Related Experiment Video
Updated: Jul 5, 2025
The Mouse Stroke Unit Protocol with Standardized Neurological Scoring for Translational Mouse Stroke Studies
Published on: February 7, 2025
Validation of the Machine Learning-Based Stroke Impact Scale With a Cross-Cultural Sample
Shih-Chieh Lee1, Chia-Yeh Chou2, Po-Ting Chen3
1Shih-Chieh Lee, PhD, is Assistant Professor, School of Occupational Therapy, College of Medicine, National Taiwan University, Taipei, Taiwan, and Occupational Therapist, Department of Psychiatry, National Taiwan University Hospital, Taipei, Taiwan.
The machine learning-based Stroke Impact Scale (ML-SIS) offers comparable scores and reliability to the original Stroke Impact Scale-Third Edition (SIS 3.0), except for the Emotion domain. This makes ML-SIS a valuable tool for efficient stroke assessments.
Area of Science:
- Neuroscience
- Rehabilitation Medicine
- Health Informatics
Background:
- The Stroke Impact Scale-Third Edition (SIS 3.0) is a comprehensive measure of stroke recovery.
- A 28-item machine learning-based version (ML-SIS) was developed for efficiency.
- The ML-SIS requires validation in an independent sample.
Purpose of the Study:
- To cross-validate the ML-SIS against the original SIS 3.0 in an independent stroke survivor cohort.
- To assess the test-retest reliability of the ML-SIS.
- To determine the comparability of domain scores between ML-SIS and SIS 3.0.
Main Methods:
- The study involved 263 individuals with stroke from five hospitals in Taiwan.
- Comparability was assessed using R-squared, mean absolute error, and RMSE.
- Test-retest reliability was evaluated using intraclass correlation coefficients (ICC) with 144 participants completing a second assessment after 2 weeks.
Main Results:
- High R-squared values (0.87-0.95) and low error metrics indicated strong comparability for most domains.
- The Emotion domain showed lower comparability (R-squared = 0.08).
- Test-retest reliability (ICC values 0.39-0.87) was similar between ML-SIS and SIS 3.0 (0.46-0.87).
Conclusions:
- The ML-SIS demonstrates strong comparability and reliability to the SIS 3.0, making it a viable alternative for stroke assessment.
- The ML-SIS can reduce the burden on patients and healthcare professionals during routine assessments.
- The Emotion domain requires further investigation for the ML-SIS's application.
