Related Experiment Video
Updated: Apr 18, 2026
14:27
Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data
Published on: June 26, 2013
16.5K
Analyzing multiple-sclerosis progression: stage-specific biomarker insights via explainable machine learning.
Selahaddin Batuhan Akben1, Ayşenur Bilirim1, Cantürk Akben2
1Department of Electrical and Electronics Engineering, Osmaniye Korkut Ata University, Osmaniye, Türkiye.
Neurodegenerative Disease Management
|April 2, 2026
Summary
This study used AI to find early signs of Multiple Sclerosis (MS) in patients with Clinically Isolated Syndrome (CIS). Key predictors include MRI lesions and oligoclonal bands, with environmental factors also playing a role.
Area of Science:
- Neurology
- Artificial Intelligence
- Biomarker Discovery
Background:
- Multiple Sclerosis (MS) is a chronic autoimmune disease.
- Early diagnosis from Clinically Isolated Syndrome (CIS) is challenging.
Purpose of the Study:
- Investigate stage-specific biomarkers for CIS-to-MS conversion.
- Utilize explainable machine learning for predictive modeling.
Main Methods:
- Prospective dataset of 273 CIS patients over 10 years.
- Stratification by Expanded Disability Status Scale (EDSS) scores (1, 2, 3).
- 10-fold cross-validation and Shapley analysis for variable importance.
Main Results:
- AI models achieved high accuracy (up to 100%) in predicting MS conversion.
- Periventricular MRI lesions and oligoclonal bands were primary predictors.
- Spinal cord lesions, motor symptoms, education, breastfeeding, and varicella history influenced risk.
Conclusions:
- AI models effectively identify stage-specific biomarkers for MS.
- MRI findings and psychosocial/environmental factors are crucial for early diagnosis and management.
Keywords:
Multiple sclerosisclinically isolated syndromedisease progressionexplainable artificial intelligencehealth psychologymachine learning
