EEG-EMG correlation for parkinson’s disease

Abstract

Parkinson's disease (PD) is a neurodegenerative disease in which disease symptoms progresses with time and mostly affected after 50 years of age. This disorder leads to loss of interconnectivity between brain and muscles. Human Biosignal like Electroencephalography (EEG) and Electromyography (EMG) acts as a tool for detecting various cognitive decline and muscle conditions in PD. Work reported here emphasized on a study of the EEG from the frontal and temporal brain in PD and healthy subjects as well as their muscle activities for the flexion and extension of the wrist. Signal acquisition was done using a dual channel Biosignal acquisition system from AD Instruments. Signals were acquired for 30 minutes each for all the subjects in the Out Patient Departments of various hospitals. Signal analysis was carried out in MATLAB platform. Various EEG and EMG features were extracted to determine brain and muscle conditions of patients. Classification was done using Artificial Neural Network for EEG and EMG features separately. A combination of EEG and EMG feature was also classified using ANN which gave a highest classification rate of 98.8% as compared to only EEG and EMG feature. This result proved that correlation and combination of various EEG and EMG features together provides more accuracy to the disease classifier. It gives an insight into the mechanism of muscle weakness and gait problems caused from the low production of dopamine in the brain of PD by giving a high classification rate as compared to others(Only EEG features and Only EMG features). Many studies have reviled the effect of EEG and EMG using various tools but here we focus mainly on correlation between them in the early stages. Detection of conditions of brain and muscle in early stage helps the clinician for proper medication to control disease progression. This work can also be further applicable for classifying the various stages of PD and also classifying them from the healthy subjects.