This talk presents findings from two experiments investigating how beliefs about the efficacy of one’s actions shape cognitive engagement and task performance. The central question is how performance feedback influences cognitive effort by modulating perceived efficacy—that is, the extent to which individuals believe their effort can meaningfully affect outcomes. Participants completed a color discrimination task while behavioral responses and task-related EEG activity were recorded.

In Experiment 1 (N = 40), participants were told that the response hardware would occasionally register incorrect answers. Consistent with this instruction, feedback was manipulated to display an incorrect response on one-third of trials. This manipulation produced slower reaction times and reduced contingent negative variation (CNV) amplitudes, indicating decreased cognitive engagement when confidence in action–outcome mappings was undermined.

In Experiment 2 (N = 80), participants received false feedback indicating that their performance was above, below, or near a threshold linked to avoiding an aversive cold-water exposure. In the performance-contingent condition, feedback signaling proximity to the threshold elicited faster and more accurate responses, along with more negative CNV amplitudes, suggesting increased cognitive effort when goal attainment was perceived as achievable. This pattern was not observed in a control condition where the aversive outcome was unrelated to performance.

Across both experiments, cognitive effort tracked perceived efficacy: engagement decreased when feedback reduced confidence in performance–outcome contingencies and increased when feedback suggested that effort could effectively influence meaningful outcomes. These results identify perceived efficacy as a central mechanism governing adaptive allocation of cognitive control.

This talk presents findings from an EEG study investigating the influence of perceived outcome-controllability on proactive control and outcome monitoring. Thirty-five participants performed in a color-discrimination task with manipulated performance feedback. Across periods with veridical feedback, periods with feedback suggesting an increased error rate due to impaired personal performance, and periods with feedback suggesting an increased error rate due to malfunctioning response buttons, we observed how the differential attribution of action-outcome disruptions shapes outcome monitoring and cognitive states of perceived agency, which in turn affect task preparation.

We used EZ-diffusion modeling to decompose behavioral performance, and EEG time-frequency analysis focussed on event-related alpha, beta and theta oscillations during task preparation and outcome monitoring intervals.

Participants responded significantly slower in the self-attribution of errors condition compared to neutral feedback, and also significantly slower in the system-attribution of errors condition compared to self-attribution of errors. Decomposition with the diffusion model indicates that a general decrease of response times with manipulated feedback can be attributed to decreased drift rates, whereas the difference between the self and system error conditions are reflected in the non-decision time.

In the EEG, the manipulated feedback was reflected in attenuated decreases of occipital alpha and sensorimotor beta power during the cue-target interval. Furthermore, system-versus self-attributed errors elicited stronger feedback-locked midfrontal theta responses.

Our findings suggest a functional dissociation within the agency inference process, where perceived controllability regulates preparatory investment of cognitive resources, while the attribution of action-outcome discrepancies seem to modulate sensory processes or motor-execution.

Threat avoidance requires affective control, especially when threats are imminent vs. distant and when outcomes are worse (negative prediction error, PE-) or better (positive prediction error, PE+) than expected. Electroencephalography (EEG) studies linked PE processing to slow-wave oscillations, namely frontal midline theta (FMθ, 1-4 Hz) and centroparietal delta activity (1-3 Hz), as well as phase-locked activity within FMθ/delta time windows. However, electrocortical threat PE processing and its modulation by threat imminence remain largely unknown. Here, we assessed feedback-locked EEG activity of 90 participants performing two reversal learning tasks. In one task, negative feedback indicated immediate threat (one loud noise burst), whereas in the other task it signalled delayed threat (accumulated noise bursts delivered at the end of the block). The colour of the inter-trial fixation cross indicated participants’ progress toward the end of each block. Single-trial analyses revealed greater FMθ sensitivity to feedback valence and PE processing in immediate vs. delayed threat. This effect was further qualified by a task x feedback valence x PE interaction, suggesting stronger PE+ processing for immediate threat and stronger PE- processing for delayed threat. There was no empirical support that EEG activity was modulated by proximity to exiting (immediate threat) or entering (delayed threat) the threat encounter. However, preliminary analyses showed a general decrease in late delta activity as participants approached the end of a block. These findings suggest that threat imminence may shape neural PE representations, promoting safety-driven learning during immediate threat and a stronger focus on potential negative outcomes during distant threat.

Aggressive behaviour emerges from a complex dynamic interplay between situational provocation, individual differences, and neurocognitive processes. This preregistered study (N=114) employed single-trial multilevel modelling to examine how personality traits (trait anger, trait aggression), midfrontal theta power during punishment selection (as a marker of cognitive effort), and provoked aggression (high provocation / low provocation per block) jointly predicted subsequent punishment choices in a competitive Taylor Aggression Paradigm. In addition, the feedback-related negativity (FRN) to a (fictive) opponent facial response (smiling / suffering / neutral) to participant aggression was examined. To consolidate EEG results, three different reference schemes were used.

Results indicated that as predicted, high trait expression of trait anger or trait aggression was leading to more punishment especially in high provocation context and high provocation led to high punishment in general. In addition, trait aggression also led to high punishment in general.

Concerning midfrontal theta, we found as predicted that high midfrontal theta during trials led to less punishment as well as the predicted interaction with provocation level.

FRN responses to the facial expressions of the (fictive) opponents depending on the participants´ aggression was more negative to neutral facial responses than to any other facial reaction as partly predicted.

These findings indicate that midfrontal theta responses serve as a marker for cognitive effort and cognitive control to overcome default behaviour while behaviour is influenced by personality dispositions and situational context. Also, a lack of reaction to retaliatory aggression is negatively evaluated compared to any other response.