Poster B21, Sunday, March 25, 8:00-10:00 am, Exhibit Hall C
EEG Responses to Unexpected Outcomes of Own or Partner’s Actions in a Turn-Taking Game
Gedeon Deák1, Kevin Jenson1, Alvin Li1, Scott Makeig1; 1University of California, San Diego
To investigate real-time, naturalistic social decision-making, we collected EEG within a turn-taking game. Adult dyads took turns touching one of two animated bubbles on a touchscreen table, with the goal of eliciting a “big pop” (high reward) instead of a less-interesting (i.e., low reward) outcome. Subjects had to learn feedback contingencies by trial and error. Each session consisted of a 64-trial rule-learning block followed by two unpredictable blocks (totaling 400 trials) in which outcome contingencies were reversed in 20% of the trials (random), and finally a block of 48 trials with the original contingencies. Late Positive Complex (P3) features of scalp ERPs were modulated by reward size, and by expectancy-violation. We used adaptive-mixture independent component analysis (AMICA) to unmix EEG signals, and clustered the resulting independent components (ICs) using k-means. We then analyzed two IC clusters that accounted for the most ERP variance from 200 to 700ms. ERP amplitudes of the more centro-medial-ACC cluster were larger for high-reward outcomes, whereas amplitudes of the more rostral-ACC cluster were larger following unexpected outcomes. Notably, both of these effects were significant for own-trial ERPs, but were qualitatively similar but non-significant in observation trials (i.e., watching the partner’s actions). These results suggest that (1) reward and uncertainty processing exhibit unique dynamics in distinguishable cortical networks; (2) the LPC/P3 consists of multiple spatially, temporally and functionally distinct processes (converging evidence); and (3) cortical outcome-evaluation processes are attenuated in observation relative to “first person” action, even in a cooperative, shared activity.
Topic Area: EMOTION & SOCIAL: Person perception