Decoding the behavioral responses after online social evaluation using EEG-based multivariate pattern analysis

Effective processing of online social evaluation and the ability to adjust behavior accordingly are central to maintaining social bonds in the pervasive context of online interactions. An essential but unresolved question is how people dynamically respond to social evaluations during online interactions, and what neural mechanisms underlie these behavioral adjustments. Forty-three young adults (16 male; age from18-26) completed a novel social-media task in which ostensible peers provided positive feedback ("likes") or negative feedback ("dislikes") to participants' posts. Subsequently, participants rated the likeability of each peer's post, recorded the likability rating reaction time (RT), and indicated readiness to view the next evaluation (ready RT). We combined behavioral measures with single-trial EEG analyses and multivariate pattern analysis (MVPA). Participants increased likability ratings after receiving likes and decreased ratings after receiving dislikes. Additionally, lower self-reported pleasantness following dislikes was associated with a speeding of readiness to view subsequent evaluations. Time-resolved MVPA decoded trial-by-trial likability ratings, likability rating RT, and Change in Ready RT (ready RT_n+1 - RT_n) from distributed EEG patterns emerging in a late window (∼600-1000 ms), with primary contributions from frontal-central and central-parietal electrodes. Furthermore, single-trial LMMs showed that larger frontal Late Positive Potential (LPP) predicted lower liking after receiving likes, whereas smaller frontal LPP amplitudes predicted lower liking and slower readiness after receiving dislikes. These findings unpack the valence-specific adjustment patterns individuals employ in online contexts and their underlying dynamic neurocognitive mechanisms, demonstrating that late-stage cognitive control and regulatory processes are crucial for shaping subsequent social actions.