A Sequential Study of Emotions through EEG using HMM

Changes in emotional state of an individual are the result of changes in brain signals. The brain signals are transduced in form of time dependent electrical signals by electroencephalogram (EEG). The method for study of EEG signals for detection of emotional state of the subject is proposed in this paper. The sequential study of changes in emotions is described using the hidden Markov model (HMM). The HMM is used on the normally distributed (ND) data to depict the series of changes in emotions as the subject is evoked through video stimulations. The data set used in this paper is the standard DEAP dataset. The standard valence-arousal plane is analyzed for detection of emotional states. HMM is employed for sequentially classifying the features extracted from the given data. HMM predictions for transition and emission probabilities are compared to resulting probabilities. The aim of this paper is to propose a method to depict the transition of emotions from state into another, hence, transition matrix of the HMM are used to measure the accuracy of the system. To determine the accuracy of the system absolute mean value of the transition matrix is obtained, and accuracy is found out to be 97.21%. Hence, a model is developed using HMM to describe the sequential change in emotional state as the EEG changes.