A study presented at the 2023 International Congress of Parkinson’s Disease and Movement Disorders demonstrated the potential of a machine learning (ML) approach to predict gait dysfunction in people with Parkinson disease (PD). The study specifically determined that the multilayer perceptron (MLP) ML approach performed best in analyzing diffusion tensor imaging (DTI) of multi-spectral diffusion weighted imaging (DWI) to predict gait dysfunction in people with PD based on white matter.

Researchers developed a prediction model based on the Movement Disorder Society-Unified Parkinson’s Disease Rating Scale III (UPDRS III) using 7 quantitative DTI measures in a population of 43 PD participants. The study assessed the comparative effectiveness of multiple ML approaches when applied to 60 template regions of interest (ROIs). MLP with 5 hidden layers and rectified linear unit (ReLU) activation function yielded an area under the curve (AUC) of 0.78, demonstrating better performance in binary classification of gait dysfunction vs the linear discriminant analysis (LDA), random forest, gradient boosting machiens (GBM), and long short-term memory (LSTM) approaches.

“This study suggests that a machine learning approach of DTI analysis may have potential in predicting gait dysfunction in PD patients if confirmed by larger confirmative studies,” said Klaus Seppi, Director of the Department of Neurology at the Hospital Kufstein. “Moreover, future studies have to explore if this pattern changes with disease progression.”