A multimodal mutual information‐guided feature selection framework for predicting rehabilitation response in Parkinson's disease with postural instability and gait disorder
Randomized trial identifies multimodal predictors of rehabilitation response in Parkinson's disease, suggesting potential for personalized neurorehabilitation strategies.
Key Points
The study aims to identify key multimodal features that predict rehabilitation response in Parkinson's disease with postural instability and gait disorder.
Twenty-one PD patients with postural instability and gait disorder participated in motor-cognitive rehabilitation.
A multimodal framework was used to assess features across demographics, clinical scales, gait parameters, MRI, and EEG.
Predictive stability was evaluated using five machine learning models including support vector machine and random forest.
Multimodal predictors consistently outperformed unimodal models across classifiers.
Key features identified include functional connectivity and cortical thickness derived from MRI and EEG.
The proposed framework showed strong potential for generalization in predicting rehabilitation outcomes.