Dataset DOI: https://doi.org/10.5061/dryad.x69p8czzx

Description of the Data and File Structure

1. Study Description

This dataset was generated from an outcome assessor-blinded randomized controlled trial investigating the effects of exercise training and immersive virtual reality (VR) gaming, compared with a wait-list control condition, on gait performance, cognitive function, functional outcomes, and circulating biomarkers in individuals with relapsing-remitting multiple sclerosis (RRMS).

Participants were randomly assigned to one of three groups: exercise intervention, VR gaming intervention, or wait-list control. Outcomes were assessed at baseline (pretest) and immediately following the 8-week intervention period (post-intervention).

The data were analyzed using conventional statistical methods. Selected baseline variables were also used in exploratory supervised machine-learning analyses to investigate potential predictors of individual treatment response.

2. Study Design

Study type: Randomized controlled trial
Blinding: Outcome assessor blinded
Intervention duration: 8 weeks
Assessment points: Baseline (pretest) and post-intervention

Groups:

• Exercise group: supervised aerobic and resistance training
• Virtual reality group: immersive motor-cognitive VR gaming
• Control group: wait-list control maintaining usual activity

Total participants: 48

3. Participants

Diagnosis: Relapsing-remitting multiple sclerosis according to the 2010 McDonald criteria
Age range: 18–55 years
Disability level: Expanded Disability Status Scale (EDSS) ≤ 5.0
Disease-modifying therapy: Stable for ≥ 6 months

The released data are de-identified. Participant identifiers in the dataset are anonymized and do not contain direct personal identifiers.

4. File Contents

The dataset is provided in tabular CSV format and includes:

• Participant demographic characteristics
• Clinical disability measures
• Cognitive test scores
• Gait and functional mobility outcomes
• Balance and flexibility measures
• Serum biomarker concentrations
• Baseline and post-intervention values
• Calculated change scores

Unless otherwise specified, change scores are calculated as:

Post-intervention value − baseline value

5. Definition of Treatment Responders

Responder definitions were specified for the exploratory supervised machine-learning analyses.

Gait responder (10MWT):

A participant was classified as a gait responder if 10-meter walk test time improved by at least 20%:

(Baseline − Post-intervention) / Baseline × 100 ≥ 20%

Because lower 10MWT times indicate faster gait, this formulation expresses improvement as a positive percentage.

Cognitive responder (SDMT):

A participant was classified as a cognitive responder if the Symbol Digit Modalities Test (SDMT) score increased by at least 4 points from baseline to post-intervention.

These thresholds were selected based on clinically meaningful change criteria described in the associated manuscript. The machine-learning analyses should be interpreted as exploratory because of the modest sample size.

6. Data Processing Notes

• Data are provided in non-imputed form.
• No normalization or scaling has been applied to the deposited participant-level raw variables.
• Change-score variables are calculated as post-intervention minus baseline unless otherwise noted.
• The deposited data may be used for independent reanalysis, secondary analysis, and methodological replication subject to the CC0 license.

7. Ethical Approval and Informed Consent

The study was conducted in accordance with the Declaration of Helsinki and approved by the Royan Institute Ethics Committee.

Ethics approval ID: IR.ACECR.ROYAN.REC.1396.98

Written informed consent was obtained from all participants before study participation.

8. Data License

This dataset is released under the CC0 1.0 Universal Public Domain Dedication. The data may therefore be copied, modified, distributed, and reused in accordance with the terms of the CC0 dedication.

Files and Variables

Files

• MSRCT_Data_Set__anonymized.csv
• MSRCT_ALL_Mean_SD__anonymized.csv
• MSRCT_ALL_Class__anonymized.csv
• DisciplineSpecificMetadata.json

The CSV files contain de-identified data generated from the randomized controlled trial comparing exercise training, immersive VR gaming, and a wait-list control condition in individuals with RRMS.

Data were collected at two principal assessment time points:

• Baseline (pretest)
• Post-intervention (after 8 weeks)

The files include demographic characteristics, clinical disability measures, blood-based biomarkers, cognitive outcomes, and physical performance measures.

Identifiers and Grouping

• ID – Unique anonymized participant identifier
• group name – Intervention assignment (exercise, VR, or control)
• gender – Biological sex as coded in the deposited dataset (m, f)

Demographic and Clinical Characteristics

• age – Age in years
• EDSS (pretest) – Expanded Disability Status Scale score at baseline
• EDSS (post) – EDSS score after the intervention period
• EDSS diff – Change in EDSS score (post − pre)

Serum Biomarkers

• IGF-1 (ng/ml) pretest – Baseline insulin-like growth factor-1 concentration
• IGF-1 (ng/ml) post – Post-intervention IGF-1 concentration
• IGF-1 diff – Change in IGF-1 concentration (post − pre)
• BDNF (pg/ml) pretest – Baseline brain-derived neurotrophic factor concentration
• BDNF (pg/ml) post – Post-intervention BDNF concentration
• BDNF diff – Change in BDNF concentration (post − pre)
• NFL (pg/ml) pretest – Baseline serum neurofilament light chain (NfL) concentration
• NFL (pg/ml) post – Post-intervention serum NfL concentration
• NFL diff – Change in serum NfL concentration (post − pre)

Note: NFL is retained above where it reflects the deposited column name; the biomarker is referred to scientifically as neurofilament light chain (NfL).

Cognitive Outcomes

• ICA score pretest – Baseline Integrated Cognitive Assessment (CognICA) composite score
• ICA score post – Post-intervention ICA score
• ICA diff – Change in ICA score (post − pre)
• SDMT pretest – Baseline Symbol Digit Modalities Test score (number correct)
• SDMT post – Post-intervention SDMT score
• SDMT diff – Change in SDMT score (post − pre)

Functional and Mobility Outcomes

• Timed get up & go test (s) pretest – Baseline Timed Up and Go (TUG) test time in seconds
• Timed get up & go test (s) post – Post-intervention TUG time
• Timed get up & go test diff – Change in TUG time (post − pre)
• Three minutes step test pretest – Baseline number of steps completed
• Three minutes step test post – Post-intervention number of steps completed
• Three minutes step test diff – Change in step-test performance (post − pre)
• 10-metre timed walk test (s) pretest – Baseline 10-meter walk test (10MWT) time in seconds
• 10-metre timed walk test (s) post – Post-intervention 10MWT time
• 10-metre timed walk test diff – Change in 10MWT time (post − pre)
• Standing balance test (s) pretest – Baseline standing-balance duration in seconds
• Standing balance test (s) post – Post-intervention standing-balance duration
• Standing balance test diff – Change in standing-balance duration (post − pre)
• The sit & reach-A test (cm) pretest – Baseline sit-and-reach flexibility score in centimeters
• The sit & reach-A test (cm) post – Post-intervention sit-and-reach score
• The sit & reach-A test diff – Change in flexibility score (post − pre)

MSRCT_ALL_Class__anonymized.csv

This derived file was used in the exploratory machine-learning component of the study. It contains variables used in classification analyses of treatment response.

Variables include:

• class – Classification variable used in the derived machine-learning dataset. The exact coding should be interpreted according to the deposited data-generation workflow and associated manuscript.
• ICA score – Integrated Cognitive Assessment (CognICA) composite score
• SDMT – Symbol Digit Modalities Test score
• Timed get up & go test (s) – Timed Up and Go result in seconds
• Three minutes step test – Number of steps completed during the 3-minute step test
• 10-metre timed walk test (s) – Time required to walk 10 meters, in seconds
• Standing balance test (s) – Static standing-balance duration, in seconds
• The sit & reach-A test (cm) – Sit-and-reach flexibility score, in centimeters

DisciplineSpecificMetadata.json

This file contains discipline-specific metadata automatically generated during the Dryad submission process. It supports repository indexing and interoperability and is not required for analysis of the participant-level dataset.

Missing Values

No missing values are present in the final deposited dataset.

Code and Software

The deposited CSV files can be opened with standard tabular-data software, including Microsoft Excel, LibreOffice Calc, Google Sheets, R, or Python. No proprietary or custom software is required to access the raw data.

Statistical Analysis

Conventional statistical analyses reported in the associated manuscript were conducted using SAS version 9.4.

Exploratory Machine-Learning Analysis

Machine-learning analyses were conducted using Python 3.11, including:

• scikit-learn 1.3.0 – Logistic Regression, Random Forest, cross-validation, and model evaluation
• XGBoost 1.7.6 – Extreme Gradient Boosting classification
• SHAP – Model interpretability and feature-attribution analyses
• NumPy and Pandas – Numerical and tabular data processing

The exploratory machine-learning workflow included:

1. Preparing de-identified baseline predictor variables
2. Defining responder and non-responder outcomes according to the prespecified thresholds described above
3. Training Random Forest, XGBoost, and logistic-regression classifiers
4. Performing stratified model validation and hyperparameter tuning
5. Evaluating classification performance
6. Exploring feature contributions using SHAP values

Detailed machine-learning hyperparameters and methodological information are reported in the associated manuscript and Supplementary Methods.

Access Information

The dataset was generated by the study authors and was not derived from external or third-party datasets.

The de-identified participant-level data and accompanying documentation are publicly available through the Dryad Digital Repository:

https://doi.org/10.5061/dryad.x69p8czzx

The dataset is provided to support transparency, independent reanalysis, and reproducibility of the associated research.

All deposited data are released under the CC0 Public Domain Dedication.

Human subjects data

All data included in this dataset were collected from human participants following approval by the relevant institutional ethics committee.

Written informed consent was obtained from all participants, including explicit consent for the use and publication of de-identified research data in the public domain.

The dataset has been fully anonymized prior to deposition. All direct identifiers (such as names, contact information, national identification numbers, and exact dates of birth) were removed. Participants are represented only by randomly assigned study IDs.

The dataset contains only non-identifiable demographic variables (e.g., age in years, sex), clinical measures, biochemical markers, and functional test outcomes.

No personally identifiable information (PII) is included in this dataset, and the risk of re-identification is minimal.