ERP Correlates of Feedback Processing for Number of Misses in Gambling

Abstract

Feedback signals promote changes in behavior that may help to avoid negative out-
comes. Feedback evaluation primarily calculates the difference between expected out-
comes and the actual outcomes presented by feedback. Gambling tasks are used to in-
vestigate the processing of feedback evaluation. Studies have shown that event-related
brain potentials (ERPs) reflect a feedback signal (e.g. feedback negativity, FN), such
as monetary gain or loss, and its magnitude of reward. While feedback information can
deliver multiple results (e.g. hits and misses) immediately for choices with multiple se-
lections (e.g. bets on red/black and odd/even in roulette), it remains unclear whether
ERPs reflect the number of hits/misses indicated in the feedback. This ERP study
focused on evaluating the number of misses indicated by feedback for risky choices.
Electroencephalograms were measured while participants performed the task. In each
trial, two geometric stimuli (square or triangle, colored blue or red) were presented in
series with an interval between. During the first stimulus presentation, participants
bet on differences of color and shape between the presented and subsequent stimulus.
The second stimulus provided feedback on the result. Participants won for two hits
and lost the bet if they had single or double miss. ERP results showed that feedback
from double miss evoked a larger P3 on the front central site than feedback from single
miss, regardless of the amount lost. This indicates that the combination of FN and
P3 reflects processing of the number of misses in multiple selections in a single trial
notified by feedback.