Factors affecting the amplitude of the feedback-related negativity during the balloon analogue risk task

Abstract

When making decisions, the probability and magnitude of errors can play a major role in changing preferences. Electroencephalography (EEG) research examining the error-related negativity (ERN) and the associated feedback-related negativity (FRN) has indicated that the amplitude of each component may predict subsequent behavioral change. The current study used a version of the Balloon Analogue Risk Task (BART) that involves outcomes that are dynamically changing over time. As the balloon grows, more points are available but the probability of the balloon popping (netting zero points) is higher; the participant decides when to stop the balloon’s expansion to maximize points. The BART was adapted to facilitate the study of the FRN in dynamic environments. The purpose of Experiment 1 was to determine the effect of error magnitude on FRN amplitude during popped (incorrect) trials, whereas Experiment 2 was aimed at determining the effect of error magnitude on FRN amplitude during cashed-in (correct) trials. It was hypothesized that larger errors (i.e., the balloon popping after waiting a long time to cash-in) would result in a larger FRN than smaller errors. In Experiment 1, error magnitude did not contribute to the amplitude of the FRN. In Experiment 2, the masked points possible condition was a replication of Experiment 1. In the unmasked points possible condition, the number of points that could have been earned for each balloon was presented before participants found out how many points were earned. It was expected that there would be a larger FRN magnitude after cashed-in trials in the unmasked points possible condition compared to the masked points possible condition based on the magnitude of the error. In Experiment 2, the amplitude of the FRN was affected by the magnitude of the error on cashed-in trials in the unmasked condition, but not the masked condition. These results are seemingly at odds, and cannot be assimilated into any currently extant model of the FRN. An explanation relying on the motivational importance of errors is discussed.