Single-channel Selection for EEG-based Emotion Recognition Using Brain Rhythm Sequencing
The EEG-based emotion recognition can be achieved by our proposed brain rhythm sequencing method. Please find more details attached.
1. 'brs-matlab' can be used to interpret EEG as the brain rhythm sequence data. The results are saved as the 'ar'. In the 'ar', regarding the five brain rhythms, 1 is delta, 2 is theta, 3 is alpha, 4 is beta, 5 is gamma.
2. 'from ar to ac' can be used to rearrange the sequence data based on the channel locations, as our method aims to evaluate the accuracies based on the channels.
3. 'classification' can be used to achieve the sequence classification through similarity (by DTW). The reuslts of arousal classifcation are in 'acc1' and of valence classifcation are in 'acc2'.
In the 'acc1' or 'acc2', the size is 32 x 6, in which 32 corresponds to the 32 channels, the order can be found from the description of the dataset. 6 refers to the 6 time segments, each of 10 s.
So, the 'acc1' or 'acc2' include the classification accuracies of all 32 channels at 6 time segments, and the best one can be determined from it.
Following the above operations, the results of all subjects from different cases can be obtained.