A Deep Learning–Driven Framework for Remote Monitoring and Severity Classification in Multiple Sclerosis Using Non-Motor Symptom Questionnaires

Document Type : Research Article

Authors

1 Computer Engineering Department, Khavaran Institute of Higher Education, Mashhad, Iran.

2 Computer Engineering Department, Khavaran Institute of Higher Education

3 Faculty of Medicine, University of Medical Sciences, Mashhad, Iran.

Abstract

Multiple sclerosis (MS) is a chronic neurodegenerative disorder with unpredictable progression. Despite advancements in imaging and clinical assessment, access to regular neurological evaluations remains limited, particularly in regions with centralized healthcare services. This study proposes a novel deep learning–based approach for assessing MS severity using non-motor symptom questionnaires. Sixteen validated patient-reported outcome measures (PROMs) were administered through a web-based platform, enabling remote monitoring of emotional, behavioral, urinary, fatigue, and functional symptoms. Data from 152 fully completed questionnaires were used to train and evaluate the proposed convolutional neural network (CNN) model. The model demonstrated superior performance compared with traditional machine learning classifiers, achieving 99.5% accuracy on the test dataset while incorporating regularization to mitigate overfitting. The proposed framework demonstrates strong potential as a decision-support tool for assisting remote monitoring and early identification of symptom deterioration in patients with multiple sclerosis. The proposed approach offers an affordable, non-invasive, and scalable solution for MS monitoring, particularly beneficial where MRI and specialist access are limited. Future work will validate the model on larger and multi-center cohorts.

Keywords

Future Research on AI and IoT
Volume 2, Issue 2
October 2026
Pages 73-85

Files

History

  • Receive Date: 30 December 2025
  • Revise Date: 13 June 2026
  • Accept Date: 27 January 2026
  • First Publish Date: 01 June 2026
  • Publish Date: 01 June 2026

Share

How to cite

Statistics