Error-related brain acitivty reflects : independent systems in human error monitoring

Performance monitoring is a key function of human cognition and critical for achieving goal-directed behavior. In recent years, research has particularly focused on how the brain detects and evaluates behavioral errors. Most studies investigated two neural correlates of performance monitoring in the human scalp EEG. In particular, the error-related negativity (Ne/ERN) is a negative fronto-central deflection observed immediately after an erroneous response, representing an early and unconscious stage of error processing. The Ne/ERN is followed by the later error positivity (Pe), a broader positivity viewed as a correlate of conscious error processing. Whereas a large amount of research has been conducted on these neural correlates, fundamental questions on their relationship remain. Crucially, it is still unclear whether both components represent functionally independent processes of error monitoring or whether the two components are part of a single mechanism. The first possibility implies that the earlier Ne/ERN provides the basis for the later emergence of the Pe and error awareness in a cascade-like architecture of error monitoring. The other possibility is that the Pe and error awareness can emerge independently of the Ne/ERN, which implies that different error monitoring mechanisms exist and may proceed independently. The thesis addresses the important question whether Ne/ERN provides necessary information for error awareness and Pe. To this aim, behavioral and psychophysiological studies were conducted: (1) In a first part, we investigated whether Ne/ERN and Pe are causally related (study 1). We developed a novel experimental paradigm based on the classical letters flanker task. In this paradigm, participants have to classify targets but ignore irrelevant distractors (flankers) that are always associated with an incorrect response. Targets but not flankers are masked with varying target-masking intervals. On some trials, no target at all is presented, thus preventing the representation of a correct response. Importantly, the lack of a representation of the correct response also prevents the emergence of Ne/ERN. However, because flankers are easily visible and responses to these flankers are always incorrect, conscious detection of these flanker errors is still possible. The presence of a Pe in the absence of Ne/ERN for flanker errors provides evidence for independent error monitoring processes. (2) In a second part, we investigated whether the Ne/ERN and Pe are differentially sensitive to temporal aspects of conscious error detection. Whereas the Pe is assumed to be the earliest correlate of the emergence of conscious error perception, participants often report the feeling of having detected an error even before the erroneous response was actually executed (“early error sensations”). The first goal of this part was to investigate whether such anecdotal evidence can be measured empirically. In study 2, a series of behavioral experiments using different methodological procedures were conducted. Participants during choice tasks have to report whether errors were accompanied or not by early error sensations (i.e. early and late detected errors) or give confidence judgments about early error sensations. Participants frequently reported early errors with high level of confidence. Subsequently, we studied how error-related brain activity reflects the emergence of early error sensations. In study 3, we measured EEG activity and compared early and late detected errors. Results showed that while the Pe reliably reflects the early error sensations (larger amplitude for early errors), the Ne/ERN does not (no differences between early and late errors). Crucially, the Ne/ERN and the Pe responded differently to temporal aspects of error awareness, meaning that Ne/ERN and Pe could rely on different types of information for error detection. This again speaks for the idea of independent systems of error monitoring