Recognition of Emotion Provoked by Auditory Stimuli using EEG Signal Based on Deep Neural Networks

Abstract

Excitements are important for the proper interpretation of actions as well as relationships among individuals. Recognizing emotions through Electroencephalogram (EEG) allows recognition of emotional states without traditional methods including filling in the questionnaire. The automatic emotion recognition reflects the excitement of the individual without clinical examinations or need to visits, which plays a very important role in completing the Brain-Computer Interface (BCI) puzzle. One of the major challenges in this regard is first to select and extract the proper characteristics/features of the EEG signal in order to create an acceptable distinction between different emotional states. The process of finding the desirable feature is generally time consuming. This study presents a new approach for the automatic identification of 3-states of emotion (positive, negative and neutral) based on the auditory stimulation of EEG signals. In the proposed method, the raw EEG signal is directly applied to convolutional neural network-long short time memory (CNN-LSTM) network, without involving the extraction/selection feature. This has been a challenging process in previous literature. The proposed network architecture includes 10 convolutional layers with 3 LSTM layers followed by 2 fully connected layers. The simulation results of the proposed algorithm for classifying 2-stages (negative and positive) and 3-stages (negative, neutral and positive) of emotion for 12 active channels show the accuracy of 97.42% and 95.23% and Cohen’s Kappa coefficient of 0.96 and 0.93 respectively.

Keywords

• eeg
• emotion
• auditory stimuli
• convulsion neural network (cnn)