2025 American Academy of Neurology Abstract Website

Objective:

To analyze the role state-of-the-art machine learning (ML) techniques have played in understanding the cognitive and behavioral patterns in people with multiple sclerosis (pwMS) from hidden pattern identification to classification and prediction, and identify specific gaps for future ML-based cognitive research in multiple sclerosis (MS).

Background:

Multiple Sclerosis is a chronic neurodegenerative disease that is the leading cause of non-traumatic disability in young adults. Cognitive impairment and behavioral comorbidities are common disabling symptoms that show significant heterogeneity among pwMS. Machine learning is emerging as a powerful tool for exploring the role of cognition in the disease.

Design/Methods:

This was a scoping review designed per the PRISMA guidelines. PubMed and Web of Science were searched for experimental and observational studies that utilized any one of the machine learning methodologies and described cognitive impairment in multiple sclerosis. Reviews were excluded.

Results:

A total of 147 studies were screened for inclusion and 57 were included per inclusion criteria. Of these, 35% (20/57) used machine learning for the classification of cognitive impairment, 25% (14/57) described their role in optimizing MS diagnosis, 32% (18/57) identified hidden patterns using ML.

Conclusions:

While ML techniques have been implemented in several settings to enhance the understanding of cognitive impairment in MS, studies leveraging ML to elucidate the cognitive phenotypes, and their therapeutic associations are lacking. This gap represents a significant opportunity for future research, particularly in utilizing ML to improve personalized care for MS patients by incorporating cognitive outcomes into treatment decision-making.