کنفرانس مهندسی برق ایران

صفحه اصلی / سی و سومین کنفرانس بین المللی مهندسی برق

Parkinson’s Disease Classification Using Continuous Wavelet Transform and Ensemble Convolutional Neural Networks on EEG Signals

نویسندگان :

Seyed Pedram Monazami¹ Raheleh Davoodi²

1- دانشگاه شهیدبهشتی تهران 2- دانشگاه شهیدبهشتی تهران

کلمات کلیدی :

Parkinson’s Disease Detection،EEG،Wavelet Transform،Convolutional Neural Networks،Ensemble Learning،Time-Frequency Analysis.

چکیده :

Parkinson’s Disease is a neurodegenerative disorder that impairs motor and cognitive functions, with early detection being critical for effective treatment. Electroencephalography provides a non-invasive method to capture neural activity, revealing task-related abnormalities in PD patients. In this study, we propose a novel framework that combines Continuous Wavelet Transform (CWT) with deep learning to classify PD and healthy controls using EEG signals. CWT was employed to generate high-resolution time-frequency images from EEG epochs across different task categories, recorded during interval timing tasks with short and long durations. Separate Convolutional Neural Networks were designed and trained for each task-specific image type to autonomously extract discriminative features. The outputs of the individual CNNs were combined using an ensemble strategy and integrated into a fully connected network for the final classification. The proposed framework was evaluated on EEG data from 89 PD patients and 41 healthy controls, achieving high accuracy, precision, recall, and area under the ROC curve (AUC) across multiple task categories. The ensemble approach demonstrated superior performance by leveraging complementary features extracted from different task types and durations. The results highlight the potential of time-frequency representations and ensemble deep learning techniques in EEG-based PD detection. This study provides a robust, non-invasive diagnostic tool that integrates temporal and spectral neural dynamics, paving the way for improved early detection and monitoring of Parkinson’s Disease.