Sensor-Derived Measures of Motor and Cognitive Functions in People With Multiple Sclerosis Using Unsupervised...

Area of Science:

• Neurology
• Digital Health
• Biomedical Engineering

Background:

• Smartphones and wearables offer objective, frequent, and remote assessment of cognitive and motor function in neurological disorders.
• Selecting appropriate sensor-derived measures (SDMs) from abundant data is a significant challenge in remote monitoring.

Purpose of the Study:

• To develop and apply a framework for selecting robust SDMs for assessing cognitive and motor function in people with multiple sclerosis (MS).
• The framework includes automated data quality checks and statistical property evaluation.

Main Methods:

• Applied a selection framework to data from 85 people with MS and 68 healthy controls using smartphone-based cognitive, manual dexterity, and mobility tests.
• Extracted 47 SDMs, screened them for bias and normality, and evaluated reliability (intraclass correlation coefficient, minimal detectable change).
• Assessed convergence of selected SDMs with in-clinic measures and patient-reported outcomes.

Main Results:

• 16 of 47 SDMs (34%) met the selection criteria, demonstrating moderate-to-good reliability in remote settings.
• Selected SDMs showed varying correlations with in-clinic tests and Expanded Disability Status Scale (EDSS) scores, with mobility SDMs showing stronger correlations.
• Correlations were consistent between in-clinic and remote smartphone assessments.

Conclusions:

• Smartphone-based assessments yield high-quality SDMs suitable for remote monitoring of cognitive and motor function in MS.
• The SDM selection framework enhances interpretability and standardization, supporting future use in interventional trials.