Related Experiment Video
Updated: Jun 13, 2025
07:15
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
6.8K
Performance of machine learning models for predicting high-severity symptoms in multiple sclerosis.
Subhrajit Roy1, Diana Mincu2, Lev Proleev2
1Google Research, London, UK. subhrajitroy@google.com.
Scientific Reports
|May 25, 2025
Summary
This study used a mobile app to collect data from multiple sclerosis (MS) patients, enabling prediction of high-severity symptoms like fatigue and walking instability three months in advance.
Area of Science:
- Neurology
- Digital Health
- Machine Learning
Background:
- Current multiple sclerosis (MS) care relies on infrequent data collection, potentially missing subtle disease changes.
- Mobile technology offers continuous data collection for better understanding and prediction of MS progression.
Purpose of the Study:
- To develop and validate a mobile application for longitudinal data collection in MS patients.
- To utilize machine learning models for predicting high-severity MS symptoms three months in advance.
Main Methods:
- An observational study (MS Mosaic) involving a publicly launched mobile app for data collection over three years.
- Retrospective development and application of classical machine learning and deep learning models.
- Prediction of five high-severity symptoms: fatigue, sensory disturbance, walking instability, depression/anxiety, and cramps/spasms.
Main Results:
- The study successfully collected longitudinal data from MS subjects in the United States.
- Developed and tested predictive models for key MS symptoms.
- Demonstrated the potential for continuous, advance prediction of symptom incidence.
Conclusions:
- Mobile technology and machine learning can enhance the continuous monitoring and prediction of multiple sclerosis symptoms.
- The MS Mosaic study provides a foundation for proactive MS management through data-driven insights.
