Effect of Smoking in EEG Pattern and Time-Frequency Domain Analysis for Smoker and Non-Smoker

Abstract

As addiction is said to be a mental health condition which can be derived from Electroencephalogram, this study focuses on changes in EEG pattern due to smoking. A methodology is proposed here to identify smoker and nonsmoker by EEG domain analysis. In this research, three ANNs were built for domain analysis to differentiate smokers from non-smokers. It was found that the neural network built with the attributes of both time and frequency domain provided best quality with MSE of $6.238 \times 10^{-08}$ than the neural networks using either time or frequency domain features. On the other hand, frequency domain based ANN solely gives better performance than time domain properties based neural network. It is concluded in this paper that values of PSD and FFT is much higher in EEG of smokers than the non-smokers.