Parkinson's Disease

The detection of Parkinson's Disease (PD) using Electroencephalography (EEG) signals presents a significant challenge due to the complex nature of the EEG data. We present a detailed progress analysis of our final year research project on Parkinson's Disease identification leveraging Quantum Convolutional Neural Networks (QCNN). Employing QCNN with Gramian Angular Fields (GAF) representation, we offer a robust and generic solution. Our methodology includes an innovative approach to propose a more robust and generalized solution by capturing PD-related long-term dependencies using a larger window size. The use of Gramian Angular Fields (GAF) to represent EEG time series data in a time-frequency domain allows the tomographic inspection of EEG signals. Leveraging the power of quantum computing (QCNN) processes larger data to capture and process features from the GAF images effectively. Current results provide some insight into the PD-related patterns versus individual-specific patterns from EEGs. The results also showcase that larger window sizes are more promising for capturing the PD-related long-term dependencies from EEG signals. Using the proposed QCNN model, the classification of PD achieved 78\% accuracy.

Application: Parkinson Detect

User Guide

Datasets:

• UC San Diego Resting State EEG Data from Patients with Parkinson’s Disease

• EEG: 3-Stim Auditory Oddball and Rest in Parkinson’s 

• Parkinson’s disease: Resting state EEG 

• PD LPC Rest 

Sample Data

GitHub Code

System Usability Study for the Application

Team :  Mr. Dasun Nimantha, Mr. Dinuka Nimantha, Mr. Dinuka Wimalarathna

Research Home             Biomedical Imaging and Computational Modelling           Publications