NeuroAid: Emotion-Based EEG Analysis for Parkinson's Disease Identification

Parkinson's disease (PD) is a neurodegenerative condition characterized by intricate behavior and neuronal function changes. The intricacy of these changes makes it difficult to identify PD in its early stages. Experts frequently use manual evaluations of patients' movements, including drawing, writing, walking, tremors, facial expressions, and speech, although this process is laborious and prone to mistakes. A more promising avenue is the utilization of electroencephalogram (EEG) readings, which provide insights into changes in brain activity. Nevertheless, it is important to note that analyzing EEG signals is challenging due to their complexity, nonstationarity, and nonlinearity. In order to overcome these challenges and gain deeper insights into PD and its associated emotions, this letter aims to leverage deep neural networks (DNNs) to extract emotional data from these intricate EEG signals. With this motivation, this letter designs a novel DNN model for PD detection. Moreover, we have conducted experiments and compared the accuracy with several state-of-the-art machine learning and deep learning methods. The performance validation of the DNN model on the benchmark EEG brainwave feeling emotions dataset pointed out the effectiveness of the proposed DNN model with a maximum accuracy of 98.43%.