Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Exploration of machine learning techniques in predicting multiple sclerosis disease course.

Yijun Zhao1, Brian C Healy2,3, Dalia Rotstein2

  • 1Department of Computer Science, Tufts University, Medford, Massachusetts, United States of America.

Plos One
|April 6, 2017
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Lesion network localization of depression in multiple sclerosis.

Nature. Mental health·2026
Same author

Current modeling approaches for drug delivery to the central nervous system.

Therapeutic delivery·2026
Same author

Multiple Sclerosis and Outpatient Physician Care in the Last Years of Life: A Population-Based Study.

Neurology·2026
Same author

Plasma phosphorylated tau 217 and longitudinal trajectories of Aβ, tau, and cognition in cognitively unimpaired older adults.

Nature communications·2026
Same author

A pilot randomized clinical trial of a peer support intervention for hematopoietic stem cell transplantation.

Blood advances·2026
Same author

Assessment of the impact of outcome definitions on power in multiple sclerosis clinical trials.

Multiple sclerosis journal - experimental, translational and clinical·2026

Machine learning, specifically Support Vector Machines (SVM), can effectively predict multiple sclerosis (MS) disease course using early clinical and MRI data. This approach aids in identifying patients needing aggressive treatment.

Area of Science:

  • Neurology
  • Medical Informatics
  • Machine Learning in Medicine

Background:

  • Predicting the disease course of multiple sclerosis (MS) is crucial for effective patient management and treatment selection.
  • Current prediction methods often lack accuracy, especially in the early stages of the disease.
  • Machine learning offers potential for improved predictive modeling in complex neurological conditions.

Purpose of the Study:

  • To evaluate the efficacy of machine learning methods, particularly Support Vector Machines (SVM), in predicting the disease course of multiple sclerosis (MS).
  • To compare the performance of SVM against logistic regression (LR) using various clinical and imaging data.
  • To identify key predictors for different disease trajectories in MS patients.

Main Methods:

Related Experiment Videos

  • Utilized data from 1693 CLIMB study patients, classifying them as worsening (EDSS increase ≥1.5) or non-worsening over five years.
  • Developed SVM and logistic regression models using demographic, clinical, and MRI data from years one and two to predict EDSS at five years.
  • Compared model performance based on sensitivity, specificity, and accuracy, exploring the impact of misclassification costs and data types.

    • Main Results:

    • Baseline data alone had limited predictive power for MS disease course.
    • SVM models incorporating one year of clinical and MRI data significantly improved prediction accuracy (sensitivity 71%, specificity 68%) compared to baseline.
    • Optimized SVM models with adjusted classification costs achieved up to 86% prediction accuracy, outperforming logistic regression.
    • Key predictors for non-worsening MS included race, family history, and brain parenchymal fraction; worsening MS was predicted by brain T2 lesion volume.

    Conclusions:

    • Support Vector Machines (SVM) show significant promise for predicting multiple sclerosis (MS) disease course.
    • Integrating short-term clinical data, brain MRI, and advanced modeling techniques (cost-sensitive learning) enhances predictive accuracy.
    • This approach can facilitate early identification of patients requiring more intensive treatment strategies for MS.