Language-based social feedback processing with randomized "senders": An ERP study.

Recently, several event-related potential (ERP) studies investigated the impact of sender attributions on language-based social feedback processing. Results showed very early responses to the social context, while interactions or effects of emotional content started later. However, in these studies, sender attribution was varied across blocks, possibly inducing unspecific, anticipatory effects. Here, who was giving feedback was disclosed simultaneously with the decision itself. Participants' ERPs differentiated between attributed senders starting with the early posterior negativity. P3 and late positive potential (LPP) amplitudes were also enlarged for the "human sender". Emotion effects occurred in the P3 and LPP time windows. Further, we found an interaction on the P3: "Human" emotional feedback was selectively amplified. Source analysis localized enhanced processing of "human"-generated feedback in visual areas from around 300 ms after feedback onset and from 400 ms also in temporal regions. Enhancement of "human" emotional feedback resulted from increased activations in the left visual word form area. These findings highlight that decoding who is giving feedback precedes content processing, both in blocked and in randomly alternating situations. Further, in quasi-realistic social contexts, processing of emotional content is selectively amplified. Finally, involvement of semantic language processing structures indicates reintegration of words in a salient context.