EEG signals for emotion recognition

Abstract

This paper proposes an emotion recognition system using the electroencephalographic (EEG) signals. Both time domain and frequency domain approaches for feature extraction were evaluated using neural network (NN) and fuzzy neural network (FNN) as classifiers. Data was collected using psychological stimulation experiments. Three basic emotions namely; Angry, Happy, and Sad were selected for recognition with relax as an emotionless state. Both the time domain (based on statistical method) and frequency domain (based on MFCC) approaches shows potential to be used for emotion recognition using the EEG signals.

Evaluation of Features in Detection of Dislike Responses to Audio–Visual Stimuli from EEG Signals

“…A mean classification accuracy of up to 98.6% is reported for the best performing classifier (kNN) and feature set (DWT-db4). The results reported in Section 3 are in good agreement with the accuracy reported in previous related work [29,[31][32][33][34][35] on the DEAP dataset, summarized in Section 1. Our experimental evaluation has shown an average classification accuracy of 1-3% higher than the results reported in previous related work for the best performing classifiers (kNN and SVM).…”

“…Directly using the wavelet coefficients for classification was also evaluated in [32]. When compared to direct use of DFT and DWT coefficients, LFCC and Mel-Frequency cepstral coefficients (MFCC) [33][34][35][36][37] provide a more compact representation of the energy in the frequency bands of a signal. Cepstral coefficients were computed from the spectrum of EEG signals with overlapping [33] or without overlapping [34] among subsequent frames.…”

Dynamic screening of autistic children in various mental states using pattern of connectivity between brain regions

“…The highlighted changes in resting states including cortical under-connectivity in autism [22,78]; or more specifically, the fronto-parietal and temporo-parietal functional under-connectivity in autism [30] have been reported in previous neuroimaging studies as well. Studies have demonstrated that patterns of functional connectivity are task dependent and can be recognized for emotional states recognition as well as for identification of long-term mental stress level [75,[79][80][81][82][83]. Therefore, the contrast pattern between autistic children and age matched healthy controls during the perception of emotional faces (calm, happy and sad) are also presented in Figs.…”

Two dimensional affective state distribution of the brain under emotion stimuli

“…With significant enhancements in neuroimaging techniques, in the recent studies, brain signals are also used as the inputs for measuring valence and arousal values through machine learning methods [12] [13].…”