Publication

Abstract : Parkinson’s disease (PD) is a neurodegenerative disorder marked by motor dysfunction and cognitive challenges. This study focuses on analyzing the non-linear characteristics of electroencephalography (EEG) signals in PD patients by examining the effects of eyes open and eyes closed conditions during ON and OFF medication states. Using a publicly available EEG dataset from the University of New Mexico (UNM), which includes approximately 2 min of resting-state EEG recordings from 27 PD patients and 27 healthy controls, fuzzy recurrence plots are generated for converting the recorded signals to images. A range of deep learning techniques, including Convolutional Neural Network, ResNet, Inception, and Vision Transformer are applied for classification. While most models exhibited low accuracy, the Vision Transformer demonstrated promising performance, achieving better results exclusively for the eyes closed condition, but not for the eyes open condition. Classification performance was evaluated using metrics such as accuracy, precision, recall, and F1 score, complemented by a classification report and confusion matrix. This research contributes to the diagnosis and monitoring of PD based on EEG, highlighting the potential of deep learning techniques to analyze complex neurological data.