Individual Differences in Error Related Processing of Expectancy Violations

Abstract

Electrophysiological investigations of brain processing of feedback reveal that the anterior cingulate cortex is important in monitoring for behavioral errors. The feedback related negativity (FRN) is an event related potential that reflects this behavioral monitoring to allow for behavioral programs that are planned to be inhibited and more adaptive responses generated. Reinforcement learning theory proposes that the FRN is the expression of the action of the mesolimbic dopaminergic input to the cingulate, and that a participant's expectation of reward is reflected in the FRN, based on rational appraisal of objective probabilities. However, investigations into how the FRN responds to probability variation have shown it to be relatively insensitive to objective probabilities, possibly due to differences between participant subjective expectations and task relevant objective expectations. The relationships between error related processing and individual differences in personality and cortical baseline asymmetry measures were explored utilizing a simple pattern learning task and not a gambling related paradigm in order to limit gambling related fallacious reasoning and affective responses to gambling related activities. The current study was able to induce a FRN component, although this component was found to be independent of reward expectation and showed little relationship with personality and measures of cortical asymmetry. A second prominent event related potential was revealed during data analysis which represented a positivity occurring after the FRN, referred to as the P4. This component showed sensitivity to expectancy variation, with significantly different amplitudes between conditions for anterior midline electrodes (p < .01). The P4 component showed some relationships with measures of anxiety, appetitive motivation, and baseline cortical asymmetry, although these relationships varied by electrode and expectancy condition. However, multiple regression analysis revealed that the amplitude differences between the high and low expectancy conditions for the P4 component was able to account for 60% of the total variance in differences between expectancy conditions for the FZ electrode, F(3, 14) = 7.057, p = .004, [eta]² = .60, with other significant models revealed for the other anterior midline electrodes at diminished goodness of fit. This suggests that the event related potential components reflects more cognitive related processes, but that affective states and differences in affective and appetitive personality promoted individual differences in processing loss related feedback.