Parkinson's Disease Diagnosis Using Laplacian Score, Gaussian Process Regression and Self-Organizing Maps

Area of Science:

• Neuroscience
• Computer Science
• Biomedical Engineering

Background:

• Parkinson's disease (PD) is a neurodegenerative disorder impacting motor function.
• Machine learning models are crucial for tracking PD progression via Unified Parkinson's Disease Rating Scale (UPDRS) score prediction.

Purpose of the Study:

• To develop an advanced method for PD diagnosis and UPDRS score prediction.
• To leverage supervised and unsupervised learning for enhanced PD assessment.

Main Methods:

• Utilized Laplacian score for feature selection.
• Employed Gaussian Process Regression (GPR) for UPDRS score prediction.
• Integrated Self-Organizing Maps (SOM) for large dataset segmentation.

Main Results:

• The proposed method demonstrated high accuracy in predicting UPDRS scores from speech signals (dysphonia measures).
• SOM combined with Laplacian score and GPR (exponential kernel) yielded superior R-squared (Motor-UPDRS = 0.9489; Total-UPDRS = 0.9516) and RMSE (Motor-UPDRS = 0.5144; Total-UPDRS = 0.5105) values.

Conclusions:

• The developed approach is effective for PD diagnosis and progression monitoring.
• Speech signal analysis using this machine learning model shows significant potential for clinical application.