Decoding and Predicting Implicit Agreeing/Disagreeing IntentionBased on Electroencephalography (EEG)

Suh-Yeon Dong, Bo-Kyeong Kim, Soo-Young Lee

Abstract: A new experiment design is proposed to understand human implicit intention by using electroencephalography (EEG). EEG data is recorded using 32-channel electrodes while seeing various sentences which contain self-relevant contents. Subjects are asked to make a decision of agreement or disagreement just after sentence ending is shown. Based on their answer, support vector machine is used for pattern classification with radial basis function kernel. The classification result shows the intention to the sentences can be classified with 67.89% of maximum average accuracy. The spatial relationship of average classification accuracy shows right frontal areas have relatively high classification accuracy. Our findings indicate that covert representation of agreement or disagreement intention can be found in the EEG band power and it is also possible to predict subjects implicit intention even before making explicit expression.