A Novel Approach to Distinguish Parkinson's Disease Patients From Healthy Control Subjects Using Speech-Based Task...

Area of Science:

• Neurology
• Speech Science
• Biomedical Engineering
• Machine Learning

Background:

• Parkinson's disease (PD) frequently presents with early speech and voice impairments.
• These speech characteristics can serve as potential biomarkers for PD diagnosis.
• Speech analysis is crucial for tracking PD progression and evaluating treatment efficacy.

Purpose of the Study:

• To develop a novel method for differentiating Parkinson's disease patients from healthy controls using speech recordings.
• To explore the utility of integrating recurrence plots (RPs) with traditional speech features for enhanced diagnostic accuracy.
• To assess the proposed method's effectiveness in monitoring PD progression.

Main Methods:

• Speech task recordings from Parkinson's disease patients and healthy controls were analyzed.
• A hybrid approach combined recurrence plots (RPs) and Mel-frequency Cepstral Coefficients (MFCCs).
• Convolutional Neural Networks (CNNs) extracted features from RPs and Mel-spectrograms, followed by Support Vector Machine (SVM) classification.

Main Results:

• The proposed method demonstrated high effectiveness in distinguishing between Parkinson's disease patients and healthy controls.
• Classification accuracies exceeding 90% were achieved on the PC-GITA speech database and Greek PD tasks.
• The integration of RPs with CNNs and SVMs proved successful in enhancing speech signal representation and classification.

Conclusions:

• The novel speech analysis method shows significant promise as a biomarker for Parkinson's disease diagnosis.
• This approach offers a reliable, non-invasive tool for monitoring disease progression and therapeutic response.
• Further validation on diverse datasets is recommended to solidify its clinical applicability.