The effect of working memory load on feedback processing: Evidence from an event-related potentials (ERP) study

Abstract

<p id=C2>Feedback processing plays an important role in behavior modification and knowledge acquisition. Previous research has explored the neurophysiological basis and psychological functions of feedback processing and proposed corresponding theoretical models, but little is known about how working memory (WM) load affects feedback processing. Studies have reported electrophysiological indicators, such as the reward positivity (RewP) and the related theta and delta oscillations, the P3 and the late positive potential (LPP), during brain processing feedback. This study will further examine how WM load modulates these electrophysiological components and their corresponding cognitive functions. <break/>In the present study, we used a dual-task paradigm to investigate feedback processing under different WM load conditions. This study included 25 healthy college students and used a 3 (WM load: baseline vs. low WM load vs. high WM load) by 2 (feedback valence: positive vs. negative) within-participant factorial design. During the experiment, participants were asked to perform a simple gambling task and a spatial memory task simultaneously, and the magnitude of the WM load included three conditions: baseline, low WM load and high WM load. The RewP generated in the early stage of feedback processing and the LPP generated in the late stage of feedback processing, as well as the delta and theta oscillations related to feedback evaluation, were analyzed. <break/>The behavioral results showed that the accuracy of the low WM load condition was significantly higher than that of the high WM load condition. The electrophysiological results showed that the amplitudes of the RewP were sensitive to feedback valence, with positive feedback evoking larger RewP than negative feedback, but the RewP was not affected by the WM load. There was no difference in the P3 amplitude under the different WM load conditions. For the LPP, there was a significant interaction between the WM load and feedback valence. Further analysis revealed that, in the high WM load condition, the LPP amplitude was larger for positive feedback than for negative feedback. The theta power differences between negative feedback and positive feedback were larger in the low WM load condition than in the high WM load condition. For delta oscillation, the power was increased after positive feedback compared to after negative feedback, but there was no difference at different WM load levels. <break/>The RewP results indicate that the participants process feedback valence information well under all three WM load conditions in the experiment. The LPP results suggest that the participants assigned additional emotional motivation to the feedback outcome as a result of their cognitive efforts under high WM load conditions. The ERP results for the time domain dimension showed that the effect of the WM load on feedback processing was most noticeable in the later stages of feedback processing. Moreover, these observations support the argument that the RewP and theta power reflect distinct cognitive phenomena; namely, the RewP reflects the processing of feedback valence in the anterior cingulate cortex (ACC), whereas theta oscillations reflect the role of the ACC in cognitive control. The WM load selectively modulates the cognitive control process in the ACC.