Enhanced Assessment of Gait Dynamics in Multiple Sclerosis:

A Signal Processing Approach for Extracting Range of Motion

Using Wearable IMUs

Dunya Moradi

, Kardelen Akar

2,4

, Hussein Youssef

2,4

, Ayse Altintas

2,4

, Atay Vural

2,4

and Beren Semiz

3,4

Graduate School of Sciences and Engineering, Koc University, Istanbul, Turkey

Department of Neurology, Koc University, Istanbul, Turkey

Department of Electrical and Electronics Engineering, Koc University, Istanbul, Turkey

Koc University Research Center for Translational Medicine (KUTTAM), Koc University, Istanbul, Turkey

Keywords: Multiple Sclerosis, Inertial Measurement Units, Range of Motion, Timed 25-Foot Walk (T25FW), Biomedical

Signal Processing.

Abstract: This study investigates the gait dynamics and motor impairment severity in individuals with multiple sclerosis

(MS) by analyzing lower limb range of motion (ROM) using data from inertial measurement units (IMUs)

collected during the Timed 25-Foot Walk (T25FW) clinical task. Forty-eight participants were categorized

into two MS groups based on motor impairment severity (16 MS patients with low motor impairment, 16 MS

patients with moderate to severe motor impairment) and 16 healthy control group. IMU raw data of

accelerometer and gyroscope from the feet sensors with respect to the lumbar region, were processed using a

Butterworth filter and an Extended Kalman Filter to obtain accurate orientation, followed by quaternion to

Euler angle conversion for calculating ROM. When the ROM-extracted statistical and time domain features

were compared, there were significant differences in ROM characteristics among groups, particularly

highlighting the increased variability and reduced control in participants with severe motor impairments.

ROM-extracted features such as kurtosis, skewness, and entropy underscored the asymmetrical and irregular

motion patterns in advanced MS cases. These findings support the potential of IMU-derived ROM metrics as

biomarkers for tracking MS disease progression and tailoring rehabilitation.

1 INTRODUCTION

Multiple Sclerosis (MS) is a chronic autoimmune

disorder that affects the central nervous system,

leading to a wide range of physical and cognitive

impairments. One of the key manifestations of MS is

the disruption of motor functions, which can severely

impact the quality of life for those affected. Patients

with MS often experience muscle weakness, spasticity,

and fatigue, all of which can hinder their ability to

perform daily activities. These motor impairments lead

to a reduction in mobility, increasing the risk of

secondary complications such as falls, contractures,

and decreased overall functional capacity (LaRocca,

2011), (Heesen et al., 2008).

Gait asymmetry in MS patients can arise from

various factors, including spasticity, muscle

weakness, or impaired motor coordination.

Spasticity, characterized by increased muscle tone

and resistance to movement, is a common symptom

of MS and can significantly affect gait dynamics

(Coca-Tapia et al., 2021; GÜLŞEN et al., 2024).

Range of motion (ROM) is a critical metric in

assessing the flexibility and functional capabilities of

joints. It quantifies the extent to which a joint can move

through its intended motion patterns, providing

valuable insights into a patient's musculoskeletal

health. In the context of MS, monitoring ROM serves

several important purposes, such as assessing mobility,

informing personalized rehabilitation, monitoring

disease progression, preventing complications, and

enhancing quality of life (Soucie et al., 2011).

Wearable devices facilitate real-world patient

monitoring and provide valuable biomarkers for

symptoms and behaviors associated with gait

disorders, thereby enhancing clinical assessments and

972

Moradi, D., Akar, K., Youssef, H., Altintas, A., Vural, A. and Semiz, B.

Enhanced Assessment of Gait Dynamics in Multiple Sclerosis: A Signal Processing Approach for Extracting Range of Motion Using Wearable IMUs.

DOI: 10.5220/0013260300003911

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 18th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2025) - Volume 1, pages 972-979

ISBN: 978-989-758-731-3; ISSN: 2184-4305

enabling personalized treatment plans (Huang et al.,

2024). Recent advancements in wearable

technologies and motion analysis techniques, such as

inertial measurement units (IMUs), have made it

possible to accurately assess ROM in real time. These

innovations provide clinicians and researchers with

valuable tools to evaluate joint movements in

dynamic environments, paving the way for a more

comprehensive understanding and management of

MS-related motor impairments (Grisbrook et al.,

2018), (Blandeau et al., 2023).

Considering the preliminary evidence suggesting

potential benefits of exercise programs on lower limb

flexibility in MS patients (Torres-Pareja et al., 2019),

we believe that a detailed analysis of ROM features

will enhance our understanding of these patients'

physical capabilities. By extracting and analyzing

these beneficial ROM features, we can uncover critical

insights into the mobility dynamics of individuals with

MS, which may inform and refine treatment protocols.

Accurate measurement and monitoring of ROM can

facilitate the identification of specific movement

impairments and provide objective data that healthcare

professionals can use to tailor rehabilitation. Thus,

leveraging the advancements in motion analysis

technology can play a pivotal role in transforming the

management of motor impairments associated with

MS, ensuring that therapeutic approaches are both

effective and individualized.

In this study, we aimed to process IMU data

collected from MS patients during the T25FW test to

evaluate their gait and movement patterns.

Specifically, we focused on calculating the ROM of

the feet relative to the lumbar region, which served as

a reference point for assessing the dynamics of lower

body movements during walking. This analysis was

crucial in identifying and quantifying gait

abnormalities in MS patients, aiding in tracking

disease progression, evaluating balance and

coordination issues, and designing personalized

rehabilitation protocols.

2 METHODOLOGIES

2.1 Participants

The study was conducted under a protocol approved

by the Koc University Institutional Review Board

(2020.418.IRB.157) and all participants provided

written consent. We recruited a total of 48

participants, including 32 individuals with multiple

sclerosis (pwMS) and 16 healthy controls. The pwMS

group was divided into two distinct categories based

on their performance in the T25FW task. Group A,

comprising 16 patients (7 male, 9 female), completed

the T25FW in under 5 seconds, indicating mild motor

impairment. In

contrast, Group B consisted of 16

patients (3 male, 13 female) with moderate to severe

motor impairment, evidenced by a T25FW duration

exceeding 9 seconds. The healthy control group

included 16 participants (8 male, 8 female) and served

as a baseline for comparison. This categorization

enabled us to evaluate differences in gait impairment

levels among pwMS and to assess the precision of our

analytical approach in distinguishing between various

stages of motor impairment and healthy gait patterns

Table 1: Participants Mean T25FW Score and Age.

Mean±SD

Healthy

control

Group A Group B

T25FW

duration

(seconds)

4.95±0.79 4.24±0.39 14.17±4.31

Age

(years)

54.15±12.84 38.01±9.12 56.37±11.47

2.2 IMU Data Collection

In this study, IMU data were collected using the

APDM MobilityLab system (APDM Inc., Portland,

OR, USA) during T25FW task. IMU data were

sampled at a frequency of 128 Hz and the sensors

were synchronized in real-time using a dedicated

wireless access point, ensuring precise temporal

alignment. Data processing was performed using

Python for signal filtering, feature extraction and

statistical analysis.

For the T25FW, participants walked 7.62 meters

as quickly as possible (Figure 1), with the time

recorded manually using a stopwatch. Throughout the

task, three OPAL sensors (Oregon, USA) were placed

on participants’ feet and lumbar region to measure

spatiotemporal gait parameters, trunk and turning

angles, and balance during quiet standing (Figure 2).

The system comprised wireless IMU sensors, a

docking station, and a wireless access point for real-

time data synchronization and transmission.

Figure 1: 25 Foot walking Clinical task.

Enhanced Assessment of Gait Dynamics in Multiple Sclerosis: A Signal Processing Approach for Extracting Range of Motion Using

Wearable IMUs

973

Figure 2: Sensor placement for IMU data collecting.

2.3 Calculation of Range of Motion

A Butterworth bandpass filter with a low cutoff

frequency of 0.01 Hz and a high cutoff frequency of

20 Hz was applied to the IMU raw data. For an

accurate orientation estimation, an Extended Kalman

Filter (EKF) was used. The EKF was implemented to

fuse accelerometer and gyroscope data, leveraging

their complementary strengths. The prediction step

utilized gyroscope data to estimate orientation

changes, while the update step corrected these

estimates using accelerometer data, reducing the

impact of drift and noise. The process equations were

based on quaternion kinematics. This iterative

approach ensures accurate orientation estimation,

particularly in dynamic conditions. This was critical

for accurately calculating the ROM from IMU raw

data (Senesh & Wolf, 2009), (Das et al., 2018),

(Keskinoğlu & Aydın, 2021). Precise orientation data

were essential for determining how much the feet

rotated relative to the lumbar region during walking.

Hence, fused accelerometer and gyroscope data could

achieve more accurate 3D orientation estimates for

each foot relative to the lumbar region. Initially,

quaternions were used to represent orientation due to

their ability to avoid gimbal lock, which can occur

when using Euler angles. Quaternions offer a more

stable representation of 3D rotations, especially when

continuous motion is involved, as they do not suffer

from the singularities that Euler angles do. However,

for more intuitive interpretation and ROM

calculation, the quaternions were later converted into

Euler angles.

Euler angles were preferred for this study as they

provide a clearer, more intuitive understanding of

each foot motion relative to the lumbar region,

making it easier to interpret and compute ROM in

clinically relevant terms (Agnew, 1944). The

conversion from quaternions to roll, pitch, and yaw,

calculated according to the Equations (2), (3), and (4),

ensures mathematical precision, facilitating accurate

ROM measurement.

A quaternion is typically expressed as Equation

(1):

𝑄𝑤𝑥𝑖𝑦𝑗𝑧𝑘

(1)

Where

𝑤

is the scalar component and 𝑖, j and k

are the vector components along the x, y, and z axes,

respectively (Latimer, 1948). Therefore, the

calculation of the three Euler angles (roll, pitch, and

yaw) is possible by using quaternion Q=(w,x,y,z), as

in the Equations (2), (3), and (4). In this context, roll

is the rotation around the X-axis, pitch is the rotation

around the Y-axis and yaw is the rotation around the

Z-axis (Diebel, 2006).

𝑅𝑜𝑙𝑙  arctan













(2)

𝑃𝑖𝑡𝑐ℎ  𝑎𝑟𝑐𝑠𝑖𝑛 2𝑤𝑦  𝑧𝑥

(3)

𝑌𝑎𝑤  arctan













(4)

The ROM was calculated using a sliding window

approach, where each set of Euler angles (roll, pitch,

and yaw) was divided into overlapping segments with

a window size of 640 samples and a step size of 128

samples, resulting in a 20 percent step size between

consecutive windows. This method captures the

variability in motion over time by calculating the

ROM based on the maximum and minimum values

within these 5-second windows, offering a detailed

temporal analysis of joint motion. Key phases of gait,

such as heel strike or toe-off, which are often

impaired in MS patients, are highlighted through this

approach. For each window, the ROM was

determined as in the Equation (5), by computing the

difference between the maximum and minimum

values of the Euler angles (roll, pitch, and yaw) for

both the right and left feet, relative to the lumbar

region. This relative comparison aids in

understanding the mobility constraints imposed by

MS, and by capturing the variability in joint angles, it

provides insights into the functional limitations

experienced by MS patients.

𝑅𝑂𝑀  max



𝑗𝑜𝑖𝑛𝑡 𝑎𝑛𝑔𝑙𝑒𝑠



min



𝑗𝑜𝑖𝑛𝑡 𝑎𝑛𝑔𝑙𝑒𝑠



(5)

2.4 Statistical Analysis

Beyond visualizing ROM patterns, key time-domain

and statistical features were extracted to provide a

more detailed comparison between groups. These

features included the standard deviation, skewness,

entropy, peak to peak and time to peak of the ROM

BIOSIGNALS 2025 - 18th International Conference on Bio-inspired Systems and Signal Processing

974

values, offering insights into movement consistency,

variability, and distribution (Table 2). By analyzing

these metrics, we aimed to capture both the

magnitude and irregularities in joint motions,

enhancing the ability to detect subtle motor

impairments in MS patients and provide unique

insights into the ROM characteristics across different

groups and capture various aspects of gait dynamics.

Skewness reflects asymmetry in movement, which is

a common biomarker in MS patients, while entropy

measures the complexity and irregularity of motion

patterns. These metrics collectively provide a

comprehensive view of movement variability, motor

control, and functional capacity, which are critical for

assessing MS-related impairments (Giannakopoulos

& Pikrakis, 2014).

Table 2: ROM-extracted features.

Statistical features Time domain features

Skewness Peak to Peak

Standard deviation Entropy

Time to Peak

3 RESULTS

Figure 3 effectively visualizes comparison of all types

of calculated ROM from IMU raw data for both the

right and left feet, across the three groups which

included to the study. The sub-plots focus on the

pitch, yaw, and roll ROM measurements. This

visualization serves as a powerful tool to highlight

variations in mobility and functionality among the

two groups of MS patients and healthy control

participants.

According to the Figure 3, the ROM duration is

noticeably longer for Group B, indicating that

individuals with moderate to severe motor

impairment took more time to complete the walking

task compared to Group A and the Healthy group.

Group B also exhibits greater variability in their

ROM patterns, particularly in roll and pitch,

reflecting difficulty in maintaining consistent

movement and stability during gait. In contrast, the

Healthy group consistently demonstrates smoother,

more controlled movement patterns across all graphs,

with fewer sharp fluctuations and shorter task

durations, highlighting superior motor control and

balance. Group A shows intermediate performance,

with higher initial ROM values in some movements,

particularly in Roll and Yaw, but exhibits more

variability than the Healthy group. Despite this,

Group A maintains better control than Group B,

which shows irregular movement patterns with

sudden spikes and dips in ROM, indicating a lack of

consistent motor control. Additionally, there are

notable differences between Right and Left ROMs

Figure 3: Comparison of extracted types of ROMs for both feet across the three groups (walking time in seconds and average

ROM in degree angles).

Enhanced Assessment of Gait Dynamics in Multiple Sclerosis: A Signal Processing Approach for Extracting Range of Motion Using

Wearable IMUs

975

across groups A and B, with Group B showing greater

variability on one side, suggesting asymmetry in

movement impairment, pointing to uneven motor

function and higher irregularities in roll and pitch

ROM compared to Group A and the Healthy group.

This trend underscores the progressive impact of

motor impairments on walking dynamics and

mobility in MS patients and suggest a progressive

loss of motor control and balance in advanced MS

stages. Additionally, asymmetry between the right

and left feet is more pronounced in Group B,

reflecting uneven motor impairments

3.1 ROM-Derived Features Extraction

3.1.1 ROM Peak to Peak Value Across the

Groups

The peak-to-peak (P2P) ROM measure reflects the

difference between maximum and minimum values,

providing an indicator of movement amplitude. A

higher P2P value suggests increased variability in

movement, which may indicate either broader motion

capacity or challenges with movement control,

particularly relevant in the context of MS-related

motor impairments. The corresponding P2P values

are presented in Table 3. Group B shows consistently

higher P2P values than both Group A and the Healthy

group across all ROM dimensions and for both feet.

This pattern points to greater fluctuations in

movement within the advanced-stage MS group,

highlighting reduced control, increased variability,

and a lack of stability in rotational movements. In

contrast, Group A demonstrates intermediate P2P

values, indicating some variability but with relatively

more control than Group B, while the Healthy group

consistently exhibits the lowest P2P values, reflecting

smoother, more controlled movement patterns.

Table 3: The comparison of ROM peak to peak value across

the three groups.

Measure Healthy

Controls

Group A Group B

ht Roll 0.51 0.57 0.59

Right Pitch 0.11 0.17 0.20

Right Yaw 0.06 0.07 0.19

Left Roll 0.43 0.58 0.84

Left Pitch 0.12 0.18 0.29

Left Yaw 0.14 0.18 0.24

3.1.2 ROM Standard Deviation Value

Across the Groups

Standard Deviation (SD) quantifies the level of

variation or dispersion in ROM data, where a higher

SD value reflects greater variability in movement

patterns. In this study, examining the dispersion of

ROM values is crucial for assessing movement

consistency, which is a key indicator of mobility

impairment in MS patients. As seen in Table 4, Group

B demonstrates higher SD values across five of six

ROM types compared to Group A and the Healthy

group, highlighting increased variability and reduced

movement consistency. This heightened variability

likely reflects the challenges in balance and

coordination associated with advanced MS,

underscoring the progressive impact of the disease on

motor control. In contrast, Group A shows moderate

SD values, suggesting better control than Group B but

with more variability than the Healthy group, which

consistently displays the lowest SD values, indicative

of smooth and stable movement patterns.

Table 4: The comparison of ROM Standard Deviation value

across the three groups.

Measure Healthy

Controls

Group A Group B

ht Roll 0.08 0.07 0.11

ht Pitch 0.02 0.03 0.04

Right Yaw 0.02 0.03 0.04

Left Roll 0.07 0.09 0.17

Left Pitch 0.03 0.03 0.05

Left Yaw 0.02 0.02 0.05

3.1.3 ROM Entropy Value Across the

Groups

As the entropy measures the complexity and

unpredictability of the ROM data, higher entropy

values suggest a more complex movement pattern,

which could be indicative of compensatory strategies

in MS patients. Conversely, lower entropy may

indicate more stereotyped and less varied movements,

potentially reflecting reduced motor function. In our

study, Group B consistently shows higher entropy

across all ROM types, indicating that their movements

are less predictable and more erratic (Table 5). This

could be due to a lack of coordination and stability as

the disease progresses. Group A shows moderately

increased entropy compared to the Healthy group,

which makes sense as early-stage MS patients may

already experience some irregularities in movement,

though not as severe as those in advanced stages.

Healthy individuals have the lowest entropy, indicating

controlled and predictable movement patterns.

BIOSIGNALS 2025 - 18th International Conference on Bio-inspired Systems and Signal Processing

976

Table 5: The comparison of ROM Entropy value across the

three groups.

Measure Healthy

Controls

Group A Group B

ht Roll 1.14 1.25 3.48

ht Pitch 1.14 1.28 3.48

Right Yaw 1.14 1.25 3.48

Left Roll 1.14 1.28 3.57

Left Pitch 1.14 1.25 3.54

Left Yaw 1.14 1.29 3.54

3.1.4 ROM Skewness Value Across the

Groups

As the skewness assesses the asymmetry of the

distribution of ROM data, it can help to identify if a

particular group has a tendency toward higher or

lower ROM measurements. According to Table 6,

both Group A and Group B exhibit high negative

skewness values in roll ROM, indicating a leftward

(negative) bias in their movement distribution. This

could suggest that their rolling motions tend to be

skewed toward a limited range, potentially due to

muscular or motor control issues. Healthy individuals

have slightly less negative skewness, indicating more

balanced motion distributions. The statistical analysis

of ROM data across different groups has revealed

significant insights into the movement capabilities of

MS patients, particularly highlighting the differences

between MS patients with mild motor impairments

and patients with moderate to severe motor

impairments. The measures of central tendency,

variability, energy, complexity, and distribution

shape all underscore the functional limitations

experienced by individuals with MS. Such insights

are crucial for developing targeted rehabilitation

strategies aimed at improving mobility and quality of

life for patients suffering from this condition.

Table 6: The comparison of ROM Skewness value across

the three groups.

Measure Healthy

Controls

Group A Group B

Right Roll -1.25 -3.39 -3.64

Right Pitch -1.88 -2.43 -0.13

ht Yaw 0.71 1.02 1.18

Left Roll -1.53 -3.81 -2.54

Left Pitch 0.34 -2.70 -0.56

Left Yaw 0.24 0.38 -0.92

3.1.5 ROM Time to Peak Value Across the

Groups

The Healthy Group reaches peak ROM the fastest,

with notably low times for Right Roll, Right Yaw and

Left Roll, which aligns with normal motor control and

strength for efficient movement (Table 7). Group B

has a longer time-to-peak in most ROM measures

(Right Roll, Right Pitch, Left Roll and Left Yaw)

indicating that severe MS patients, take longer to

reach their maximum ROM. This likely reflects

greater challenges in motor control and muscular

strength.

Table 7: The comparison of ROM Time to Peak value

across the three groups in seconds (s).

Measure Healthy

Controls

Group A Group B

ht Roll 4 12 19

ht Pitch 11 9 32

ht Yaw 2 11 4

Left Roll 7 13 23

Left Pitch 5 7 5

Left Yaw 11 11 25

4 DISCUSSIONS

4.1 Interpretation of the Results from

Clinical Perspective

Overall, the extracted features clearly illustrate that

Group B, which represents individuals with severe

motor impairment, demonstrated more pronounced

and significant differences in their ROM

characteristics when compared to Group A and

Healthy group, and showed highly irregular ROM

patterns, particularly in Roll movements, where sharp

peaks and drops are evident. Group A demonstrates

more stability than Group B but less control than the

Healthy group. This observation is logical, as

advanced stages of MS are typically associated with

greater functional limitations and variability in

movement patterns, leading to a more evident

expression of motor deficits. Interestingly, while

Group A is characterized by lower motor impairment,

it consistently exhibited intermediate values across in

most extracted features such as standard deviation,

peak to peak, entropy. This indicates that even in the

early stages of MS, individuals may experience subtle

yet meaningful variations in their movement patterns

that can be overlooked in traditional clinical

assessments. Group A patients, as the early-stage

patients tended to complete the task with increased

speed and exerted additional effort, likely as an

attempt to demonstrate their ability to perform certain

movements with ease. The presence of these

intermediate values highlights the importance of

utilizing quantitative measures to capture nuances in

Enhanced Assessment of Gait Dynamics in Multiple Sclerosis: A Signal Processing Approach for Extracting Range of Motion Using

Wearable IMUs

977

motor performance, as reliance on observable

changes alone may lead to an underestimation of

impairments. Overall, our findings underscore the

utility of detailed ROM assessments in differentiating

movement patterns among various stages of MS.

From a clinical perspective, the ROM metrics

analyzed in this study provide actionable insights into

motor impairments in MS patients. The increased

variability and entropy observed in Group B may

indicate a need for targeted interventions aimed at

improving balance and coordination. Similarly, the

asymmetrical gait patterns highlighted by skewness

metrics could guide clinicians in tailoring

rehabilitation programs to address specific motor

deficits. By integrating these objective measures into

routine clinical practice, healthcare professionals can

enhance the accuracy of disease monitoring and

personalize treatment plans.

Comparing these results to earlier studies on MS

and mobility limitations, this study uniquely

quantifies ROM variability calculated from IMU raw

signals, providing a more granular view of the

asymmetrical and erratic movement patterns in

patients with MS by extracting different time domain

and statistical features. In contrast to prior research,

which primarily relied on the directly reported

spatiotemporal gait parameters provided by IMU

recording systems, our study adopted a more

sophisticated approach by independently processing

the raw IMU signals to extract ROM metrics. Rather

than depending solely on system-generated

parameters, we applied signal processing techniques,

including quaternion to Euler angle conversion, to

obtain precise, real-time ROM measurements. This

methodology allowed us to capture nuanced joint

motion patterns, providing a more granular analysis

of movement dynamics, which may reveal subtle

variations in mobility that standard IMU reported

parameters could overlook. The ROM metrics

explored here, such as peak-to-peak, entropy,

kurtosis, and skewness, provide clinicians with an

enriched understanding of MS patients’ movement

consistency and variability. For patients with

moderate to severe impairment, the ability to detect

high variability and asymmetrical movements could

support targeted physical therapy interventions,

aiming to improve balance and reduce fall risks.

Additionally, this study suggests the potential role of

IMUs and ROM analysis in developing predictive

models for motor function decline in MS, which can

assist clinicians in modifying treatment strategies.

MS disease classification primarily rely on

spatiotemporal gait parameters, such as step length

and walking speed, reported by IMU systems, which

often lack the sensitivity to detect subtle motor

impairments. However, this study introduces a novel

approach that leverages ROM-based metrics which

capture the variability, complexity, and asymmetry in

gait dynamics and by incorporating these features,

this study aims to enhance the sensitivity and

specificity of MS disease classification.

4.2 Limitations

While the study presents promising results,

limitations include the small sample size and potential

variability in sensor placement, which may influence

ROM accuracy.

5 FUTURE DIRECTIONS

Future research could expand on this work by

incorporating a larger, more diverse patient

population and exploring longitudinal ROM changes

post-rehabilitation to assess therapy efficacy. Further,

integrating machine learning techniques with ROM

metrics could enhance the predictive power of these

measures, potentially leading to automated, real-time

analysis in clinical settings.

6 CONCLUSIONS

In conclusion, our analysis of range of motion (ROM)

features and their visualizations has significantly

enhanced our understanding of the differences among

the three participant groups. Group B’s heightened

ROM variability reflects impaired motor control,

which is consistent with prior research linking

advanced MS stages to increased motor instability

and reduced gait consistency. These findings

underscore the potential for using ROM metrics as

biomarkers for tracking MS progression and tailoring

rehabilitation approaches. By identifying and

analyzing these subtle differences, clinicians can

improve diagnostic accuracy and better tailor

interventions to meet the specific needs of patients at

different stages of the disease, ultimately enhancing

their mobility and quality of life.

ACKNOWLEDGEMENTS

This study was conducted using the service and

infrastructure of Koc University Research Center for

Translational Medicine (KUTTAM).

BIOSIGNALS 2025 - 18th International Conference on Bio-inspired Systems and Signal Processing

978

REFERENCES

Agnew, R. P. (1944). Euler Transformations. American

Journal of Mathematics, 66(2), 313. https://doi.org/

10.2307/2371990

Blandeau, M., Guichard, R., Hubaut, R. & Leteneur, S.

(2023). IMU positioning affects range of motion

measurement during squat motion analysis. Journal of

Biomechanics, 153(April). https://doi.org/10.1016/

j.jbiomech.2023.111598

Coca-Tapia, M., Cuesta-Gómez, A., Molina-Rueda, F. &

Carratalá-Tejada, M. (2021). Gait pattern in people

with multiple sclerosis: A systematic review.

Diagnostics, 11(4), 1–11. https://doi.org/10.3390/diag

nostics11040584

Das, P., Chakravarty, K., Chowdhury, A., Chatterjee, D.,

Sinha, A. & Pal, A. (2018). Improving joint position

estimation of Kinect using anthropometric constraint

based adaptive Kalman filter for rehabilitation.

Biomedical Physics and Engineering Express, 4(3).

https://doi.org/10.1088/2057-1976/aaa371

Diebel, J. (2006). Representing attitude: Euler angles, unit

quaternions, and rotation vectors. Matrix, 58, 1–35.

ftp://sbai2009.ene.unb.br/Projects/GPS-

IMU/George/arquivos/Bibliografia/79.pdf

Giannakopoulos, T. & Pikrakis, A. (2014). Introduction to

Audio Analysis: A MATLAB Approach. In

Introduction to Audio Analysis: A MATLAB Approach

(pp. 1–266). https://doi.org/10.1016/C2012-0-03524-7

Grisbrook, T., Imms, C., Walmsley, C. P., Campbell, A.,

Elliott, C. & Williams, S. A. (2018). Measurement of

upper limb range of motion using wearable sensors: a

systematic review. Sports Medicine - Open.

Gülşen, Ç., Söke, F., Aydin, F., Özcan Gülşen, E., Yilmaz,

Ö., Koçer, B., Çürük, E., Demirkaya, Ş. & Yücesan, C.

(2024). Effect of task difficulty on dual-task cost during

dual-task walking in people with multiple sclerosis.

Gait and Posture, 114(March), 95–100.

https://doi.org/10.1016/ j.gaitpost.2024.09.003

Heesen, C., Böhm, J., Reich, C., Kasper, J., Goebel, M. &

Gold, S. M. (2008). Patient perception of bodily

functions in multiple sclerosis: Gait and visual function

are the most valuable. Multiple Sclerosis, 14(7), 988–

991. https://doi.org/10.1177/1352458508088916

Huang, J., Wang, H., Wu, Q., Yin, J., Zhou, H. & He, Y.

(2024). Clinical research on neurological and

psychiatric diagnosis and monitoring using wearable

devices: A literature review. Interdisciplinary

Medicine, March. https://doi.org/10.1002/inmd.202300

Keskinoğlu, C. & Aydın, A. (2021). Wearable wireless

low-cost electrogoniometer design with Kalman filter

for joint range of motion measurement and 3D

modeling of joint movements. Proceedings of the

Institution of Mechanical Engineers, Part H: Journal of

Engineering in Medicine, 235(2), 222–231.

https://doi.org/10.1177/0954411920971398

LaRocca, N. G. (2011). Impact of Walking Impairment in

Multiple Sclerosis. The Patient: Patient-Centered

Outcomes Research, 4(3), 189–201. https://doi.org/

10.2165/11591150-000000000-00000

Latimer, C. G. (1948). Quaternion algebras. Duke

Mathematical Journal, 15(2), 357–366. https://doi.org/

10.1215/S0012-7094-48-01534-8

Senesh, M. & Wolf, A. (2009). Motion estimation using

point cluster method and Kalman filter. Journal of

Biomechanical Engineering, 131(5). https://doi.org/

10.1115/1.3116153

Soucie, J. M., Wang, C., Forsyth, A., Funk, S., Denny, M.,

Roach, K. E. & Boone, D. (2011). Range of motion

measurements: Reference values and a database for

comparison studies. Haemophilia, 17(3), 500–507.

https://doi.org/10.1111/j.1365-2516.2010.02399.x

Torres-Pareja, M., Sánchez-Lastra, M. A., Iglesias, L.,

Suárez-Iglesias, D., Mendoza, N. & Ayán, C. (2019).

Exercise interventions for improving flexibility in

people with multiple sclerosis: A systematic review and

meta-analysis. Medicina (Lithuania), 55(11).

https://doi.org/10.3390/medicina55110726

Enhanced Assessment of Gait Dynamics in Multiple Sclerosis: A Signal Processing Approach for Extracting Range of Motion Using

Wearable IMUs

979