santos PCR, Beretta VS, Simieli L, Vitorio R, et al. Double obstacles increase gait asymmetry during obstacle crossing in people with Parkinson’s disease and healthy older adults: A pilot study. Sci Rep. 2020 Feb 10;10:2272. Przybyla A, Good DC, Sainburg RL. Dynamic dominance varies with handedness: reduced interlimb asymmetries in left-handers. Exp Brain Res. 2012 Feb 1;216(3):419–31. Papadatou-Pastou M, Ntolka E, Schmitz J, Martin M, Munafò MR, Ocklenburg S, et al. Human handedness: A meta-analysis. Psychol Bull. 2020 Jun;146(6):481–524. Lin CC, Creath RA, Rogers MW. Variability of Anticipatory Postural Adjustments During Gait Initiation in Individuals With Parkinson Disease. J Neurol Phys Ther. 2016 Jan;40(1):40–6. Gorgolewski KJ, Auer T, Calhoun VD, Craddock RC, Das S, Duff EP, et al. The brain imaging data structure, a format for organizing and describing outputs of neuroimaging experiments. Sci Data. 2016 Jun 21;3(1):160044. Jeung S, Cockx H, Appelhoff S, Berg T, Gramann K, Grothkopp S, et al. Motion-BIDS: an extension to the brain imaging data structure to organize motion data for reproducible research. Sci Data. 2024 Jul 2;11(1):716. Chénier F. Kinetics Toolkit: An Open-Source Python Package to Facilitate Research in Biomechanics. Journal of Open Source Software. 2021 Oct 23;6(66):3714. Dietz V, Michel J. Human bipeds use quadrupedal coordination during locomotion. Ann N Y Acad Sci. 2009 May;1164:97–103. Hagberg AA, Schult DA, Swart PJ. Exploring Network Structure, Dynamics, and Function using NetworkX. In: Varoquaux G, Vaught T, Millman J, editors. Proceedings of the 7th Python in Science Conference. Pasadena, CA USA; 2008. p. 11–5. Saegner K, Romijnders R, Conceição P. pykinectome [Internet]. Neurogeriatrics Kiel; Available from: https://github.com/neurogeriatricskiel/pykinectome 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 Mar;17(3):261–72. Nichols TE, Holmes AP. Nonparametric permutation tests for functional neuroimaging: A primer with examples. Hum Brain Mapp. 2001 Oct 15;15(1):1–25. Milewski JPC& TM. 13 Permutation Testing | PSY317L Guidebook [Internet]. [cited 2025 Jul 15]. Available from: https://bookdown.org/curleyjp0/psy317l_guides5/permutation-testing.html Prokopenko D, Hecker J, Silverman EK, Pagano M, Nöthen MM, Dina C, et al. Utilizing the Jaccard index to reveal population stratification in sequencing data: a simulation study and an application to the 1000 Genomes Project. Bioinformatics. 2016 May 1;32(9):1366–72. Peterson SJ, Foley S. Clinician’s Guide to Understanding Effect Size, Alpha Level, Power, and Sample Size. Nutrition in Clinical Practice. 2021;36(3):598–605. 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 NeuroEngineering Rehabil. 2005 Jul 31;2(1):23. König N, Singh NB, Baumann CR, Taylor WR. Can Gait Signatures Provide Quantitative Measures for Aiding Clinical Decision-Making? A Systematic Meta-Analysis of Gait Variability Behavior in Patients with Parkinson’s Disease. Front Hum Neurosci [Internet]. 2016 Jun 30 [cited 2025 Jun 13];10. Available from: https://www.frontiersin.org/journals/human-neuroscience/articles/10.3389/fnhum.2016.00319/full Roemmich RT, Nocera JR, Vallabhajosula S, Amano S, Naugle KM, Stegemöller EL, et al. Spatiotemporal variability during gait initiation in Parkinson’s disease. Gait & Posture. 2012 Jul 1;36(3):340–3. Lowry KA, Smiley-Oyen AL, Carrel AJ, Kerr JP. Walking stability using harmonic ratios in Parkinson’s disease. Mov Disord. 2009 Jan 30;24(2):261–7. Rochester L, Yarnall AJ, Baker MR, David RV, Lord S, Galna B, et al. Cholinergic dysfunction contributes to gait disturbance in early Parkinson’s disease. Brain. 2012 Sep;135(Pt 9):2779–88. Hausdorff JM, Cudkowicz ME, Firtion R, Wei JY, Goldberger AL. Gait variability and basal ganglia disorders: Stride-to-stride variations of gait cycle timing in parkinson’s disease and Huntington’s disease. Movement Disorders. 1998;13(3):428–37. Kleiner AFR, Souza Pagnussat A, Pinto C, Redivo Marchese R, Salazar AP, Galli M. Automated Mechanical Peripheral Stimulation Effects on Gait Variability in Individuals With Parkinson Disease and Freezing of Gait: A Double-Blind, Randomized Controlled Trial. Archives of Physical Medicine and Rehabilitation. 2018 Dec 1;99(12):2420–9. Frenkel-Toledo S, Giladi N, Peretz C, Herman T, Gruendlinger L, Hausdorff JM. Treadmill walking as an external pacemaker to improve gait rhythm and stability in Parkinson’s disease. Movement Disorders. 2005;20(9):1109–14. Majed L, Ibrahim R, Lock MJ, Jabbour G. Walking around the preferred speed: examination of metabolic, perceptual, spatiotemporal and stability parameters. Front Physiol [Internet]. 2024 Feb 9 [cited 2025 Jun 13];15. Available from: https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2024.1357172/full Padmanabhan P, Sreekanth Rao K, Gonzalez AJ, Pantelyat AY, Chib VS, Roemmich RT. The Cost of Gait Slowness: Can Persons with Parkinson’s Disease Save Energy by Walking Faster? J Parkinsons Dis. 2021;11(4):2073–84. Blandini F, Nappi G, Tassorelli C, Martignoni E. Functional changes of the basal ganglia circuitry in Parkinson’s disease. Prog Neurobiol. 2000 Sep;62(1):63–88. Groenewegen HJ. The Basal Ganglia and Motor Control. Neural Plasticity. 2003;10(1–2):108384. Christiansen CL, Schenkman ML, McFann K, Wolfe P, Kohrt WM. Walking economy in people with Parkinson’s disease. Movement Disorders. 2009;24(10):1481–7. Schenkman M, Hall D, Kumar R, Kohrt WM. Endurance Exercise Training to Improve Economy of Movement of People With Parkinson Disease: Three Case Reports. Physical Therapy. 2008 Jan 1;88(1):63–76. Mancini M, Afshari M, Almeida Q, Amundsen-Huffmaster S, Balfany K, Camicioli R, et al. Digital gait biomarkers in Parkinson’s disease: susceptibility/risk, progression, response to exercise, and prognosis. NPJ Parkinsons Dis. 2025 Mar 21;11:51. King LA, Horak FB. Lateral Stepping for Postural Correction in Parkinson’s Disease. Archives of Physical Medicine and Rehabilitation. 2008 Mar 1;89(3):492–9. Wegen EEH van, Emmerik REA van, Wagenaar RC, Ellis T. Stability Boundaries and Lateral Postural Control in Parkinson’s Disease. Motor Control. 2001 Jul 1;5(3):254–69. Bryant MS, Rintala DH, Hou JG, Collins RL, Protas EJ. Gait variability in Parkinson’s disease: levodopa and walking direction. Acta Neurol Scand. 2016 Jul;134(1):83–6. Tosserams A, Bloem BR, Ehgoetz Martens KA, Helmich RC, Kessels RPC, Shine JM, et al. Modulating arousal to overcome gait impairments in Parkinson’s disease: how the noradrenergic system may act as a double-edged sword. Transl Neurodegener. 2023 Mar 26;12:15. McDonald LM, Griffin HJ, Angeli A, Torkamani M, Georgiev D, Jahanshahi M. Motivational Modulation of Self-Initiated and Externally Triggered Movement Speed Induced by Threat of Shock: Experimental Evidence for Paradoxical Kinesis in Parkinson’s Disease. PLoS One. 2015;10(8):e0135149. Wu T, Hallett M, Chan P. Motor automaticity in Parkinson’s disease. Neurobiol Dis. 2015 Oct;82:226–34. Redgrave P, Rodriguez M, Smith Y, Rodriguez-Oroz MC, Lehericy S, Bergman H, et al. Goal-directed and habitual control in the basal ganglia: implications for Parkinson’s disease. Nat Rev Neurosci. 2010 Nov;11(11):760–72. Ban R, Ahn J, Simpkins C, Lazarus J, Yang F. Dynamic gait stability in people with mild to moderate Parkinson’s disease. Clinical Biomechanics. 2024 Aug 1;118:106316. Warmerdam E, Hansen C, Romijnders R, Hobert MA, Welzel J, Maetzler W. Full-Body Mobility Data to Validate Inertial Measurement Unit Algorithms in Healthy and Neurological Cohorts. Data. 2022 Oct;7(10):136. Additional Declarations No competing interests reported. Supplementary Files Supplementarymaterialkinectome.docx Cite Share Download PDF Status: Published Journal Publication published 12 Dec, 2025 Read the published version in Journal of NeuroEngineering and Rehabilitation → Version 1 posted Editorial decision: Revision requested 29 Sep, 2025 Reviews received at journal 26 Sep, 2025 Reviews received at journal 23 Sep, 2025 Reviewers agreed at journal 05 Sep, 2025 Reviewers agreed at journal 03 Sep, 2025 Reviewers invited by journal 03 Sep, 2025 Editor assigned by journal 22 Aug, 2025 Submission checks completed at journal 22 Aug, 2025 First submitted to journal 20 Aug, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7415461","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":510970544,"identity":"4a8b7cc6-c1b3-4c6c-a1bc-1ac93d92d39e","order_by":0,"name":"Karolina Saegner","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABD0lEQVRIie2Pv0rEQBCHJwRSBbbdsHj3CnMsBI7zz6usBHabAw9s7AwIe6VtXsNGLAMLXnNim8IiaVJ7BFFR8PYUJRarVwrux8wWw3z7YwA8nj8I3Tw19iYEIghhUz8poq8k+VZKf4LlL0oyv2lqMbsbAjFN83C1e8QrmXUzmAxcCosVR4HtKKeS852lHF9W8poVoLhLGYCMqEAT5BRSlmiDaaV0GIM5zF0Kad+Vg5wsHl8S/Ya8UGedVU5dCqMfKfbPaRqsdIlIZcmsIhwGJEUb2ltMpun0mAU6Q7psJYtRjVwp9FYG9f2r2Tsni4vVs95HMpe8i08mQ1fKF5Fte/Un6N78RvC05aLH4/H8L9ap61A9ZQisFwAAAABJRU5ErkJggg==","orcid":"","institution":"University Hospital Schleswig-Holstein, Campus Kiel and Kiel University","correspondingAuthor":true,"prefix":"","firstName":"Karolina","middleName":"","lastName":"Saegner","suffix":""},{"id":510970545,"identity":"85388a35-7ca5-4355-92bc-8f8d92e639ad","order_by":1,"name":"Robbin Romijnders","email":"","orcid":"","institution":"University Hospital Schleswig-Holstein, Campus Kiel and Kiel University","correspondingAuthor":false,"prefix":"","firstName":"Robbin","middleName":"","lastName":"Romijnders","suffix":""},{"id":510970546,"identity":"20509daf-ba77-4bbb-9c4f-d878474a96a0","order_by":2,"name":"Inga Ruff","email":"","orcid":"","institution":"University Hospital Schleswig-Holstein, Campus Kiel and Kiel University","correspondingAuthor":false,"prefix":"","firstName":"Inga","middleName":"","lastName":"Ruff","suffix":""},{"id":510970547,"identity":"4f4f2699-dbfc-47f1-922d-831e4c959058","order_by":3,"name":"Julius Welzel","email":"","orcid":"","institution":"University Hospital Schleswig-Holstein, Campus Kiel and Kiel University","correspondingAuthor":false,"prefix":"","firstName":"Julius","middleName":"","lastName":"Welzel","suffix":""},{"id":510970548,"identity":"2d59d110-428f-4600-bab1-05e42f1e9e0d","order_by":4,"name":"Clint Hansen","email":"","orcid":"","institution":"University Hospital Schleswig-Holstein, Campus Kiel and Kiel University","correspondingAuthor":false,"prefix":"","firstName":"Clint","middleName":"","lastName":"Hansen","suffix":""},{"id":510970549,"identity":"ea1ff00c-443f-4d13-8880-80a591f30e0c","order_by":5,"name":"Elke Warmerdam","email":"","orcid":"","institution":"University of Twente","correspondingAuthor":false,"prefix":"","firstName":"Elke","middleName":"","lastName":"Warmerdam","suffix":""},{"id":510970551,"identity":"e0c09a28-3f4e-4776-bf0f-e3790419679e","order_by":6,"name":"Pedro Conceição","email":"","orcid":"","institution":"Kiel University","correspondingAuthor":false,"prefix":"","firstName":"Pedro","middleName":"","lastName":"Conceição","suffix":""},{"id":510970552,"identity":"8edf83fa-e352-46dc-8aee-c83987570ff1","order_by":7,"name":"Walter Maetzler","email":"","orcid":"","institution":"University Hospital Schleswig-Holstein, Campus Kiel and Kiel University","correspondingAuthor":false,"prefix":"","firstName":"Walter","middleName":"","lastName":"Maetzler","suffix":""}],"badges":[],"createdAt":"2025-08-20 08:53:36","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7415461/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7415461/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s12984-025-01835-1","type":"published","date":"2025-12-12T15:57:11+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":90924424,"identity":"55a4b8a3-31f4-4e75-8444-49d980a35e02","added_by":"auto","created_at":"2025-09-09 15:21:29","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":576709,"visible":true,"origin":"","legend":"\u003cp\u003eScheme of the kinectome analysis. a) original marker positions; b) marker positions after reduction; c) acceleration data of the most and least affected ankles; data cutting into one full left and one full right gait cycles = one kinectome; d) example kinectomes in anteroposterior, mediolateral, and vertical directions; e) representing the kinectome as a graph; f) strongest pattern in the graph (dark blue). Fig. 1a and Fig. 1b adapted from Warmerdam et al. (2021) with permission [11]\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7415461/v1/a61c0884bd00f8e7b6082c6f.png"},{"id":90924428,"identity":"d37486d9-1b51-42b6-9291-db8dbc70879a","added_by":"auto","created_at":"2025-09-09 15:21:29","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":110419,"visible":true,"origin":"","legend":"\u003cp\u003eStandard deviation kinectomes of Control and PD groups obtained from preferred walking speed trials, AP direction. Warmer colours indicate higher standard deviation. The P-value is obtained from permutation testing. AP – anteroposterior direction, asis – anterior superior iliac spine, las – least affected side, mas – most affected side, PD – Parkinsons’s disease group, psis – posterior superior iliac spine\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7415461/v1/a35f3ee9c85bb5d69e4c1149.png"},{"id":90924422,"identity":"1e36e4de-c8f2-4d27-a06c-9cd9983eef50","added_by":"auto","created_at":"2025-09-09 15:21:29","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":524566,"visible":true,"origin":"","legend":"\u003cp\u003eSignificantly different coordination patterns between PD and controls at preferred walking speed. The patterns with largest effect sizes are shown. Connected nodes indicate the strongest coordination deficit in pwPD; coloured but unconnected nodes appear in other significant patterns. Red indicates weaker inter-segmental coordination in the PD group than in the control group. The colour intensity reflects the frequency each node appeared in all significant patterns for the given speed, direction, and segmental level (more vivid - higher frequency). ASIS – anterior superior iliac spine; PSIS – posterior superior iliac spine\u003c/p\u003e","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7415461/v1/3c3ac28ff5b534d2709fc6f5.jpeg"},{"id":90925591,"identity":"7ddf1e5a-3eba-4005-b12a-4647d56ed6e9","added_by":"auto","created_at":"2025-09-09 15:29:29","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":242823,"visible":true,"origin":"","legend":"\u003cp\u003eSignificantly different coordination patterns between PD and controls at fast walking speed. The patterns with largest effect sizes are shown. Connected nodes indicate the strongest coordination deficit in pwPD; coloured but unconnected nodes appear in other significant patterns. Red indicates weaker, and blue indicates stronger inter-segmental coordination in the PD group than in the control group. The colour intensity reflects the frequency each node appeared in all significant patterns for the given speed, direction, and segmental level (more vivid - higher frequency). ASIS – anterior superior iliac spine; PSIS – posterior superior iliac spine\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7415461/v1/9193ba31d4d8b0a8c42b75d8.png"},{"id":98244834,"identity":"5697dd6c-d05f-4b5b-a88d-b12c3029344d","added_by":"auto","created_at":"2025-12-15 16:15:32","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2022783,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7415461/v1/a7a8cb4f-5500-4790-aedf-6f7fb1a381dd.pdf"},{"id":90924426,"identity":"be4c1163-d0e1-463e-9936-13f5088c4e85","added_by":"auto","created_at":"2025-09-09 15:21:29","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":633657,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterialkinectome.docx","url":"https://assets-eu.researchsquare.com/files/rs-7415461/v1/6ecd0ef9e4c62586e03bd9a2.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Inter-segmental coordination patterns in Parkinson’s disease are particularly disturbed during preferred walking speed: a data-driven network approach","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eParkinson\u0026rsquo;s disease (PD) is the second most common (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) and the fastest growing (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) neurodegenerative disease worldwide. Typical motor symptoms include bradykinesia, resting tremor, rigidity, freezing of gait, and unstable posture (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e). Furthermore, people with PD (pwPD) often have gait disturbances, such as shuffling gait, increased difficulty lifting feet from the ground, decreased stride length and velocity (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e), impaired foot placement (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e), and reduced range of walking speed (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). Importantly, these disturbances are associated with activity limitations during daily life (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eTypically gait of pwPD is evaluated with parameters that can be extracted from the lower extremities and the lower back, including velocity, stride and step length and width, cadence, gait phase duration, and variability and asymmetry of these metrics, as well as the range of motion of the joints of the lower extremities (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e). However, for an overall view of the entire human body as an integrated system it is important to capture many simultaneous interactions arising from coordinated work of different musculoskeletal segments. This can be analysed with information coming from more than only one area of the body (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e) and, for complex analyses, using the network theory (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e). The latter offers a framework for investigating complex systems, consisting of many interconnected elements, and how they interact with each other. This approach allows insights into the overall behaviour of the whole system (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eTroisi Lopez et al. (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e) have used principles from network theory to explore human locomotion. For that purpose, they have established a \u003cem\u003ekinectome\u003c/em\u003e, where human gait is represented as a matrix of acceleration correlations between the body segments. The kinectome was interpreted as a network, where the nodes corresponded to body segments, and edges were defined by the correlation of acceleration signals between the segments (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e). Using a variety of analysis methods, such as modularity and various topological measures, they were able to distinguish between pwPD and controls and to predict clinical symptoms in PD (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e). More specifically, they showed a reduced joint coordination in pwPD compared to controls (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e), and were able to demonstrate an effect of dopaminergic treatment on this coordination (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eHowever, to our best knowledge, how the full body is involved in walking speed-related deficits has not been adequately investigated to date. It has already been shown that rigidity contributes to reduced forward limb propulsion in PD, thus negatively affecting walking speed and step length (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e). Furthermore, pwPD have difficulties in modulating their walking speed (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e) and adapting to changes in walking speed (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e). To address the questions about the effect of different walking speed on inter-segmental coordination, which is defined as a relationship between different body segments during movement, we used the aforementioned network theory (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e) and applied it to a dataset produced in our lab (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e). We hypothesized that the inter-segmental coordination during walking differs between pwPD and healthy controls in lower and upper parts of the body, with the differences depending on different walking speeds.\u003c/p\u003e"},{"header":"2. Methods","content":"\u003cp\u003e\u003cem\u003e2.1 Participants, in-and exclusion criteria\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eParticipants aged 18 years or older who were able to walk independently without the use of walking aids were eligible for inclusion. Individuals were excluded if they scored below 15 on the Montreal Cognitive Assessment or if they had other movement disorders known to impact mobility, based on evaluation by a movement disorder specialist (WM). For full details regarding inclusion and exclusion criteria, see Warmerdam et al. (11).\u003c/p\u003e\n\u003cp\u003eA total of 58 participants were included in the present analysis. This comprised 29 pwPD (11 females), i.e., the PD group, who performed walking tasks during their self-reported optimal \u0026quot;ON\u0026quot; medication state, assessed between 30 and 120 minutes after oral intake of levodopa. The control group consisted of 29 age-matched healthy adults (14 females), with no comorbidities affecting mobility, as confirmed by the same movement disorder specialist. Demographic and clinical characteristics are summarized in Table 1.\u003c/p\u003e\n\u003cp\u003eTable 1. Demographic and clinical parameters of the participating groups\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"595\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 177px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eGroup\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 186px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eControls\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 232px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePD\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 177px;\"\u003e\n \u003cp\u003eTotal \u003cem\u003eN\u003c/em\u003e (males/females)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 186px;\"\u003e\n \u003cp\u003e29 (15/14)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 232px;\"\u003e\n \u003cp\u003e29 (18/11)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 177px;\"\u003e\n \u003cp\u003eAge (years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 186px;\"\u003e\n \u003cp\u003e67 \u0026plusmn; 12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 232px;\"\u003e\n \u003cp\u003e67 \u0026plusmn; 10\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 177px;\"\u003e\n \u003cp\u003eHeight [cm]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 186px;\"\u003e\n \u003cp\u003e176 \u0026plusmn; 10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 232px;\"\u003e\n \u003cp\u003e174 \u0026plusmn; 9\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 177px;\"\u003e\n \u003cp\u003eWeight [kg]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 186px;\"\u003e\n \u003cp\u003e79 \u0026plusmn; 16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 232px;\"\u003e\n \u003cp\u003e82 \u0026plusmn; 18\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 177px;\"\u003e\n \u003cp\u003eBMI [kg/m\u003csup\u003e2\u003c/sup\u003e]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 186px;\"\u003e\n \u003cp\u003e26 \u0026plusmn; 5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 232px;\"\u003e\n \u003cp\u003e27 \u0026plusmn; 5\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 177px;\"\u003e\n \u003cp\u003eDisease duration (years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 186px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 232px;\"\u003e\n \u003cp\u003e9 \u0026plusmn; 6\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 177px;\"\u003e\n \u003cp\u003eHoehn \u0026amp; Yahr (1-5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 186px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 232px;\"\u003e\n \u003cp\u003e3 \u0026plusmn; 1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 177px;\"\u003e\n \u003cp\u003eMedication dose (LEDD)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 186px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 232px;\"\u003e\n \u003cp\u003e713 \u0026plusmn; 301\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 177px;\"\u003e\n \u003cp\u003eMDS-UPDRS III\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 186px;\"\u003e\n \u003cp\u003e4 \u0026plusmn; 4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 232px;\"\u003e\n \u003cp\u003e29 \u0026plusmn; 21\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eThe values are displayed as mean \u0026plusmn; standard deviation. PD \u0026ndash; Parkinson\u0026rsquo;s disease; BMI \u0026ndash; Body Mass Index; LEDD \u0026ndash; Levodopa equivalent daily dose; MDS-UPDRS III \u0026ndash; Motor part of the Movement Disorder Society Unified Parkinson\u0026rsquo;s Disease Rating Scale\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e2.2 Determining the most and least affected sides\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe most and least affected upper extremities and lower extremities were determined based on the motor part of the Movement Disorder Society Unified Parkinson\u0026rsquo;s Disease Rating Scale (MDS-UPDRS III) (20) scores, namely the sum of items 3.3 (rigidity), 3.4 (finger tapping), 3.5 (hand movements), and 3.6 (pronation and supination of the hand) for left and right upper extremity, and the sum of 3.3 (rigidity), 3.7 (toe tapping), and 3.8 (leg agility) for left and right lower extremity. A difference of \u0026ge;1 point was considered as having lateralization (21) . If the MDS-UPDRS III scores between the sides were even, as well as for the control group, the non-dominant hand was taken as the most affected side due to lower motor performance on the non-dominant side (22). For participants missing the handedness data (9 PD and 6 controls), the left hand was taken as non-dominant due to handedness distribution in the general population, where around 90% of all people are right-handed (23).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e2.3 Equipment, data acquisition and data dimensionality reduction\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eAll the motor tasks were recorded using a twelve-camera optical motion capture system recording with 200 Hz (Qualisys AB, G\u0026ouml;teborg, Sweden). A total of 47 reflective markers were placed on the body (Fig. 1a). Full details on the experimental protocol can be found in Warmerdam et al. (11). For the present study, the four head markers were calculated into one midpoint, the 3-marker cluster on the sternum was calculated into one midpoint, the 4-marker clusters on the thighs and shanks were calculated into one midpoint per segment and per side, and the upper arm, forearm, and heel marker locations were excluded from further data analysis. The resulting 22 marker data (Fig. 1b), representing body segments, were taken for further analysis.\u003c/p\u003e\n\u003cp\u003e2.4 Motor assessment\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAll study participants were instructed to walk a straight path along a 5 m long and 1 m wide walkway. There were three walking trials, namely at preferred (\u0026ldquo;Please walk at your normal walking speed\u0026rdquo;), fast (\u0026ldquo;Please walk as fast as possible, without running, falling or feeling unsafe\u0026rdquo;), and slow (\u0026ldquo;Please walk half of your normal walking speed\u0026rdquo;) speeds. The walking speed and the differences between the trials can be found Supplementary Figure 1. The beginning and end of the walkway were marked by cones with reflective markers. The participants were asked to start walking approximately 2 m before the start cones and finish walking approximately 2 m behind the end cones to make sure that steady state walking was measured within the 5 m gait trial and any variation in gait appearing due to gait initiation and termination (7,24) is excluded from the data acquisition.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e2.5 Marker data pre-processing\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eMarker data were prepared in a format following the Brain Imaging Data Structure (BIDS) (25,26) and were pre-processed using Python (Python Software Foundation. Python Language Reference, version 3.11.2). The walking data were trimmed to only contain the samples collected between the start and end cones. These data were then cut so each gait acquisition\u003cem\u003e\u0026nbsp;\u003c/em\u003econtains one full left and one full right gait cycle (Fig. 1c). Missing values were linearly interpolated using the fill_missing_samples function from Kinetics Toolkit (27) with the threshold of 271 samples (1.35 s). The remaining data were filtered using a 2\u003csup\u003end\u003c/sup\u003e order Butterworth filter with the cut-off frequency of 6 Hz. Next, the data were rotated using Principal Component Analysis to align the x direction of the markers with the anteroposterior (AP) movement direction. Lastly, the data, originally containing positions, was differentiated twice to get the acceleration of the markers.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e2.6 Building the kinectome\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eTime series containing acceleration data of each marker during each gait acquisition in the AP, mediolateral (ML), and vertical (V) directions were first normalised to 500 samples, and then were used to build person, trial and gait acquisition specific kinectomes. The kinectomes were built in each movement direction separately by correlating the data from each marker with that from every other marker, resulting in a mirrored 22✕22✕3 correlation matrix (where 22 is the number of body segments, and 3 is the number of movement directions, Fig. 1d). This matrix will be referred to as the kinectome from now on\u0026nbsp;(14). In this paper we present the results obtained using Pearson\u0026apos;s correlation.\u003c/p\u003e\n\u003cp\u003eThen for each person an average kinectome was built by averaging the kinectomes from multiple gait kinectomes obtained from the respective trial in a respective movement direction (e.g., averaging kinectomes one, two, and three in Fig. 1c). For each person standard deviation of the kinectomes was obtained by calculating the standard deviation between all gait kinectomes (Fig. 1c) for each trial and movement direction separately. This resulted in each person having one person-specific average and one person-specific standard deviation kinectome per trial and direction (total of 9 average and 9 standard deviation kinectomes per person).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e2.7 Group kinectome analysis\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe individual average and standard deviation kinectomes were used to build group-specific average and standard deviation kinectomes, respectively, for each walking speed and each movement direction separately for the control and PD groups, by averaging all individual kinectomes between the people within the groups. The latter kinectomes were used for permutation analysis.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e2.8 Representing the kinectome as a graph\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eEach person-specific average kinectome was represented as a graph (Fig. 1e) for further analysis. The body segments were represented as nodes, and the acceleration correlations between them as edges between the nodes. These graphs were non-directional, i.e., there was no particular direction of the edges between the nodes; and complete, i.e., each node was connected with each other node, based on the assumption that motor coordination involves relationships between anatomically distant joints (e.g. contralateral limb swing during gait) (28). Network analysis was done\u0026nbsp;\u003cbr\u003eusing NetworkX package (Version 3.4.2) (29) for Python.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e2.9 Patterns\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eA pattern of a graph refers to a specific arrangement of its nodes and edges (Fig. 1f). Here we analysed path patterns, i.e., sequences of connected nodes and edges. In this paper, a path refers to a sequence of nodes connected by edges, and a pattern is a path represented in the kinematic data. We have looked at the maximum-weighted patterns in the respective person-specific average kinectome that started from a given node and had a fixed length. This approach provided an unconstrained and data-driven view on respective comparisons. These comparisons were done on three arbitrarily defined levels, namely local (2-6 body segments in the pattern), referring to simple coordination networks, multi-segmental (7-14 body segments) \u0026ndash; more complex networks, and whole-body (15+ body segments) \u0026ndash; global coordination networks. The maximum-weighted patterns were chosen for the analysis because they provide the baseline for understanding coordination differences between the groups.\u003c/p\u003e\n\u003cp\u003eThe algorithm src.graph_utils.kinectome2pattern.strongest_pattern_subgraph (available on (30)) starts at a given node and constructs a path by iteratively selecting edges with the highest absolute weight while avoiding node repetition. This ensures that each subsequent step in the path represents the strongest local connection available from the current position, creating a path that follows the most prominent signal through the network structure. Note that while the starting node is fixed, there is no predetermined target node\u0026mdash;only the path length is constrained.\u003c/p\u003e\n\u003cp\u003eSince there are 22 bone segments taken for the analysis, the algorithm looped over each body segment and took it as a starting node. Pattern lengths ranging from 2 to 20 nodes were considered. First, the strongest pattern for each individual was found. Then, within each group, patterns of the same length and starting node were averaged to determine the strongest group-specific pattern. Finally, the values of this pattern were then taken from each individual in both groups and compared statistically.\u003c/p\u003e\n\u003cp\u003eThis strategy enabled an exploratory analysis of emerging patterns; rather than specifying a particular pattern, it allowed the algorithm to first identify the strongest pattern and then assess the differences between these patterns across groups.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e2.10 Statistics\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eStatistical analysis was done using SciPy (31) for Python and JASP (JASP Team (2023). JASP (Version 0.18)). The characteristics of the groups are presented as mean \u0026plusmn; standard deviation. Normality distribution was checked using the Shapiro-Wilk test. Outliers were removed using the Z-score method (removing values three standard deviations above or below the group mean) for normally distributed data, or by calculating the interquartile range and removing values 1.5 times greater or less than the interquartile range for non-normally distributed data.\u003c/p\u003e\n\u003cp\u003eChi-square test was used to compare the gender distribution between the groups. To compare all other variables between the groups, t-test and Mann-Whitney U tests were used for normally and non-normally distributed data, respectively.\u003c/p\u003e\n\u003cp\u003eThe group-specific average and standard deviation kinectomes were compared through permutation testing, by randomly shuffling the control group kinectomes 5000 times. Spearman\u0026rsquo;s rho was used to generate the distribution of the differences that are to be expected by chance alone (32,33). To assess the stability of the kinectomes across sample sizes, we performed bootstrap resampling (n=5000) at varying percentages (10-90%) of the original dataset. For each bootstrap iteration, kinectomes for each walking speed and movement direction were computed using the subset of the full sample, and permutation testing was applied to these subsets to build a bootstrap distribution of correlations. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThere were two pattern analyses done, namely strongest patterns found in the control group were compared to the same patterns in the PD group, and strongest patterns found in the PD group were compared to those in the control group. The patterns derived from the control group are presented in the result section, and the results comparing PD group\u0026rsquo;s patterns can be found in Supplementary Table 1. Patterns with an overlap of 80% as determined by Jaccard index (34) were seen as the same pattern. Higher values of the pattern indicate stronger correlation between the body segments of that particular pattern. The patterns between the groups were compared using t-test or Mann-Whitney U test for normally and non-normally distributed data, respectively. The effect size was calculated using Cohen\u0026rsquo;s d. Effect size measure was interpreted as small, medium, or large if the effect size was 0.2\u0026lt;x\u0026lt;0.5, 0.5\u0026lt;x\u0026lt;0.8, and x\u0026ge;0.8, respectively (35). To adjust for multiple comparisons when comparing the patterns of the same length, but different starting nodes, a Bonferroni correction was used (n=22). All comparisons were made between the groups within the same walking speed and same movement direction.\u003c/p\u003e"},{"header":"3. Results","content":"\u003cp\u003eThere were no significant differences between the groups in gender distribution, age, height, weight, or body mass index (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;.05). The MDS-UPDRS III scores differed between the control and PD groups, with the latter group having higher scores. Details are shown in Table 1.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e3.1 Comparing the group-specific kinectomes\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eAverage and standard deviation kinectomes\u0026rsquo; comparisons between the PD and control groups using permutation testing did not reveal any significant differences. The correlations between the average and between the standard deviation kinectomes (correlation values are displayed in Table 2) are unlikely to have occurred by chance, and the observed correlation is stronger than what could be expected by random chance.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable 2. Spearman\u0026rsquo;s rho values between the group-specific kinectomes for preferred, fast, and slow walking speeds\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"594\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSpearman\u0026rsquo;s rho\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 153px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePreferred\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 153px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFast\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"4\" valign=\"top\" style=\"width: 157px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSlow\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003eAP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003eML\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003eV\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003eAP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003eML\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003eV\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003eAP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003eML\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003eV\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 4px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAverage kinectomes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e.99\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e.93\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e.98\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e.91\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e.92\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e.98\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e.94\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e.98\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 4px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eStandard deviation kinectomes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e.73\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e.63\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e.83\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e.88\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e.61\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e.66\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e.88\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e.74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 51px;\"\u003e\n \u003cp\u003e.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 4px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eSpearman\u0026rsquo;s rho values between control and PD standard deviation kinectomes were consistently lower than those for the average kinectomes across all walking speeds and directions. A visual representation of a standard deviation kinectomes can be found in Figure 2. Moreover, the standard deviation values were consistently higher in the PD group than in the control group. These findings indicate that while the overall structure of the inter-segmental coordination is similar between the groups, the variability in this inter-segmental coordination is higher in PD group than in controls. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e3.1.1 Bootstrapping\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eBootstrap analysis demonstrated high correlation stability across all conditions. The correlations approached asymptotic values by 40-60% sample size for average and 80-90% for standard deviation kinectomes. Using 80% of the original sample for the subset size, the observed correlations exceeded bootstrap means (Supplementary Fig. 3 and 4).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e3.2 Graph patterns\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003ePattern analysis revealed distinct differences in movement coordination patterns between PD and control groups across movements in AP, ML, and V directions mainly during preferred and fast walking, while slow walking had only one significantly different pattern. Full patterns can be found in Supplementary Table 2.\u003c/p\u003e\n\u003cp\u003eThe results consistently show altered inter-segmental coordination in the PD group across various organizational levels (local, multi-segmental, and whole-body). Importantly, the manifestation and characteristics of these deficits varied with walking speed.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003ePreferred walking speed\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eAt preferred walking speed, the inter-segmental coordination deficits of the PD group were most apparent (Fig. 3). The reduction in coordination was consistent and multi-level in both the AP and ML directions. In the following, the main findings are summarized:\u003c/p\u003e\n\u003cul class=\"decimal_type\"\u003e\n \u003cli\u003eAP direction: the PD group showed weaker coordination than controls across all organizational levels, with the most significantly different patterns observed in this speed and direction (n=146):\u003cul class=\"decimal_type\"\u003e\n \u003cli\u003eOn a local level (n=26), patterns mostly involving axial segments (head, trunk, pelvis), both upper limbs, and the less affected lower limb, were weaker in the PD group (effect sizes: 0.53-1.86, mean=0.8). This suggests deficits in basic postural coordination.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eOn a multi-segmental level (n=80), patterns involving multiple body regions and including contralateral limb movement were weaker in the PD group (effect sizes: 0.53-1.67, mean=0.78). These findings suggest PD-related deficits in simultaneous multi-segmental coordination during walking.\u003c/li\u003e\n \u003cli\u003eOn a whole-body level (n=40), patterns involving most body segments showed the most pronounced deficits in the PD group (effect sizes: 0.53-1.35, mean=0.7). This suggests that pwPD cannot effectively coordinate whole-body inter-segmental movements during walking.\u0026nbsp;\u003c/li\u003e\n \u003c/ul\u003e\n \u003c/li\u003e\n \u003cli\u003eML direction: Weaker inter-segmental coordination was observed across all organizational levels in the PD group compared to controls, although notably fewer patterns were significantly different (n=16) compared to the AP direction.\u003cul class=\"decimal_type\"\u003e\n \u003cli\u003eOn a local level (n=5), patterns involving sternum, the less affected shoulder and thigh, as well as between the most affected upper limb and several segments in both most and least affected lower limbs, were weaker in the PD group (effect sizes: 0.54-1.59, mean=1.33).\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eOn a multi-segmental level (n=7), patterns primarily including the pelvis, and upper and lower limbs, were weaker in the PD group (effect sizes: 0.54-1.59, mean=1.33). This suggests impaired synchronization between the more and less affected body sides during walking.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eOn a whole-body level (n=4), patterns involving most body segments showed the largest deficits in the PD group (effect sizes: 1-1.04, mean=1.02). This suggests that whole-body coordination in the ML direction, crucial for stability during walking, is affected by PD.\u0026nbsp;\u003c/li\u003e\n \u003c/ul\u003e\n \u003c/li\u003e\n \u003cli\u003eV direction: No significant differences were observed.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cem\u003eFast Walking Speed\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eFewer and distinct significant differences between the PD group and controls were observed compared to preferred speed, and they were observed exclusively on a multi-segmental level:\u003c/p\u003e\n\u003cul class=\"decimal_type\"\u003e\n \u003cli\u003e\u0026nbsp;AP direction: Two patterns on a multi-segmental level were found to be significantly weaker in the PD group (effect sizes: 0.51-1, mean=0.76) (Fig. 4, AP direction, subplot b). The patterns involved core segments, both upper limbs and the less affected lower limb, suggesting an impaired synchronization between the core and limb swing during walking.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eML direction: No significant differences.\u003c/li\u003e\n \u003cli\u003eV Direction: Two patterns on a multi-segmental level were found to be significantly different (Fig. 4, V direction, subplot b). The patterns involved sternum, most affected side of the pelvis and both upper and lower limbs. Interestingly, these values were \u003cem\u003ehigher\u003c/em\u003e in the PD group (effect sizes: 0.96-99, mean=0.98). This suggestsuggesting compensatory activation of muscles and potentially also increased rigidity within complex networks mainly including upper and lower limbs, as well as pelvic segments.\u0026nbsp;\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003e\u003cem\u003eSlow Walking Speed\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eAt slow walking speed, minimal significant differences in inter-segmental coordination were observed between the groups:\u003c/p\u003e\n\u003cul class=\"decimal_type\"\u003e\n \u003cli\u003eAP and V direction: No significant differences were found.\u003c/li\u003e\n \u003cli\u003eML direction: One pattern was found to be different in the slow walking speed (Supplementary Figure 2, ML direction, subplot a), involving the less affected shoulder and thigh, with an effect size of 1.17.\u003c/li\u003e\n\u003c/ul\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThis study describes the inter-segmental coordination of the whole body during walking in pwPD and age-matched controls, and how this coordination changes at different walking speeds. The coordination between the segments was evaluated by the means of kinectomes, containing pairwise acceleration interactions between different body segments. Analysis methods were based on the network theory to evaluate the inter-segmental coordination differences between pwPD and controls (\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe high correlations in the average kinectomes suggest that the fundamental inter-segmental coordination structure remains intact in PD. This result reflects our clinical experience: pwPD are able to propel themselves forward during their entire mobile phase of the disease, while maintaining the swing of contralateral limbs. As shown with our collected data (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e) (Supplementary Fig.\u0026nbsp;1), as well as previous research, pwPD are also able to modulate their walking speed (\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e), although to a lesser extent than healthy older adults (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eHowever, movement variability during gait, as seen in reduced correlations in the standard deviation kinectomes, as well as their overall higher standard deviation in the PD group, was increased especially during the preferred walking speed. Increased variability in gait of pwPD has been previously shown by multiple studies, however they mostly investigated discrete spatial and temporal gait parameters (\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e) like swing time (\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e), stride time (\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e), stride length (\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e), and stride-to-stride variability (\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e). The kinectome allows us to look at the inter-segmental coordination (and the variability thereof) by analysing simultaneous kinematic relationships across multiple body segments, as well as all three movement directions. This approach helps to dive deeper into pathological mechanisms during gait, but also to evaluate potential adaptability and compensatory mechanisms, providing insights for allied therapeutic health interventions (\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe analysis of strongest inter-segmental coordination patterns revealed the predominance of coordination deficits at preferred walking speed, suggesting that pwPD experience impaired motor control during their most comfortable walking speed, which is typically the most often used and most efficient for healthy people (\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e). Our results therefore suggest that in pwPD, the motor coordination system is more challenged during preferred than fast and slow walking speed. This consistent and multi-level (starting from pairs of two body segments and ending at patterns involving all body segments) deficit of inter-segmental coordination strength indicates fundamental impairments in integrating movements, which are crucial for dynamic stability (\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e) and overall gait execution. These coordination deficits likely reflect the underlying pathophysiology of PD where the degeneration of dopaminergic neurons affects the whole basal ganglia network (\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e), which is responsible for selecting optimal motor strategies (\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e). In addition, the reduced coordination between the segments in pwPD, compared to controls, could make gait less efficient and require increased effort, contributing to further symptoms well known to occur in PD: fatigue (\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e) and reduced walking endurance (\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eFurthermore, the motor coordination deficits were most pronounced in the AP and ML directions. Deficits in the axial segments, especially in the AP direction, suggest issues with core postural control during ambulation, which is a known difficulty in pwPD (\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e). Our data show that, in AP direction, the hip segments are centrally involved in significantly weaker patterns pwPD, suggesting that especially interventions targeting pelvic coordination (and its coordinated \u0026ldquo;interaction\u0026rdquo; with distant body parts) could help these people to keep their gait stable during forward movement. Furthermore, the decreased limb coordination in ML direction that we observed in our PD group, point to the presence of lateral stability impairments during gait in this disease. This is of particular relevance, as lateral stability deficits in pwPD have been linked to falls (\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e, \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eInterestingly, the fast walking speed revealed a completely different picture of gait coordination pattern in pwPD. The patterns showed a shift from severely impaired gait coordination, pronounced in the AP and ML direction, as observed in the preferred walking condition, towards generally much \u0026ldquo;better\u0026rdquo; gait coordination during the fast walking condition. Previous studies have shown that, in pwPD, the variabilities of stride length and stride time are lower during fast walking than during preferred walking (\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e), bringing the gait pattern of pwPD during fast walking closer to that of healthy controls. We have observed similar results. During the fast walking condition, only two patterns in the AP direction and none in the ML direction were significantly weaker in the PD group than in the control group, as opposed to 146 significant patterns in the AP direction and 16 in the ML direction during preferred speed (Supplementary Table\u0026nbsp;2).\u003c/p\u003e\u003cp\u003eWhy did we observe most differences of gait coordination between PD and controls in preferred walking speed? One explanation could be that walking at a speed above the \u0026ldquo;comfort zone\u0026rdquo; improves the arousal state (\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e) and with that resources to coordinate gait. Previously it was shown that a known phenomenon in pwPD called \u0026ldquo;paradoxical kinesis\u0026rdquo; occurs in situations when a person is highly influenced by external cues or higher arousal (\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e, \u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e) and is able to produce movements like healthy people. Another reason could be that higher walking speeds lead to a proportionally reduced energy consumption (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e), potentially making brain reserves free for gait coordination. Moreover, PD is known to be a disease which affects automatic movements \u0026ndash; pwPD lose previously acquired automatic skills and have difficulty restoring them (\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e). Importantly, these automatic movements are associated with posterior putamen, a region in basal ganglia which is primarily affected by PD due to the loss of dopamine (\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e). Preferred walking speed can be seen as automatic protocol for walking, and may therefore be particularly affected by PD. Fast walking speed, as assessed in this study, could have served as the cue to move in a non-automated manner and demonstrate more control-like performance.\u003c/p\u003e\u003cp\u003eWhy pwPD, based on the pattern analysis, showed an even higher inter-limb coordination than controls in the fast walking condition in the V direction, is not yet clear. We hypothesize that, to maintain stability and forward propulsion at higher walking speed, pwPD may stiffen multiple segments and put them to work together, sacrificing fluid, relaxed movements to increase stability and meet the more demanding walking speed requirements. Future studies should confirm our results and, if this is the case, investigate whether this high interlimb coordination in the V axis indeed reflects better coordination, or whether this adaptation comes with a loss of adaptability and thus an increased risk of negative consequences, such as falls.\u003c/p\u003e\u003cp\u003eThe differences in gait coordination were least significant in the slow walking condition. We hypothesize that, at slow walking speed, the motor system has sufficient time to plan and execute movements, bringing the movement pattern closer to that observed in healthy adults. It is already known that, in order to compensate for reduced dynamic stability during walking, pwPD adopt strategies with shorter step length (\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e), which could have allowed them choosing a control-like inter-segmental integration on a whole-body level.\u003c/p\u003e\u003cp\u003eThere are a few considerations worth mentioning regarding the pattern analysis. We have chosen to compare the maximum-weighted patterns between the groups as baseline, however future work examining weaker patterns could reveal compensatory mechanisms, fatigue, less optimal motor strategies, or subtle early stage changes during the disease progression. Furthermore, a sub-analysis of the patterns themselves could identify where do the strongest group patterns rank on an individual list, and if they are significantly stronger than the other possible patterns. In addition, future studies should inlcude measures of varability (e.g. by bootstrapping) to understand how generalizable the results from such high-level analysis are.\u003c/p\u003e\u003cp\u003eThis study has several strengths and limitations. Among the strengths are the evaluation of inter-segmental coordination of the entire body, different walking speeds, and including bootstrapping to assess the stability of the kinectomes. The limitations include the evaluation of the PD group only after intake of medication. Different coordination patterns may be revealed in the off-medication state. Other limitations include high disease variability, the focus of relatively short bouts of straight walking, and a relatively high proportion of pwPD without lateralization. Furthermore, the data was collected in a clinical setting, namely a movement analysis laboratory, with no obstacles and no disturbances, which certainly occur in usual daily lives.\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eIn conclusion, our results from a data-driven network analysis indicate that inter-segmental coordination of gait at preferred walking speed is severely affected in pwPD. We hypothesise that this alteration is due to deficits in automaticity and arousal networks. Furthermore, the distribution of the patterns that are significantly different from controls provide insights into the pathomechanistic aspects of body control during walking in pwPD, and potentially interesting treatment options. Interestingly, these coordination deficits were much less obvious during fast and slow walking, respectively. These findings support the implementation of different walking speeds in mobility training in pwPD. Our direction-specific analyses suggest that exercises which help in improving coordination in AP and ML directions, including contralateral limb swing and pelvic segmental coordination, have the most potential in addressing inter-limb coordination deficits in PD.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eElectronic supplementary material\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAvailable.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe study was approved by the ethics committee of the Medical Faculty of Kiel University (D438/18) and was conducted in accordance with the principles of the Declaration of Helsinki. The study is registered in the German Clinical Trials Register (DRKS00022998, registered on 04 Sep 2020). All study participants have read and signed an informed consent prior to the study.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors have read and agreed to the final version of the manuscript.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePart of the data analysed during the current study are available at (59). The remaining data contains patient information and are available from the corresponding author on reasonable request. The data analysis code\u003ca href=\"https://github.com/karsae/pykinectome\"\u003e\u0026nbsp;\u003c/a\u003eis available at (30). It is platform independent, written in Python programming language. Requirements include Python version 3.11 or higher.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNone declared\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor\u0026lsquo;s contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eValidation (1); formal analysis (2); investigation (3); resources (4); data curation (5); writing\u0026mdash;original draft preparation (6); writing\u0026mdash;review and editing (7); visualization (8); supervision (9). K.S.: 1, 2, 3, 6, 7, 8; R.R.: 2, 3, 7; I.R.: 7; J.W.: 3, 7; C.H.: 1, 7, 9; E.W.: 5, 7; P.C.: 2, 3, 7; W.M.: 4, 7, 9. All authors have read and agreed to the final version of the manuscript.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLandesprojekt LPW21-E/1.2.2/179.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eZhu J, Cui Y, Zhang J, Yan R, Su D, Zhao D, et al. Temporal trends in the prevalence of Parkinson\u0026rsquo;s disease from 1980 to 2023: a systematic review and meta-analysis. The Lancet Healthy Longevity. 2024 Jul 1;5(7):e464\u0026ndash;79. \u003c/li\u003e\n\u003cli\u003eFeigin VL, Abajobir AA, Abate KH, Abd-Allah F, Abdulle AM, Abera SF, et al. Global, regional, and national burden of neurological disorders during 1990\u0026ndash;2015: a systematic analysis for the Global Burden of Disease Study 2015. The Lancet Neurology. 2017 Nov 1;16(11):877\u0026ndash;97. \u003c/li\u003e\n\u003cli\u003eBloem BR, Okun MS, Klein C. Parkinson\u0026rsquo;s disease. The Lancet. 2021 Jun 12;397(10291):2284\u0026ndash;303. \u003c/li\u003e\n\u003cli\u003eMoustafa AA, Chakravarthy S, Phillips JR, Gupta A, Keri S, Polner B, et al. Motor symptoms in Parkinson\u0026rsquo;s disease: A unified framework. Neurosci Biobehav Rev. 2016 Sep;68:727\u0026ndash;40. \u003c/li\u003e\n\u003cli\u003eStuopelytė A, \u0026Scaron;akalienė R. Gait Training Methods and ChangesiIn Gait Parameters in Parkinson`s Disease (Literature Review). Reabilitacijos mokslai: slauga, kineziterapija, ergoterapija [Internet]. 2012 Dec 27 [cited 2025 May 23];2(7). Available from: https://journals.lsu.lt/reabilitacijos-mokslai/article/view/857\u003c/li\u003e\n\u003cli\u003eLang C, Hausdorff JM, Bruijn SM, Brodie MA, Okubo Y, Maetzler W, et al. Foot placement coordination is impaired in people with Parkinson\u0026rsquo;s Disease [Internet]. bioRxiv; 2025 [cited 2025 Aug 6]. p. 2025.06.11.659042. Available from: https://www.biorxiv.org/content/10.1101/2025.06.11.659042v1\u003c/li\u003e\n\u003cli\u003evon der Recke F, Warmerdam E, Hansen C, Romijnders R, Maetzler W. Reduced Range of Gait Speed: A Parkinson\u0026rsquo;s Disease-Specific Symptom? Journal of Parkinson\u0026rsquo;s Disease. 2023 Jan 1;13(2):197\u0026ndash;202. \u003c/li\u003e\n\u003cli\u003eLang C, van Dieen JH, Brodie MA, Welzel J, Maetzler W, Singh NB, et al. Complexities and challenges of translating intervention success to real world gait in people with Parkinson\u0026rsquo;s disease. Front Neurol [Internet]. 2024 Oct 9 [cited 2025 Aug 6];15. Available from: https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2024.1455692/full\u003c/li\u003e\n\u003cli\u003eBou\u0026ccedil;a-Machado R, Jalles C, Guerreiro D, Pona-Ferreira F, Branco D, Guerreiro T, et al. Gait Kinematic Parameters in Parkinson\u0026rsquo;s Disease: A Systematic Review. Journal of Parkinson\u0026rsquo;s Disease. 2020 Jul 28;10(3):843\u0026ndash;53. \u003c/li\u003e\n\u003cli\u003edi Biase L, Di Santo A, Caminiti ML, De Liso A, Shah SA, Ricci L, et al. Gait Analysis in Parkinson\u0026rsquo;s Disease: An Overview of the Most Accurate Markers for Diagnosis and Symptoms Monitoring. Sensors (Basel). 2020 Jun 22;20(12):3529. \u003c/li\u003e\n\u003cli\u003eWarmerdam E, Romijnders R, Geritz J, Elshehabi M, Maetzler C, Otto JC, et al. Proposed Mobility Assessments with Simultaneous Full-Body Inertial Measurement Units and Optical Motion Capture in Healthy Adults and Neurological Patients for Future Validation Studies: Study Protocol. Sensors. 2021 Jan;21(17):5833. \u003c/li\u003e\n\u003cli\u003eBarab\u0026aacute;si AL. Network science. Philos Trans A Math Phys Eng Sci. 2013 Mar 28;371(1987):20120375. \u003c/li\u003e\n\u003cli\u003eBoonstra TW, Danna-Dos-Santos A, Xie HB, Roerdink M, Stins JF, Breakspear M. Muscle networks: Connectivity analysis of EMG activity during postural control. Sci Rep. 2015 Dec 4;5:17830. \u003c/li\u003e\n\u003cli\u003eTroisi Lopez E, Sorrentino P, Liparoti M, Minino R, Polverino A, Romano A, et al. The kinectome: A comprehensive kinematic map of human motion in health and disease. Ann N Y Acad Sci. 2022 Oct;1516(1):247\u0026ndash;61. \u003c/li\u003e\n\u003cli\u003eNewman MEJ. Modularity and community structure in networks. Proceedings of the National Academy of Sciences. 2006 Jun 6;103(23):8577\u0026ndash;82. \u003c/li\u003e\n\u003cli\u003eTroisi Lopez E, Liparoti M, Minino R, Romano A, Polverino A, Carotenuto A, et al. Kinematic network of joint motion provides insight on gait coordination: An observational study on Parkinson\u0026rsquo;s disease. Heliyon. 2024 Aug 15;10(15):e35751. \u003c/li\u003e\n\u003cli\u003eRomano A, Liparoti M, Minino R, Polverino A, Cipriano L, Carotenuto A, et al. The effect of dopaminergic treatment on whole body kinematics explored through network theory. Sci Rep. 2024 Jan 22;14(1):1913. \u003c/li\u003e\n\u003cli\u003ePistacchi M, Gioulis M, Sanson F, De Giovannini E, Filippi G, Rossetto F, et al. Gait analysis and clinical correlations in early Parkinson\u0026rsquo;s disease. Funct Neurol. 2017 Apr 6;32(1):28\u0026ndash;34. \u003c/li\u003e\n\u003cli\u003eSaegner K, Romijnders R, Hansen C, Holder J, Warmerdam E, Maetzler W. Inter-joint coordination with and without dopaminergic medication in Parkinson\u0026rsquo;s disease: a case-control study. J Neuroeng Rehabil. 2024 Jul 13;21:118. \u003c/li\u003e\n\u003cli\u003eGoetz CG, Tilley BC, Shaftman SR, Stebbins GT, Fahn S, Martinez-Martin P, et al. Movement Disorder Society-sponsored revision of the Unified Parkinson\u0026rsquo;s Disease Rating Scale (MDS-UPDRS): Scale presentation and clinimetric testing results: MDS-UPDRS: Clinimetric Assessment. Mov Disord. 2008 Nov 15;23(15):2129\u0026ndash;70. \u003c/li\u003e\n\u003cli\u003eOrcioli-Silva D, Barbieri FA, dos Santos PCR, Beretta VS, Simieli L, Vitorio R, et al. Double obstacles increase gait asymmetry during obstacle crossing in people with Parkinson\u0026rsquo;s disease and healthy older adults: A pilot study. Sci Rep. 2020 Feb 10;10:2272. \u003c/li\u003e\n\u003cli\u003ePrzybyla A, Good DC, Sainburg RL. Dynamic dominance varies with handedness: reduced interlimb asymmetries in left-handers. Exp Brain Res. 2012 Feb 1;216(3):419\u0026ndash;31. \u003c/li\u003e\n\u003cli\u003ePapadatou-Pastou M, Ntolka E, Schmitz J, Martin M, Munaf\u0026ograve; MR, Ocklenburg S, et al. Human handedness: A meta-analysis. Psychol Bull. 2020 Jun;146(6):481\u0026ndash;524. \u003c/li\u003e\n\u003cli\u003eLin CC, Creath RA, Rogers MW. Variability of Anticipatory Postural Adjustments During Gait Initiation in Individuals With Parkinson Disease. J Neurol Phys Ther. 2016 Jan;40(1):40\u0026ndash;6. \u003c/li\u003e\n\u003cli\u003eGorgolewski KJ, Auer T, Calhoun VD, Craddock RC, Das S, Duff EP, et al. The brain imaging data structure, a format for organizing and describing outputs of neuroimaging experiments. Sci Data. 2016 Jun 21;3(1):160044. \u003c/li\u003e\n\u003cli\u003eJeung S, Cockx H, Appelhoff S, Berg T, Gramann K, Grothkopp S, et al. Motion-BIDS: an extension to the brain imaging data structure to organize motion data for reproducible research. Sci Data. 2024 Jul 2;11(1):716. \u003c/li\u003e\n\u003cli\u003eCh\u0026eacute;nier F. Kinetics Toolkit: An Open-Source Python Package to Facilitate Research in Biomechanics. Journal of Open Source Software. 2021 Oct 23;6(66):3714. \u003c/li\u003e\n\u003cli\u003eDietz V, Michel J. Human bipeds use quadrupedal coordination during locomotion. Ann N Y Acad Sci. 2009 May;1164:97\u0026ndash;103. \u003c/li\u003e\n\u003cli\u003eHagberg AA, Schult DA, Swart PJ. Exploring Network Structure, Dynamics, and Function using NetworkX. In: Varoquaux G, Vaught T, Millman J, editors. Proceedings of the 7th Python in Science Conference. Pasadena, CA USA; 2008. p. 11\u0026ndash;5. \u003c/li\u003e\n\u003cli\u003eSaegner K, Romijnders R, Concei\u0026ccedil;\u0026atilde;o P. pykinectome [Internet]. Neurogeriatrics Kiel; Available from: https://github.com/neurogeriatricskiel/pykinectome\u003c/li\u003e\n\u003cli\u003eVirtanen 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 Mar;17(3):261\u0026ndash;72. \u003c/li\u003e\n\u003cli\u003eNichols TE, Holmes AP. Nonparametric permutation tests for functional neuroimaging: A primer with examples. Hum Brain Mapp. 2001 Oct 15;15(1):1\u0026ndash;25. \u003c/li\u003e\n\u003cli\u003eMilewski JPC\u0026amp; TM. 13 Permutation Testing | PSY317L Guidebook [Internet]. [cited 2025 Jul 15]. Available from: https://bookdown.org/curleyjp0/psy317l_guides5/permutation-testing.html\u003c/li\u003e\n\u003cli\u003eProkopenko D, Hecker J, Silverman EK, Pagano M, N\u0026ouml;then MM, Dina C, et al. Utilizing the Jaccard index to reveal population stratification in sequencing data: a simulation study and an application to the 1000 Genomes Project. Bioinformatics. 2016 May 1;32(9):1366\u0026ndash;72. \u003c/li\u003e\n\u003cli\u003ePeterson SJ, Foley S. Clinician\u0026rsquo;s Guide to Understanding Effect Size, Alpha Level, Power, and Sample Size. Nutrition in Clinical Practice. 2021;36(3):598\u0026ndash;605. \u003c/li\u003e\n\u003cli\u003eFrenkel-Toledo S, Giladi N, Peretz C, Herman T, Gruendlinger L, Hausdorff JM. Effect of gait speed on gait rhythmicity in Parkinson\u0026rsquo;s disease: variability of stride time and swing time respond differently. J NeuroEngineering Rehabil. 2005 Jul 31;2(1):23. \u003c/li\u003e\n\u003cli\u003eK\u0026ouml;nig N, Singh NB, Baumann CR, Taylor WR. Can Gait Signatures Provide Quantitative Measures for Aiding Clinical Decision-Making? A Systematic Meta-Analysis of Gait Variability Behavior in Patients with Parkinson\u0026rsquo;s Disease. Front Hum Neurosci [Internet]. 2016 Jun 30 [cited 2025 Jun 13];10. Available from: https://www.frontiersin.org/journals/human-neuroscience/articles/10.3389/fnhum.2016.00319/full\u003c/li\u003e\n\u003cli\u003eRoemmich RT, Nocera JR, Vallabhajosula S, Amano S, Naugle KM, Stegem\u0026ouml;ller EL, et al. Spatiotemporal variability during gait initiation in Parkinson\u0026rsquo;s disease. Gait \u0026amp; Posture. 2012 Jul 1;36(3):340\u0026ndash;3. \u003c/li\u003e\n\u003cli\u003eLowry KA, Smiley-Oyen AL, Carrel AJ, Kerr JP. Walking stability using harmonic ratios in Parkinson\u0026rsquo;s disease. Mov Disord. 2009 Jan 30;24(2):261\u0026ndash;7. \u003c/li\u003e\n\u003cli\u003eRochester L, Yarnall AJ, Baker MR, David RV, Lord S, Galna B, et al. Cholinergic dysfunction contributes to gait disturbance in early Parkinson\u0026rsquo;s disease. Brain. 2012 Sep;135(Pt 9):2779\u0026ndash;88. \u003c/li\u003e\n\u003cli\u003eHausdorff JM, Cudkowicz ME, Firtion R, Wei JY, Goldberger AL. Gait variability and basal ganglia disorders: Stride-to-stride variations of gait cycle timing in parkinson\u0026rsquo;s disease and Huntington\u0026rsquo;s disease. Movement Disorders. 1998;13(3):428\u0026ndash;37. \u003c/li\u003e\n\u003cli\u003eKleiner AFR, Souza Pagnussat A, Pinto C, Redivo Marchese R, Salazar AP, Galli M. Automated Mechanical Peripheral Stimulation Effects on Gait Variability in Individuals With Parkinson Disease and Freezing of Gait: A Double-Blind, Randomized Controlled Trial. Archives of Physical Medicine and Rehabilitation. 2018 Dec 1;99(12):2420\u0026ndash;9. \u003c/li\u003e\n\u003cli\u003eFrenkel-Toledo S, Giladi N, Peretz C, Herman T, Gruendlinger L, Hausdorff JM. Treadmill walking as an external pacemaker to improve gait rhythm and stability in Parkinson\u0026rsquo;s disease. Movement Disorders. 2005;20(9):1109\u0026ndash;14. \u003c/li\u003e\n\u003cli\u003eMajed L, Ibrahim R, Lock MJ, Jabbour G. Walking around the preferred speed: examination of metabolic, perceptual, spatiotemporal and stability parameters. Front Physiol [Internet]. 2024 Feb 9 [cited 2025 Jun 13];15. Available from: https://www.frontiersin.org/journals/physiology/articles/10.3389/fphys.2024.1357172/full\u003c/li\u003e\n\u003cli\u003ePadmanabhan P, Sreekanth Rao K, Gonzalez AJ, Pantelyat AY, Chib VS, Roemmich RT. The Cost of Gait Slowness: Can Persons with Parkinson\u0026rsquo;s Disease Save Energy by Walking Faster? J Parkinsons Dis. 2021;11(4):2073\u0026ndash;84. \u003c/li\u003e\n\u003cli\u003eBlandini F, Nappi G, Tassorelli C, Martignoni E. Functional changes of the basal ganglia circuitry in Parkinson\u0026rsquo;s disease. Prog Neurobiol. 2000 Sep;62(1):63\u0026ndash;88. \u003c/li\u003e\n\u003cli\u003eGroenewegen HJ. The Basal Ganglia and Motor Control. Neural Plasticity. 2003;10(1\u0026ndash;2):108384. \u003c/li\u003e\n\u003cli\u003eChristiansen CL, Schenkman ML, McFann K, Wolfe P, Kohrt WM. Walking economy in people with Parkinson\u0026rsquo;s disease. Movement Disorders. 2009;24(10):1481\u0026ndash;7. \u003c/li\u003e\n\u003cli\u003eSchenkman M, Hall D, Kumar R, Kohrt WM. Endurance Exercise Training to Improve Economy of Movement of People With Parkinson Disease: Three Case Reports. Physical Therapy. 2008 Jan 1;88(1):63\u0026ndash;76. \u003c/li\u003e\n\u003cli\u003eMancini M, Afshari M, Almeida Q, Amundsen-Huffmaster S, Balfany K, Camicioli R, et al. Digital gait biomarkers in Parkinson\u0026rsquo;s disease: susceptibility/risk, progression, response to exercise, and prognosis. NPJ Parkinsons Dis. 2025 Mar 21;11:51. \u003c/li\u003e\n\u003cli\u003eKing LA, Horak FB. Lateral Stepping for Postural Correction in Parkinson\u0026rsquo;s Disease. Archives of Physical Medicine and Rehabilitation. 2008 Mar 1;89(3):492\u0026ndash;9. \u003c/li\u003e\n\u003cli\u003eWegen EEH van, Emmerik REA van, Wagenaar RC, Ellis T. Stability Boundaries and Lateral Postural Control in Parkinson\u0026rsquo;s Disease. Motor Control. 2001 Jul 1;5(3):254\u0026ndash;69. \u003c/li\u003e\n\u003cli\u003eBryant MS, Rintala DH, Hou JG, Collins RL, Protas EJ. Gait variability in Parkinson\u0026rsquo;s disease: levodopa and walking direction. Acta Neurol Scand. 2016 Jul;134(1):83\u0026ndash;6. \u003c/li\u003e\n\u003cli\u003eTosserams A, Bloem BR, Ehgoetz Martens KA, Helmich RC, Kessels RPC, Shine JM, et al. Modulating arousal to overcome gait impairments in Parkinson\u0026rsquo;s disease: how the noradrenergic system may act as a double-edged sword. Transl Neurodegener. 2023 Mar 26;12:15. \u003c/li\u003e\n\u003cli\u003eMcDonald LM, Griffin HJ, Angeli A, Torkamani M, Georgiev D, Jahanshahi M. Motivational Modulation of Self-Initiated and Externally Triggered Movement Speed Induced by Threat of Shock: Experimental Evidence for Paradoxical Kinesis in Parkinson\u0026rsquo;s Disease. PLoS One. 2015;10(8):e0135149. \u003c/li\u003e\n\u003cli\u003eWu T, Hallett M, Chan P. Motor automaticity in Parkinson\u0026rsquo;s disease. Neurobiol Dis. 2015 Oct;82:226\u0026ndash;34. \u003c/li\u003e\n\u003cli\u003eRedgrave P, Rodriguez M, Smith Y, Rodriguez-Oroz MC, Lehericy S, Bergman H, et al. Goal-directed and habitual control in the basal ganglia: implications for Parkinson\u0026rsquo;s disease. Nat Rev Neurosci. 2010 Nov;11(11):760\u0026ndash;72. \u003c/li\u003e\n\u003cli\u003eBan R, Ahn J, Simpkins C, Lazarus J, Yang F. Dynamic gait stability in people with mild to moderate Parkinson\u0026rsquo;s disease. Clinical Biomechanics. 2024 Aug 1;118:106316. \u003c/li\u003e\n\u003cli\u003eWarmerdam E, Hansen C, Romijnders R, Hobert MA, Welzel J, Maetzler W. Full-Body Mobility Data to Validate Inertial Measurement Unit Algorithms in Healthy and Neurological Cohorts. Data. 2022 Oct;7(10):136. \u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"journal-of-neuroengineering-and-rehabilitation","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jner","sideBox":"Learn more about [Journal of NeuroEngineering and Rehabilitation](http://jneuroengrehab.biomedcentral.com/)","snPcode":"12984","submissionUrl":"https://submission.nature.com/new-submission/12984/3","title":"Journal of NeuroEngineering and Rehabilitation","twitterHandle":"@BioMedCentral","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Neurogeriatrics, gait analysis, kinectome, network, graphs","lastPublishedDoi":"10.21203/rs.3.rs-7415461/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7415461/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground:\u003c/strong\u003e Human gait involves complex coordination between musculoskeletal segments. This coordination is disturbed in Parkinson's disease (PD) and likely influenced by different walking speeds.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eObjectives: \u003c/strong\u003eTo investigate inter-segmental coordination during different walking speeds in people with PD (pwPD) using an unconstrained and data-driven network theory approach.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e Twenty-nine pwPD and 29 controls walked at preferred, fast and slow speeds. Data was collected using optical motion capture. Body segment accelerations were correlated pairwise to build kinectomes for each speed and movement direction. Anatomical body segments were defined as nodes and their co-accelerations as edges to build network graphs. The kinectomes and maximum-weighted graph patterns were compared between groups.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003e Permutation testing revealed no significant kinectome differences between groups across speeds or directions. Coordination deficits in the PD group were observed predominantly at preferred walking speed (162 significantly different graph patterns) in anteroposterior and mediolateral directions. At fast walking speed, 4 significantly different graph patterns were found in anteroposterior and vertical directions. Slow walking speed showed 1 significantly different pattern in mediolateral direction.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions:\u003c/strong\u003e PD affects inter-segmental coordination, becoming most apparent at preferred walking speed. This is surprising and highly relevant, as it is the most common gait condition in real life. 'Non-preferred' walking speeds in PD exhibit more control-like patterns, which could inform future treatment studies. The direction-specific coordination deficits could provide novel insights into patho- and compensatory mechanisms in PD gait.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTrial registration: \u003c/strong\u003eThe study is registered in the German Clinical Trials Register (DRKS00022998, registered on 04 Sep 2020).\u003c/p\u003e","manuscriptTitle":"Inter-segmental coordination patterns in Parkinson’s disease are particularly disturbed during preferred walking speed: a data-driven network approach","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-09 15:21:24","doi":"10.21203/rs.3.rs-7415461/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-09-29T04:55:36+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-09-26T05:26:08+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-09-24T00:28:05+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"152262661668100834066498124983182972336","date":"2025-09-05T11:56:36+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"163958362458420375268589723978071843973","date":"2025-09-03T05:03:19+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-03T04:05:46+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-08-22T10:11:13+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-08-22T10:09:49+00:00","index":"","fulltext":""},{"type":"submitted","content":"Journal of NeuroEngineering and Rehabilitation","date":"2025-08-20T08:40:26+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"journal-of-neuroengineering-and-rehabilitation","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"jner","sideBox":"Learn more about [Journal of NeuroEngineering and Rehabilitation](http://jneuroengrehab.biomedcentral.com/)","snPcode":"12984","submissionUrl":"https://submission.nature.com/new-submission/12984/3","title":"Journal of NeuroEngineering and Rehabilitation","twitterHandle":"@BioMedCentral","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"fc2271f9-4fd6-4a66-b72d-326c13bfe4f5","owner":[],"postedDate":"September 9th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-12-15T16:10:16+00:00","versionOfRecord":{"articleIdentity":"rs-7415461","link":"https://doi.org/10.1186/s12984-025-01835-1","journal":{"identity":"journal-of-neuroengineering-and-rehabilitation","isVorOnly":false,"title":"Journal of NeuroEngineering and Rehabilitation"},"publishedOn":"2025-12-12 15:57:11","publishedOnDateReadable":"December 12th, 2025"},"versionCreatedAt":"2025-09-09 15:21:24","video":"","vorDoi":"10.1186/s12984-025-01835-1","vorDoiUrl":"https://doi.org/10.1186/s12984-025-01835-1","workflowStages":[]},"version":"v1","identity":"rs-7415461","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7415461","identity":"rs-7415461","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}