Gait Analysis in Multiple Sclerosis: A Scoping Review of Advanced Technologies for Adaptive Rehabilitation and Health Promotion

Abstract

Background/Objectives: Multiple sclerosis (MS) often leads to gait impairments, even in early stages, and can affect autonomy and quality of life. Traditional assessment methods, while widely used, have been criticized because they lack sensitivity to subtle gait changes. This scoping review aims to map the landscape of advanced gait analysis technologies—both wearable and non-wearable—and evaluate their application in detecting, characterizing, and monitoring possible gait dysfunction in individuals with MS. Methods: A systematic search was conducted across PubMed and Scopus databases for peer-reviewed studies published in the last decade. Inclusion criteria focused on original human research using technological tools for gait assessment in individuals with MS. Data from 113 eligible studies were extracted and categorized based on gait parameters, technologies used, study design, and clinical relevance. Results: Findings highlight a growing integration of advanced technologies such as inertial measurement units, 3D motion capture, pressure insoles, and smartphone-based tools. Studies primarily focused on spatiotemporal parameters, joint kinematics, gait variability, and coordination, with many reporting strong correlations to MS subtype, disability level, fatigue, fall risk, and cognitive load. Real-world and dual-task assessments emerged as key methodologies for detecting subtle motor and cognitive-motor impairments. Digital gait biomarkers, such as stride regularity, asymmetry, and dynamic stability demonstrated high potential for early detection and monitoring. Conclusions: Advanced gait analysis technologies can provide a multidimensional, sensitive, and ecologically valid approach to evaluating and detecting motor function in MS. Their clinical integration supports personalized rehabilitation, early diagnosis, and long-term disease monitoring. Future research should focus on standardizing metrics, validating digital biomarkers, and leveraging AI-driven analytics for real-time, patient-centered care.

1. Introduction

Multiple sclerosis (MS) is a chronic autoimmune disease of the central nervous system that frequently impairs walking ability, balance, and coordination—functional domains that are essential for autonomy and quality of life. Gait deterioration may emerge even in the early stages of the disease, as a result of pyramidal, cerebellar, or sensory system involvement, leading to motor weakness, spasticity, ataxia, or proprioceptive deficits [1]. Although walking function is routinely assessed in clinical practice using tools such as the Timed 25-Foot Walk (T25FW) and the 6-Minute Walk Test (6MWT), these protocols typically yield summary values—like total distance or average speed—that may overlook dynamic within-task changes or compensatory fluctuations over time [2]. For instance, individuals with MS often demonstrate a gradual deceleration during prolonged walking tasks, yet the biomechanical underpinnings of this trend—such as stride length reduction or contact time increase—remain poorly understood. Moreover, many gait alterations in MS are too subtle to be detected through standard neurological examination or global disability scores like the Expanded Disability Status Scale (EDSS) [3].

In response to the limitations of traditional gait assessments, a wide spectrum of advanced gait analysis technologies has been developed to capture the complexity of walking in MS. These tools span from non-wearable laboratory-based systems to wearable and mobile technologies capable of real-world data collection [4]. These technologies can be broadly grouped into three main categories: (1) stationary laboratory-based systems, which offer high-precision biomechanical data; (2) wearable sensor-based systems, which enable continuous monitoring during daily activities; and (3) mobile- or smartphone-based solutions, which allow for accessible and scalable gait assessment in real-world settings. Among the non-wearable systems, three-dimensional motion capture provides high-precision tracking of joint angles and segment trajectories; force platforms measure ground reaction forces and center-of-pressure dynamics; and pressure-sensitive walkways quantify stride parameters with temporal granularity. Such systems have revealed, for example, decreased ankle plantarflexion and increased pelvic tilt in low-disability MS [5] and altered mechanical energy distribution from ankle to hip during walking [6]. Force-based analyses have also identified greater toe-off friction [7] and impaired gait initiation kinetics [8] as markers of early motor adaptation.

In parallel, wearable technologies offer ecologically valid insights into gait by enabling continuous, ambulatory monitoring. Inertial measurement units (IMUs) placed on the lower back, shank, foot, or trunk capture spatiotemporal parameters, joint coordination, turning dynamics, and variability in diverse contexts. Studies using these sensors have demonstrated high validity and reliability in measuring gait features across walking speeds and surfaces [9], and have successfully captured gait deterioration patterns over time in early-stage MS [10,11]. Additional wearable platforms include pressure-sensitive insoles, which provide plantar pressure distribution and center-of-pressure paths. Systems have shown strong concordance with lab-based instruments [12], offering accessible tools for clinical and home-based monitoring. More recently, smartphone-based gait assessments have emerged as cost-effective and scalable options. These platforms integrate accelerometers, gyroscopes, and user-friendly apps to evaluate metrics such as gait speed, step timing, and mediolateral sway, even in people with EDSS = 0 [13]. Some tests correlate with structural imaging markers [14] and demonstrate predictive value for clinical disability [15].

Beyond raw measurement, advanced analytical techniques have expanded the interpretive power of these systems. Cyclogram-based analyses visualize inter-joint coordination and have identified impairments in hip–knee–ankle coupling in early MS [16]. Semiogram frameworks integrate multiple gait dimensions (e.g., balance, propulsion, rhythm) into a single diagnostic profile, allowing differentiation between progressive MS subtypes [17]. Machine learning approaches, using regression-normalized features, have accurately classified gait deviations in MS with over 90% accuracy [18,19,20], while Lyapunov exponents and entropy-based metrics quantify local dynamic stability and movement regularity under stress. These tools and analytic models allow for a high-resolution understanding of gait as a reflection of neural function, compensatory strategy, and systemic resilience. They shift the clinical focus from broad functional outcomes (e.g., speed or distance) to subtle but meaningful markers of motor control, offering new opportunities for early diagnosis, individualized rehabilitation, and remote disease monitoring. This trend reflects a broader movement within neurorehabilitation toward the integration of technological tools—not only for assessment but also for enhancing engagement, personalization, and long-term monitoring across neurological conditions. Recent studies have emphasized the growing relevance of technology-assisted strategies in rehabilitation contexts, including cognitive and physical recovery frameworks [21,22].

This scoping review explores the current landscape of advanced gait analysis technologies and their application in assessing gait dysfunction in individuals with MS. Drawing from a broad range of studies that utilize both wearable and laboratory-based systems, it highlights the most commonly assessed gait parameters and examines their relevance to clinical outcomes, disease severity, and functional decline. In parallel, it surfaces emerging themes—such as digital biomarkers, cognitive-motor interference, and adaptive motor strategies—that reflect a growing shift toward multidimensional, neuro-informed gait evaluation. By mapping current evidence and methodological practices, the review aims to clarify research priorities and support the use of gait analysis as a tool for personalized, data-driven MS care.

2. Materials and Methods

2.1. Search Strategy

This scoping review was conducted in accordance with the 22-item Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) checklist [23]. The methodological protocol was designed a priori but was not registered in a public repository. A comprehensive literature search was conducted to identify relevant studies examining gait analysis in individuals with MS. Two independent reviewers (A.T. and P.V.) performed a comprehensive literature search across two electronic databases to identify all relevant studies published in the last decade. The search strategy utilized the advanced search functions of both databases, employing compound keyword combinations and Boolean operators. The search was performed across two major electronic databases, PubMed and Scopus, using predefined keywords and Boolean operators. The primary search terms included (“gait analysis” AND “multiple sclerosis”) for both databases, adapted to their respective syntaxes. Search strategies were adapted to the syntax of each database, and specific filters were applied to ensure the inclusion of peer-reviewed, English-language articles published within the predefined time frame. Details of the search strategy are presented in

Table 1. Literature Search Strategy.

Database	Search Date	Search Terms	Boolean Operators	Filters Applied	Timeframe
PubMed	8 May 2025	(gait analysis) AND (multiple sclerosis)	AND	Title/Abstract; English; Humans	Last 10 years
Scopus	8 May 2025	TITLE-ABS-KEY (“gait analysis” AND “multiple sclerosis”)	AND	Language: English; Publication Stage: Final; Document Type: Article	2015–2025

.

2.2. Selection Criteria

The screening and selection process of titles and abstracts was carried out using the Rayyan web-based platform for systematic reviews, which facilitated blinded decisions, tagging of exclusion reasons, and resolution of conflicts. In cases of disagreement during the screening or selection stages, a third reviewer (S.P.) was consulted, and consensus was reached through discussion. Only original, peer-reviewed full-text articles published in English were included in the present review. We excluded narrative or systematic reviews, clinical guidelines, conference abstracts, editorials, commentaries, unpublished materials, and clinical trials. Animal studies were also excluded. Aside from the requirement that participants be adults (≥18 years of age) and that the sample size be greater than 10 individuals, no other restrictions were imposed regarding study design or participant characteristics. The detailed inclusion and exclusion criteria are presented in

Table 2. Inclusion and Exclusion Criteria.

Inclusion Criteria	Exclusion Criteria
Studies published in the last decade	Studies not involving adults
Peer-reviewed original articles	Protocols, case reports, or conference abstracts
Articles written in English	Articles not available in full text
Human studies involving individuals with MS	Studies involving animals
Studies focused on gait analysis	Articles addressing only conventional clinical motor assessments
Sample size > 10 participants	Studies with <10 participants
Final publication stage	Retracted articles, personal views, or opinion pieces
Available in PubMed or Scopus databases	Clinical trials or studies not within scope of review

.

2.3. Data Extraction

Data extraction was performed using a predefined data form created in Excel. We recorded the authors, the year of publication, the type of study, the sample size, the method used for gait analysis, the main gait parameters assessed, the main outcomes, and whether a control group was included. This structured approach allowed for consistent and systematic comparison across the included studies.

2.4. Data Analysis

No statistical analysis or meta-analysis was performed due to the high heterogeneity observed among studies. Thus, a descriptive analysis was conducted by categorizing studies based on key variables such as study design, sample characteristics, MS subtype, gait analysis technology used, and gait parameters assessed. Frequencies and distributions were calculated for each category. Data were synthesized in tabular format and visualized using figures to facilitate thematic comparison across studies.

3. Results

3.1. Database Searches

A total of 672 records were initially identified through database searches: 500 from PubMed and 172 from Scopus. After removing 130 duplicate entries, 542 records remained for screening. Titles and abstracts were reviewed based on the predefined inclusion and exclusion criteria, leading to the exclusion of 198 non-relevant records. The remaining 344 full-text articles were assessed for eligibility. Among these, 231 studies were excluded for specific reasons. As a result, 113 studies were deemed eligible and were included in the final synthesis, as illustrated in Figure 1. Most studies were conducted in Europe and North America, with a predominance of cross-sectional designs. The majority of participants were individuals with relapsing-remitting MS (RRMS), although progressive subtypes were also represented. Commonly used technologies included wearable IMUs, 3D motion capture systems, and pressure-sensitive walkways. The most frequently assessed gait parameters included spatiotemporal metrics, gait variability, coordination, and dynamic stability. The detailed information and key characteristics of the studies included in this review are provided in Table S1.

3.2. Study Origin

The included studies originated from a diverse set of geographical regions, reflecting a global research interest in gait analysis among individuals with MS. The majority of the studies were conducted in Europe, contributing 61 [2,5,7,8,9,10,12,16,17,19,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72] out of the total 113 studies, followed by the United States with 30 studies [6,11,14,18,24,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98]. Asia accounted for 19 studies [1,13,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115], while Australia contributed 3 studies [116,117,118] (Figure 2). This geographical distribution underscores the concentration of research activity in Western countries, particularly in Europe and North America, where access to advanced motion analysis technology and clinical research infrastructure is more widespread. Given the complexity of gait impairments in MS and the need for objective, technology-driven assessment tools, it is noteworthy that regions with established neurorehabilitation and biomedical engineering programs appear to lead the field. However, the growing number of contributions from Asia suggests an expanding global recognition of the importance of quantitative gait evaluation in this population.

3.3. Study Design

The majority of the included studies (Figure 3) followed a cross-sectional design, accounting for 65.49% of the total sample. Observational studies made up 13.27%, while prospective studies comprised 7.08% of the studies. Smaller proportions were found for randomized controlled trials (RCTs) at 6.19% and experimental studies at 3.54%. Non-randomized interventional studies and secondary analyses were the least represented, each accounting for 0.88% of the total. This distribution reflects a strong emphasis on assessing gait characteristics at a single time point, with relatively few studies employing longitudinal or interventional designs

3.4. MS Patient Groups and Demographic Profiles

Across the 113 included studies, a total of 9052 individuals participated in gait analysis research related to MS. Of these, sex-specific data were available for the majority of cases, with 5839 females and 2669 males. This reflects a female predominance, which aligns with the well-documented epidemiological profile of MS. The demographic composition highlights the importance of considering sex-related differences in gait patterns within this population. The age of participants across the included studies ranged from 18 to 79.7 years, reflecting a wide spectrum of individuals with MS, from early adult onset to late-stage cases. The mean age, based on the average of reported study means, was approximately 47.5 years with a standard deviation around 10.5 years. This age profile is consistent with the natural history of the disease, which typically manifests in early to middle adulthood and progresses over time. Some studies reported separate age data for intervention groups or disease severity subgroups, highlighting the heterogeneity of the studied populations. Among the included studies, the majority of participants were diagnosed with RRMS, which is characterized by clearly defined attacks (relapses) of neurological dysfunction followed by periods of partial or complete recovery. Secondary progressive MS (SPMS) typically evolves from RRMS and involves a gradual worsening of neurological function over time, with or without relapses. Primary progressive MS (PPMS), in contrast, is marked by steadily worsening symptoms from the onset without early relapses or remissions. While RRMS was the most commonly reported subtype, both SPMS and PPMS were represented to a lesser extent in the included studies. Several studies included mixed populations or did not clearly specify the MS subtype. In total, RRMS appears to predominate in the reviewed literature, consistent with its higher global prevalence compared to progressive forms. However, the inclusion of progressive subtypes in many studies enhances the generalizability of the findings across the MS disease spectrum.

3.5. Reference Groups

A total of 1630 healthy control participants were included across the studies that incorporated a control group. Among those for whom sex was reported, 916 were female and 550 were male. Approximately 68 studies involved control groups. The inclusion of healthy controls in more than half of the studies enhanced the comparability of gait parameters and allowed for more robust inferences regarding MS-specific alterations in gait performance. None of the studies included a disease-control group other than MS patients. This absence may be attributed to the fact that most studies focused on identifying and quantifying gait alterations within the MS population rather than conducting differential analyses with other neurological disorders. While such an approach supports internal consistency, it limits conclusions regarding the specificity of gait features to MS. Including disease-control groups, such as individuals with Parkinson’s disease or stroke, could help distinguish MS-specific gait patterns from those shared across neurodegenerative conditions. This represents an important avenue for future comparative research.

3.6. Gait Parameters

Figure 4 presents a structured framework categorizing gait-related metrics into three primary domains: (1) Biomechanics and Neuromechanical Output, (2) Gait Control, Variability, and Adaptability, and (3) Gait Performance and Pattern. This classification was informed by prior conceptual models in neurorehabilitation and movement analysis that emphasize the multidimensional structure of gait [119,120]. Each domain encompasses distinct measurement dimensions essential for comprehensive gait assessment. The biomechanics domain includes variability metrics (e.g., stride time CV, entropy), joint kinematics (e.g., range of motion in hip, knee, ankle), and kinetic/force data such as ground reaction forces (GRF) and joint power. The gait control and adaptability domain incorporates functional performance tests (e.g., 6MWT, TUG), asymmetry indices, and sensor-derived metrics from IMUs and smartphones. Lastly, the gait performance and pattern domain focuses on spatiotemporal parameters (e.g., gait speed, stride length), stability metrics (e.g., center of pressure sway, Lyapunov exponents), and cognitive-motor interference during dual-task conditions. This figure highlights the multidimensional nature of gait analysis and underscores the importance of integrating multiple measurement modalities—spanning from wearable technology to high-resolution lab-based tools—to capture the full spectrum of gait function across different populations and settings.

Figure 5 presents a pie chart illustrating the distribution of studies based on the number of gait parameters analyzed. The majority of studies (50.6%) examined a relatively small number of parameters (1–9), suggesting a focused analytical approach in over half of the cases. About 29.4% of studies assessed between 10 and 19 parameters, indicating a broader, though still manageable, scope of analysis. A smaller proportion, 11.8%, evaluated 20 to 39 parameters, reflecting more comprehensive gait assessments. Only 8.2% of studies analyzed more than 39 parameters, demonstrating the rarity of highly detailed or multidimensional gait profiling.

3.7. Gait Measurement Tools

A wide range of gait measurement tools have been employed across studies to capture spatiotemporal, kinematic, and kinetic gait parameters with high accuracy. Among the most frequently used systems is the GAITRite® electronic walkway, valued for its ease of use and reliability in capturing step-related metrics [11,29,31,34,37,50,59,75,85,101,102,104,106,107,109,117,118,121]. Three-dimensional motion capture systems (e.g., Vicon, BTS Bioengineering, Qualisys) offer gold-standard data through high-speed cameras and marker-based tracking, often combined with force plates to assess kinetics [5,6,7,8,9,16,40,41,43,44,51,54,58,64,65,66,67,68,69,70,71,86,97,100,105,110,113,116]. Wearable inertial sensors such as Opal IMUs, Xsens MTw, and Physilog^® 5 are extensively used for ambulatory assessments, enabling real-world gait analysis outside laboratory settings. These are frequently placed on the lower back, shanks, or feet to monitor acceleration, angular velocity, and orientation [2,24,25,26,33,35,36,39,45,46,47,49,52,53,55,56,58,60,63,73,78,79,80,81,84,88,91,98,99,108,111,112]. Pressure-sensitive systems like the Zeno™ Walkway and Footscan^® allow for detailed plantar pressure distribution analysis [12,27,77,89,90]. Additional tools include smartphone-based sensors [13,14,48,57,96], video-based 2D analysis [61,62,72,115], electromyography (EMG) for muscle activation [45,46,66,67,72], and instrumented treadmills (e.g., C-Mill) [18,38,63,83,87,93,103,106,114]. Some setups integrate multimodal systems—combining IMUs, pressure insoles, and video capture—to enhance ecological validity and sensor fusion for advanced gait modeling [19,95].

3.8. Study Setting for Gait Analysis

The setting in which gait analysis is conducted significantly influences the type, accuracy, and applicability of the collected data. The majority of studies are carried out in controlled laboratory environments, where researchers can utilize high-precision equipment such as 3D motion capture systems, force plates, and instrumented treadmills. These settings are ideal for detailed biomechanical assessments under standardized conditions [5,6,7,8,9,11,16,17,18,19,27,28,29,30,31,37,38,39,40,41,42,43,44,45,48,51,53,54,55,56,58,64,65,66,67,68,69,70,71,72,73,76,77,80,82,83,84,86,87,88,89,90,91,92,93,94,95,98,99,100,103,104,105,107,110,112,113,114,115,116,118]. However, a growing number of studies are expanding into clinical laboratories, aiming to evaluate gait in patient populations using tools like the GAITRite^® system and wearable inertial sensors [1,2,10,12,13,24,32,46,47,49,50,52,59,60,61,62,63,79,85,97,101,102,106,108,109,117,121]. In parallel, real-world, and free-living environments, including participants’ homes, these have gained attention for their ability to capture ecologically valid gait patterns. In these settings, wearable IMUs, smartphones, and pressure insoles are commonly used to monitor gait during daily activities, offering insights into functional mobility that may not be apparent in laboratory settings [14,33,36,57,74,81,82]. Some studies employ hybrid designs, combining laboratory assessments with remote monitoring to balance measurement precision and real-world relevance [25,26,34,35,75,78,111]. To ensure comparability across environments, these studies often rely on consistent sensor placement, alignment of gait metrics, and normalization techniques. In some cases, machine learning models are used to harmonize data streams or account for context-specific variability. However, the lack of standardized integration protocols remains a challenge for cross-setting synthesis and clinical translation. This variety in study settings reflects the evolving need to understand human gait both in standardized contexts and in everyday life.

4. Discussion

4.1. Digital Gait Biomarkers: From Measurement to Meaning

Recent studies increasingly frame gait in MS not merely as a measure of locomotion speed, but as a multidimensional digital phenotype—one that encapsulates underlying neural processes, compensatory mechanisms, and disease burden. This reconceptualization of gait as a brain-derived output provides fertile ground for developing sensitive biomarkers that move beyond conventional metrics like gait velocity and step length. Among these advanced metrics is gait smoothness—quantified using the Log Dimensionless Jerk (LDJ), a unitless metric derived from the third derivative of position (jerk), normalized by movement duration and amplitude. LDJ reflects the efficiency and coordination of movement, with lower values indicating smoother gait patterns, and has shown promise as a reliable and responsive indicator [122]. Specifically, mediolateral smoothness demonstrated both high discriminative ability and responsiveness to rehabilitation in individuals with MS [30]. Similarly, coordination-based measures like the Phase Coordination Index (PCI), a measure that captures both the variability and accuracy of left–right step timing, reflecting bilateral coordination during gait, have revealed persistent bilateral coordination deficits even in mildly affected individuals [88], suggesting sensitivity to early motor system disruption. Moreover, variability and asymmetry have emerged as early indicators of both fall risk and disease progression. For example, step length variability was significantly higher in fallers, and such variability also showed predictive value for energy inefficiency before falls occur. The Phase Coordination Index, gait asymmetry, and other markers of gait regularity worsened with increasing disability levels and during extended walking tasks such as the 6MWT [48,60,112]. These findings reinforce the idea that dynamic motor coordination and rhythm—rather than static spatiotemporal parameters—may reflect underlying neurophysiological decline. The concept of “gait as a brain output” is also supported by evidence linking digital gait features with cortical integrity. For instance, impairments in dual-task turning and stride velocity correlated with cortical thinning and white matter changes in regions such as the precuneus [99], while smartphone-based metrics correlated with motor cortex activity and disease severity [14,96].

These findings highlight the clinical potential of digital gait biomarkers as sensitive indicators of both neurodegeneration and neuroplasticity. Their use in predictive modeling, particularly in forecasting relapse or monitoring treatment response, could redefine gait analysis from a passive observational tool into an active clinical decision aid.

4.2. Real-World Gait as a Diagnostic Lens

Traditional gait assessments—often performed under structured, laboratory conditions—fail to capture the complex, adaptive, and emotionally mediated nature of real-world walking in individuals with MS. Increasing evidence highlights that daily walking behaviors, including pauses, short bouts, turns, and transitions, offer deeper insight into the interaction between physical capacity and environmental/contextual demands. For example, Kushner et al. [81] demonstrated that 28% of falls occurred within one second of gait initiation, and that frequent pauses during walking nearly tripled fall risk. These real-world markers—especially increased movement complexity and trajectory irregularity—suggest that navigation instability, rather than pure motor capacity, underlies functional decline. Similarly, Arpan et al. [74] demonstrated that daily-life gait and turning metrics—captured using wearable inertial measurement units (IMUs) placed on the lower back—could predict fall risk with 85% sensitivity. Key predictive features included turning velocity, stride variability, and trunk rotation during free-living ambulation. This finding underscores the diagnostic value of ecologically valid gait features derived from continuous real-world monitoring. The variability of gait across short walking bouts—a hallmark of real-world ambulation—has also been linked to disability levels in MS. Storm et al. [25] found that in free-living conditions, shorter walking bouts revealed more pronounced gait deficits, including reduced pace and increased temporal fluctuations. Cantu et al. [33] further validated that wearable sensors can accurately detect real-world gait events, though accuracy declines with slower walking speeds and shorter bouts, which are common among individuals with MS. Beyond motor performance, real-world gait becomes an emotional and behavioral interface. VanNostrand et al. [82] reported that gait quality metrics (e.g., stride regularity and speed) were closely associated with cognitive processing speed [123,124], disability level, and fear of falling, while total walking time and bout count bore no significant correlation—highlighting that quality, not quantity, best reflects functional status. Additionally, Jeng et al. [78] observed that sedentary behavior patterns, rather than total sedentary time, were predictive of gait deterioration, with more frequent movement interruptions being beneficial for mobility.

These observations emphasize that real-world gait sensing—through ambient or wearable technologies—holds transformative potential for personalized, continuous MS monitoring. Moving beyond episodic clinical snapshots, ambient gait analysis can reflect moment-to-moment adaptations to both internal states (e.g., fatigue, fear, mood) and external stimuli (e.g., terrain, distractions), positioning gait as a dynamic readout of holistic health.

4.3. Cognition in Motion: Dual-Task Gait and Neural Efficiency

Dual-task gait assessment—where individuals walk while simultaneously engaging in a cognitive task—has emerged as a sensitive probe of cognitive-motor integration in people with MS. Far from being a mere test of multitasking, it functions as a “stress test” for the neural system, unmasking latent deficits that remain hidden under single-task conditions. Several studies support the utility of dual-task walking in revealing early-stage dysfunction. For example, Chen et al. [99] identified that stride velocity during dual-task circular walking were among the most accurate markers distinguishing individuals with MS from controls, with both structural brain correlates (e.g., cortical thinning) and high diagnostic accuracy (sensitivities and specificities >84%). Similarly, other studies [29,39] confirmed that cognitive-motor interference is observable in both RRMS and progressive forms of MS—even when EDSS scores are similar. Beyond group comparisons, dual-task paradigms offer insight into compensatory mechanisms and neural efficiency. De Aratanha et al. [95] used fNIRS to demonstrate increased cortical activation during dual-task walking, indicating early recruitment of motor planning regions—a likely compensatory response to latent inefficiencies. The notion that walking draws upon shared cognitive networks (akin to bilingualism or musical dual-tasking) suggests that dual-task gait may function as a non-invasive behavioral marker of cerebral reserve and adaptability. Additional work [83] showed that gait variability under dual-task conditions was exacerbated in older individuals with MS but could improve within-session with short-term practice, indicating that neural plasticity and motor learning processes remain partially intact. Likewise, Cruz et al. [38] demonstrated that dual-task uncertainty increased balance instability and reduced stride length, confirming the added vulnerability introduced by cognitive demands.

These studies position dual-task gait not only as a marker of cognitive-motor performance, but also as a dynamic window into neurofunctional resilience and compensatory potential. As such, it holds promise for early diagnosis, monitoring of progression, and evaluation of interventions aimed at enhancing cerebral efficiency.

4.4. Redefining Fall Risk: Movement Disruptions at the Margins

Traditional assessments of fall risk in MS often focus on spatiotemporal gait variability during steady walking; yet, emerging evidence highlights the critical role of transitional movements—such as gait initiation, turning, and unplanned stopping—as moments of heightened vulnerability. These movement margins demand rapid coordination and anticipatory control, rendering them more sensitive to neurological disruption than continuous walking. For instance, Tajali et al. [110] showed that short-term Lyapunov exponents, which estimate the sensitivity of a dynamic system (e.g., gait pattern) to small perturbations—reflecting local dynamic stability—during single-task treadmill walking were the strongest predictors of future falls. Interestingly, dual-task conditions did not add predictive value, suggesting that motor instability at critical transitions may outweigh cognitive burden in determining fall risk. Similarly, Salehi et al. [105] demonstrated that inter-segmental coordination variability (e.g., foot–shank and shank–thigh phase relationships) was significantly higher in MS fallers than non-fallers, particularly during the swing phase, when precise anticipatory control is required. These disruptions suggest failures in internal motor timing and adaptability at transitional phases. Further study [91] found that trunk–foot coordination metrics—especially in the frontal plane—distinguished fallers from non-fallers with high accuracy (AUC = 0.92). These metrics capture not only balance stability but also adaptive control during directional changes, offering an advantage over traditional stride-to-stride variability indices. Another study [85] identified backward walking velocity as a more powerful predictor of fall status than forward walking. Backward walking challenges anticipatory postural adjustments and sensory-motor integration, functioning as a clinical analogue of transitional instability. The increased double support time and variability observed during backward gait underline its value in exposing hidden instabilities. Overall, these conclusions argue for a shift in fall-risk assessment: from evaluating general gait patterns to monitoring dynamic transitions, where anticipatory failure, rather than motor slowness per se, becomes the defining issue. Wearable sensors and task-specific protocols targeting start-stop-turn transitions offer promising avenues for more nuanced and actionable fall risk prediction.

4.5. When the Body and the Mind Disagree: The Perception–Performance Gap

In MS, mobility is not merely a function of muscle strength or gait symmetry—it is also shaped by how individuals perceive their ability to move. A growing body of evidence reveals a perception–performance gap, where self-reported mobility limitations are influenced more by internal states like fatigue [125], fear of falling, and depression than by objectively measured gait parameters. For instance, Kalron et al. [101] demonstrated that individuals with MS who reported depressive symptoms perceived greater walking impairment, despite showing no significant differences in objective gait metrics after controlling for EDSS and demographic variables. This dissociation suggests that mood states may color subjective mobility assessments independently of physical capability. Similarly, other studies [32,49] explored the complex relationship between fatigue and gait. While objective gait features like stride time and heel strike angle correlated with state fatigue, shorter tests captured perceived fatigue as well as full-length protocols—highlighting the subjective dimension of fatigue that may not be fully reflected in traditional spatiotemporal measures. Gait quality deterioration during the 6-Minute Walk Test—including increased variability and reduced toe-off angle—occurred even in participants who did not exhibit distance-based fatigability, reinforcing the mismatch between external output and internal experience. This perceptual disconnect also extends to fall risk and balance confidence [50,103,105,110]. Patient-reported outcomes outperformed both sensor-based and clinician-rated gait data in identifying fallers. This suggests that subjective awareness of instability may reflect subclinical motor or cognitive fluctuations not captured during standardized testing.

These findings support the view that interoceptive accuracy—the brain’s ability to sense and interpret bodily signals—plays a key role in the experience of mobility. When this system is distorted by depression, anxiety, or fatigue, individuals may report high levels of disability despite retaining functional gait. Such mismatches could have implications for treatment responsiveness, as interventions targeting physical performance may fail to translate into improved quality of life unless perceptual distortions are addressed.

4.6. Smart Dysfunction: Adaptive Strategies in MS Gait

Gait abnormalities in MS are often interpreted as signs of motor breakdown, yet many reflect intelligent adaptations to neurological constraints. These compensatory strategies—although visually atypical—may function to preserve mobility, reduce risk, or optimize effort. Recognizing this motor creativity under constraint is essential for redefining therapeutic goals. One such adaptation involves shifting propulsion proximally: instead of relying on ankle push-off, individuals redistribute effort to the hip. This hip-dominant strategy, observed during walking, reflects a redistribution of mechanical work that prioritizes forward momentum despite distal weakness, as evidenced by gait analyses showing reduced ankle moments and increased hip contribution [6].

Muscle coactivation provides another example. Rather than signaling dysfunction alone, increased activation at the knee during single support and at the ankle during double support appears to enhance joint stiffness and stability. Patterns of distal and proximal coactivation, depending on gait phase, may offer functional benefits for maintaining balance in destabilizing conditions [68]. Changes in trunk and pelvic motion—especially in the sagittal and frontal planes—also emerge as compensatory mechanisms. Greater excursions in these regions have been linked to reduced lower-limb strength and endurance, suggesting a redistribution of mechanical demand to maintain gait continuity. Notably, increased frontal plane pelvis motion correlates with better performance on walking and endurance tests, such as the T25FW and 2MWT, supporting its adaptive role [92]. Similarly, more pronounced pelvis excursions during dynamic tasks like stair climbing have been associated with better functional outcomes, indicating that these kinematic changes may not require correction, but rather fine-tuning [86].

Altogether, these outcomes suggest that not all deviations from normative gait patterns are detrimental. Some may represent efficient, trainable responses to impaired control or strength. This calls for a shift in rehabilitation philosophy: from restoring “normal” gait to optimizing the most effective compensations. Identifying which patterns are adaptive and sustainable—and which cross the threshold into maladaptation—could guide more personalized, context-sensitive interventions in MS care. However, it is important to note that most of these findings stem from cross-sectional or short-duration intervention studies. Therefore, the long-term persistence, adaptability, and clinical relevance of these compensatory strategies remain uncertain. Future longitudinal research is needed to determine whether such motor adaptations are sustained over time and continue to offer functional benefits.

4.7. Measuring Motor Resilience: What Gait Reveals Post-Intervention

While improvements in walking speed or distance are common markers of rehabilitation success in MS, they may underestimate the depth of neuromotor change. A growing body of evidence highlights that subtle gait features—such as coordination, variability, and double support time—respond more sensitively to therapeutic interventions, offering a window into motor resilience and neuroplastic adaptation.

Interventions like rhythmic auditory stimulation (RAS) have shown consistent effects on spatiotemporal gait coordination. In one study, RAS led to improvements in stride length, cadence, and gait speed, while also reducing double support time and enhancing stability and rhythm—even when between-group differences in some parameters were not statistically significant [100]. These changes point to refined motor control, beyond gross performance metrics [126,127]. Similarly, hippotherapy resulted in gains not only in walking velocity but also in stride length and cadence, as captured by both performance tests and objective gait measures [75]. Such outcomes suggest an integrative neuromuscular effect, where dynamic trunk and limb engagement contribute to more stable, symmetric gait patterns. Technologically supported interventions have also shown promise. Split-belt treadmill training, for instance, encouraged propulsion reorganization—especially in the more affected limb—by increasing dorsiflexion and modulating plantarflexion timing, reflecting an adaptive shift in force distribution [87]. These changes signify a capacity for targeted motor recalibration, reinforcing the idea that the gait system in MS retains latent plastic potential. Even more unconventional approaches, such as visual biofeedback cycling, have demonstrated improvements in coordination and functional ambulation, surpassing the outcomes of traditional home exercise programs [34]. These gains—measured through increased step length and improved Functional Ambulation Profile scores—reflect reorganization at both the neural and biomechanical level.

Rather than viewing these improvements as isolated functional gains, they may best be interpreted as evidence of locomotor resilience: the nervous system’s ability to reorganize, adapt, and improve under challenge [128]. Gait thus becomes not only a behavioral outcome, but a sensitive endpoint for detecting neuroplastic shifts in response to rehabilitation.

4.8. Toward Standardization of Gait Assessment Methods

One of the main findings of this review is the marked heterogeneity in both the measurement tools and the gait parameters used across studies. Despite this variability, a pattern of convergence becomes apparent: several digital biomarkers—such as gait speed, step and stride length, gait variability, asymmetry index, and smoothness—are consistently reported across a variety of technologies, including IMUs, motion capture systems, smartphone applications, and instrumented walkways. This consistency suggests the potential to define a core set of gait parameters that could serve as standardized digital biomarkers for assessing neurological function in individuals with MS.

Our synthesis highlights five parameters—gait speed, step or stride length, gait variability, asymmetry index, and gait smoothness—as particularly promising candidates for standardization, based on both their frequency of use and clinical relevance. However, our review also revealed that few studies provide explicit computational definitions for these metrics, and most rely on device-specific or proprietary methodologies. This lack of uniformity poses a significant barrier to comparability, reproducibility, and large-scale synthesis.

To address this gap, we propose a roadmap (Figure 6) for standardization that emphasizes three foundational pillars: the development of unified computational definitions for key gait parameters, the establishment of minimum technical specifications for measurement devices (such as sampling frequency and minimum step count), and the creation of open-source algorithms for parameter extraction. Advancing in this direction could facilitate cross-platform interoperability, support more accurate and personalized rehabilitation strategies, and enable robust longitudinal remote monitoring. By identifying these issues and highlighting potential solutions, the present review offers a preliminary framework to inform future efforts toward standardization in digital gait analysis for MS.

4.9. Strenghts and Limitations of the Study

This scoping review offers a comprehensive synthesis of the current landscape of gait analysis technologies applied to individuals with MS, highlighting the evolution of digital gait biomarkers, methodological diversity, and the potential of real-world and cognitive-motor assessments. One of the primary strengths of this review lies in its systematic and rigorous methodology, which adhered to the PRISMA-ScR guidelines and involved a transparent, multi-reviewer screening process. The inclusion of a broad range of studies—spanning wearable and non-wearable technologies, various MS subtypes, and different study designs—enhances the generalizability and relevance of the findings. Furthermore, the categorization of gait domains, integration of sensor-based and clinical outcomes, and exploration of innovative concepts such as gait adaptability, cognitive-motor interference, and perception–performance discrepancies reflect a multidimensional approach to understanding gait in MS. Another notable strength is the emphasis on emerging paradigms, including the use of gait as a proxy for brain function, real-world gait monitoring, and motor compensation strategies. These themes underscore a shift toward ecological validity and patient-centered outcome measures, which are critical for personalized care and longitudinal disease monitoring.

However, several limitations must be acknowledged. First, the review was limited to studies published in English and indexed in PubMed and Scopus, which may have excluded relevant research published in other languages or databases. Second, the heterogeneity in study methodologies—including differences in gait measurement tools, analytic parameters, and participant characteristics—prevented quantitative synthesis or meta-analysis. This variability also complicates direct comparison across studies, potentially limiting the ability to draw definitive conclusions about specific tools or biomarkers. Moreover, while the review captured a wide range of gait parameters, the inconsistent reporting of cognitive, emotional, and contextual factors across studies may have underestimated the full impact of psychosocial influences on gait behavior. The lack of standardized terminology and outcome measures across studies also presents a barrier to the synthesis and clinical translation of findings.

4.10. Clinical Implications and Future Studies

This review highlights the evolving clinical role of gait analysis in MS, shifting from basic spatiotemporal assessments to a multidimensional tool that reflects neurological integrity, cognitive load, and emotional states. The growing use of wearable technologies enables real-world, continuous monitoring of gait, offering deeper insight into daily functioning and disease progression than traditional lab-based tests. Clinicians are encouraged to incorporate advanced gait metrics—such as coordination, variability, and performance during dual-task or transitional movements—into routine assessments. Furthermore, attention should be paid to the mismatch between patients’ self-perceived mobility and objective gait performance, which is often influenced by fatigue, depression, and fear of falling.

Future research should aim to standardize gait assessment protocols, validate digital biomarkers across diverse populations, and explore how gait features respond to interventions over time. Longitudinal studies combining gait data with neuroimaging, cognitive metrics, and patient-reported outcomes will be critical for developing predictive models of disease activity and treatment response. Integrating artificial intelligence and machine learning with wearable data could support personalized, real-time clinical decision-making. Ultimately, gait analysis has the potential to become a cornerstone of individualized MS management, enabling proactive care and more nuanced rehabilitation strategies tailored to each patient’s unique neurological and functional profile.

5. Conclusions

This review underscores a shift in how gait is understood in MS not just as a biomechanical process, but as a brain-driven behavior influenced by neural, cognitive, emotional, and environmental factors. Through the analysis of 113 studies, the emergence of digital gait biomarkers such as smoothness, coordination, and variability is shown to offer promising avenues for early diagnosis, risk assessment, and treatment monitoring. Real-world and dual-task assessments, along with the recognition of compensatory gait patterns, are highlighted as crucial for capturing functional impairments more accurately. The findings point toward a future in which wearable technologies, machine learning, and integrated assessment tools enable personalized, real-time gait monitoring in MS care. The review calls for further research to validate these approaches across diverse populations and clinical contexts, ultimately aiming to enhance proactive, patient-centered management strategies.