EEG Signals Acquisition, Analysis and Modeling for Classification in Healthcare

Abstract

Acquisition of electroencephalographic (EEG) from the brain using Non-invasive methods for the analyzing these signals includes visual inspection of the epoch of signals, feature extraction from these continuous recordings, and generation of some early prediction or classification. This paper studied Acquisition, Analysis of EEG Signals which can used for classification in normal and abnormal category for making a robust model used in Healthcare. The primary goal of this paper is to study various analysis and evaluation of the performance of the state of the art techniques and methods. Next goal is to try for modelling and some improvement in result of various methods which were used in our study. Biosignals were collected from Biosense BCI Bluetooth Device (single channel) and prerecorded dataset from Temple Hospital University for preprocessing and analysis. Initial part of paper carried out some improvement in the feature selection using EEGLab which is useful for the classification by machine learning technique and other part explained the preprocessing of EEG data with comparative analysis of various artifact correction and rejection algorithms. Overall results achieved by this paper shows that high accuracies can be gained with the help of long delays and using alternation in few output which limits EEG recorded information extraction from the EEG sensor used in the device and depends on its usability.