A NOVEL APPROACH FOR THE EARLY DETECTION OF PARKINSON’S DISEASE USING EEG SIGNAL

Electroencephalography is a test that applies electrophysiological monitoring methods to detect and evaluate the electrical activity of the brain. A Neuronal Dysfunction / Neuronal Death results in behavioral aberration which directly leads to abnormality in the electrical activities of the brain. The objective of this paper is to extract features and analyze the EEG Signal to detect Parkinson's disease at an early stage. This study takes into account several markers subjected to the early detection of PD that includes Welch's power spectral density, Hjorth parameters, Hurst Exponent, Information factors, and other parameters. These computed features quantify the coordination complexity of a windowed EEG data with respect to a person and are modeled to an Artificial Neural Network based classifier with ReLU and Sigmoidal Activation Function to build a classifier with 93.3% train and 88.17% test accuracy.