Parkinsons disease diagnosis through electroencephalographic signal processing and neural network classification

Parkinson's disease (PD) is the second most prevalent neurological disorder, following Alzheimer's. Despite this, there is currently no successful treatment for PD. Therefore, early detection of Parkinson's disease is crucial for preventing its progression. To address this, a computer-aided diagnosis system has been implemented to identify any abnormalities. Significant research has been conducted using speech and gait analysis. However, there is growing interest in using electroencephalographic (EEG) signals to diagnose Parkinson's disease at an early stage. This paper aims to use EEG to capture neural correlates of dysfunction in PD patients and compare with the normal ones to determine whether a person has PD. The method is to preprocess the EEG dataset using MATLAB and EEGLAB and to analyze and classify the preprocessed data using MLP neural networks, which has good expressiveness and adaptability. Our dataset contains 25 sets of data with 11 healthy people and 14 Parkinson's Disease patients. Experiments show that the model has an average test accuracy of 96.8% and average test loss of 12.8%.