Recognising Consciousness Problems in Brain Injuries Using EEG Coupling and Machine Learning

Abstract

The cognitive electroencephalogram has lately received a lot of attention for looking into whether EEG properties may be employed as novel predictors for recovery in mild brain damage detection. To solve this issue, this paper proposes a computer-aided technique for automatic DoC identification based on information extracted from electroencephalogram data.It adds a new connection metric called Power Spectral Density Difference, which is based on a recursive Cosine function. The following processing stages.As a result, developing an approach for painstakingly flagging and collecting clean EEG data in order to obtain high-quality discriminative features utilising PCA for feature selection is critical. The technique then uses an ensemble Machine learning approach to classify brain-injured individuals into DoCclasses. Our proposed method for implementing deep learning algorithms with excellent accuracy and prediction status.