Predicting participants’ attitudes from patterns of event-related potentials during the reading of morally relevant statements – An MVPA investigation

Morality and language are hardly separable, given that morality-related aspects such as knowledge, emotions, or experiences are connected with language on different levels. One question that arises is: How rapidly do neural processes set in when processing statements that reflect moral value containing information? In the current study, participants read sentences about morally relevant statements (e.g., ‘Wars are acceptable’) and expressed their (dis)agreement with the statements while their electroencephalogram (EEG) was recorded. Multivariate pattern classification (MVPA) was used during language processing to predict the individual's response. Our results show that (1) the response (‘yes’ vs. ‘no’) could be predicted from 180 ms following the decision-relevant word (here acceptable), and (2) the attitude (pro vs. contra the topic) could be predicted from 170 ms following the topic word (here wars). We suggest that the successful MVPA classification is due to different brain activity patterns evoked by differences in activated mental representations (e.g. valence, arousal, etc.) depending on whether the attitude towards the topic is positive or negative and whether it is in accordance with the presented decisive word or not.