A Novel Experimental Study to Enhance the Attentional State using EEG Signals

Abstract

In this paper, we propose a simple low-complex classification framework for the cognitive enhancement with the sustained attention stimuli using Electroencephalography (EEG) signals. The visual stimuli comprise of four face images: two happy (one male and one female) and two unhappy (one male and one female). The neuronal response is decoded using a combination of discrete wavelet transform (DWT) and ensemble classifier. The features are extracted by decomposition of recorded EEG signals using Daubechies wavelet filter (db4) and used the statistical methods such as the absolute mean value, power, and standard deviation for classification. The proposed methodology is validated on in-house recorded visual attention EEG (VA-EEG) dataset using six subjects (three males, three females) and evaluated the performance on six binary combinations of facial stimuli. The performance results show that the binary combination of male happy (MH) and female happy (FH) facial stimuli aids in cognitive enhancement for the people suffering from cognitive symptoms. The proposed low-complex feature extraction classification framework obtained a mean classification accuracy (CA) and a mean kappa value of 86.58% and 0.72, respectively.