An Experimental Investigation on the Classification Accuracy of Visual, Audio and Cognitive Stimuli-based EEG Signals for BCI Application

The term Evoked Potential (EP) describes the reaction that the brain produces in response to external stimuli. The brain's electrical potentials are measured to determine the strength of the generated response. This study is to analyze the performance of visual, auditory, and cognitive evoked potential using an ElectroEncephaloGram (EEG). EEG is used to record electrical activity in the brain. Evoked potentials provide information on the efficiency of the sensory systems throughout specific periods. There are various types of stimuli, but the study employed, visual, auditory, and cognitive evoked potential. The electrical signal generated at the visual cortex in response to visual stimulation is measured by the Visual Evoked Potential/Response (VEP/VER). AEPs are electro-physiological reactions that the auditory system produces in response to sound stimuli. Cognitive Evoked Potential (CEP), responds to a variety of fun activities that stimulate thinking, concentration, and memory in the nerve of the brain. Machine Learning models were used after the feature extraction for classification. The outcomes demonstrated that the visual evoked potential generates classification accuracy of 96.5% whereas auditory and cognitive produces 93.8% and 95.2% respectively. Accuracy, Precision, Recall, F1 score, and Co-hen's kappa are the performance metrics used in this study.