Abstract:
Electroencephalogram (EEG) functional connectivity microstates represent quasi-stable global neuronal activity and are considered as the building blocks of brain dynamics. Therefore, the microstate sequence analysis is a promising method for understanding the brain dynamics behind various emotional states. Recent studies have shown that EEG microstates sequence is non-Markov and non-stationary, which also reflects the importance of temporal dynamics between different emotional states. However, the characteristics based on microstate probability statistics cannot well represent the dynamic change of EEG signals. These inspire us to use recurrence analysis to model time series of microstates and capture non-obvious correlations in time series. To this end, we propose an emotion decoding model based on recurrence analysis of EEG functional connectivity microstate sequences. Firstly, the microstate of each frame signal is established by using the time domain correlation between channels, and the typical microstate patterns are obtained through clustering. Then, the original EEG signals are mapped onto microstate time series according to typical microstate patterns, and the time series are analyzed recursively to construct recursive charts for characterizing the EEG dynamic characteristics. Furthermore, convolutional neural networks (CNNs) are used to predict the regression of emotions based on the valence or arousal value. Finally, it is shown via experimental results on DEAP dataset that the regression effect of the mean square error (MSE) of the model in two dimensions of valence and arousal is 3.45 and 2.79, respectively, which is better than the MSE (3.87 and 3.25) of the traditional statistical method.