Parkinson's disease detection and classification using EEG based on deep CNN-LSTM model.

Area of Science:

• * Neurology
• * Computational Neuroscience
• * Medical Diagnostics

Background:

• * Parkinson's disease (PD) is characterized by progressive motor function loss.
• * Electroencephalogram (EEG) signals are valuable for early brain disorder diagnosis.
• * Current methods require enhanced signal representation for improved Parkinson's disease classification.

Purpose of the Study:

• * To develop advanced deep neural network (DNN) models for Parkinson's disease diagnosis using EEG signals.
• * To improve the accuracy of classifying individuals with Parkinson's disease based on EEG data.
• * To explore hybrid architectures combining Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) for enhanced feature extraction.

Main Methods:

• * Two hybrid deep neural network (DNN) models were designed, integrating CNN and LSTM.
• * A parallel model processed structural and contextual features of EEG signals separately before combination.
• * A series model utilized CNN for structural feature extraction followed by LSTM for contextual dependency analysis.

Main Results:

• * The parallel hybrid model achieved 97.6% specificity, 97.1% sensitivity, and 98.6% accuracy.
• * The series hybrid model demonstrated superior performance with 99.1% specificity, 98.5% sensitivity, and 99.7% accuracy.
• * Both models successfully performed 3-class classification: PD patients with medication, PD patients without medication, and healthy controls.

Conclusions:

• * Hybrid DNN architectures combining CNN and LSTM offer superior performance for Parkinson's disease diagnosis via EEG.
• * The proposed models provide highly accurate and sensitive classification, aiding in early detection and management of Parkinson's disease.
• * These findings highlight the potential of advanced signal processing and machine learning in neurological disorder diagnostics.