Abstract
Abstract
Sense of agency (SoA), the experience of controlling one’s actions and their consequences, is crucial for self-representation and adaptive goal-directed behavior. Much of today’s cognitive work is performed through interaction with systems that are not entirely reliable or are prone to operator error. Against this background, it is of particular interest to understand how perceived outcome-controllability and attribution of action-outcome disruptions feed back into cognitive processing as states of perceived agency. In this EEG study, we manipulated performance feedback in a color-discrimination task to dissociate self-attributed from system-attributed errors. Thirty-five participants completed blocks with veridical feedback, feedback suggesting increased error rates due to impaired personal performance, and feedback indicating malfunctioning response buttons. Behavioral performance was decomposed using the EZ-diffusion model, and time–frequency analyses focused on preparatory alpha and beta oscillations and feedback-locked theta activity. Both manipulated feedback conditions led to slower responses compared to veridical feedback. Diffusion modeling revealed that general performance slowing was driven by reduced drift rates, whereas differences between self- and system-attributed errors were reflected in nondecision time. In the EEG, manipulated feedback attenuated cue-related decreases in occipital alpha and sensorimotor beta power during the cue-target interval. In addition, system- versus self-attributed errors elicited stronger feedback-related midfrontal theta responses. Our findings suggest a functional dissociation within the agency inference process, where perceived controllability regulates preparatory investment of cognitive resources on a global level, while the attribution of action-outcome discrepancies seem to modulate evaluative processing.