Voice-Based Early Diagnosis of Parkinson's Disease Using Spectrogram Features and AI Models.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Computational Linguistics

Background:

• Parkinson's disease (PD) is a neurodegenerative disorder impacting motor functions, notably speech.
• Voice analysis presents a non-invasive, efficient, and economical method for PD diagnosis and tracking.
• Acoustic features of speech are altered in individuals with PD, offering potential biomarkers.

Purpose of the Study:

• To develop and evaluate an automated system for distinguishing between individuals with and without Parkinson's disease using speech signals.
• To compare the performance of various machine learning (ML) and deep learning (DL) models in classifying PD based on vocal features.
• To investigate the utility of acoustic features for early detection and monitoring of Parkinson's disease.

Main Methods:

• Utilized a voice dataset of 81 samples from PD patients and non-PD individuals for training and evaluation.
• Extracted acoustic features including Mel-frequency cepstral coefficients (MFCCs), spectrograms, Mel spectrograms, and waveform representations.
• Trained and assessed ML models (SVM, XGBoost, logistic regression) and DL models (DNN, CNN-LSTM, CNN-GRU, BiLSTM).

Main Results:

• Deep learning models significantly outperformed traditional ML models in classifying PD speech.
• The Bidirectional Long Short-Term Memory (BiLSTM) model achieved the highest accuracy of 97% and an Area Under the Curve (AUC) of 0.95.
• Comprehensive feature extraction enabled robust classification, highlighting the effectiveness of the automated system.

Conclusions:

• The developed automated system demonstrates high efficacy in distinguishing PD from non-PD speech.
• Deep learning approaches, especially BiLSTM, show significant potential for accurate and early diagnosis of Parkinson's disease.
• Integrating acoustic feature analysis with DL methods offers a promising avenue for the continuous monitoring and management of Parkinson's disease.