Recognition of two emotional states of joy and sadness using phase lag index and SVM classifier

Due to the preceding studies in recent years, emotion recognition has not been done only by using local features of a single channel. Since the process of hearing and understanding it, requires the cooperation of different brain regions, it is expected that investigation of brain connectivity among channels can be an appropriate tool for emotion recognition beside the local analysis of each channel. The aim of this research was to evaluate the possibility of recognition of two music induced emotional states, joy and sadness, using functional connectivity features extracted from EEG signals. To achieve this goal, phase lag index (PLI) features of every channel pairs (6 features) and support vector machine (SVM) algorithm were used. Based on the results, among all extracted features, the most distinction between the emotional states of joy and sadness was obtained using the connectivity features of channel pair C
₃
-F
₃
and also channel pair C
₄
-F
₄
with the classification accuracies of 68.18% and 54.55%, respectively. According to the fact that these two features are related to the connectivity of two similar channels in two hemispheres, it seems that the inter-hemispheric connections are more related to the emotional states of joy and sadness rather than the intra-hemispheric connections. Furthermore, the classification precision was higher when using 6 features simultaneously (80%). This result represents the importance of brain network analysis.