On the positive side of error processing: Error-awareness positivity revisited

Department of Cognitive Science, The Hebrew University of Jerusalem, Jerusalem, Israel.
European Journal of Neuroscience (Impact Factor: 3.18). 05/2009; 29(7):1522-32. DOI: 10.1111/j.1460-9568.2009.06690.x
Source: PubMed


Performance errors are indexed in the brain even if they are not consciously registered, as demonstrated by the error-related negativity (ERN or Ne) event-related potential. It has recently been shown that another response-locked potential, the error positivity (Pe), follows the Ne, but only in those trials in which the participants consciously detect making the error ('Aware Errors'). In the present study we generalize these findings to an auditory task and investigate possible caveats in the interpretation of the Pe as an index of error awareness. In an auditory Go/No-Go error-awareness task (auditory EAT) participants pressed an additional 'fix error' button after noticing that they had made an error. As in visual tasks, the Ne was similar for aware ('fixed') and unaware ('unfixed') errors, while the Pe was enhanced only for Aware Errors. Within subjects, the Ne and Pe behaved in similar fashions for auditory and visual errors. A control condition confirmed that the awareness effect was not due to the requirement to report error awareness. These results reinforce the evidence in favor of the Pe as a correlate of conscious error processing, and imply that this process is not modality-specific. Nevertheless, single-trial analysis suggested that the Pe may be a delayed P3b related to stimulus processing rather than to response monitoring.

Full-text preview

Available from:
  • Source
    • "ger Ne / ERN amplitudes for aware than for unaware errors ( e . g . , Scheffers and Coles , 2000 ; Maier et al . , 2008 ; Steinhauser and Yeung , 2010 ; Hewig et al . , 2011 ; Wessel et al . , 2011 ) , while other studies have not reported such an effect ( e . g . , Nieuwenhuis et al . , 2001 ; Endrass et al . , 2007 ; O ' Connell et al . , 2007 ; Shalgi et al . , 2009 ; Hughes and Yeung , 2011 ) ."
    [Show abstract] [Hide abstract]
    ABSTRACT: Detecting one's own errors and appropriately correcting behavior are crucial for efficient goal-directed performance. A correlate of rapid evaluation of behavioral outcomes is the error-related negativity (Ne/ERN) which emerges at the time of the erroneous response over frontal brain areas. However, whether the error monitoring system's ability to distinguish between errors and correct responses at this early time point is a necessary precondition for the subsequent emergence of error awareness remains unclear. The present study investigated this question using error-related brain activity and vocal error signaling responses in seven human patients with lesions in the rostral anterior cingulate cortex (rACC) and adjoining ventromedial prefrontal cortex, while they performed a flanker task. The difference between errors and correct responses was severely attenuated in these patients indicating impaired rapid error monitong, but they showed no impairment in error signaling. However, impaired rapid error monitoring coincided with a failure to increase response accuracy on trials following errors. These results demonstrate that the error monitoring system's ability to distinguish between errors and correct responses at the time of the response is crucial for adaptive post-error adjustments, but not a necessary precondition for error awareness.
    Full-text · Article · Jul 2015 · Frontiers in Human Neuroscience
  • Source
    • "Furthermore, it is thought that the accumulation of evidence that indicates errors leads to subjective awareness of error (Klein et al., 2013; Ullsperger et al., 2010). Importantly, reports have repeatedly demonstrated that compared with the ERN, Pe amplitude more strongly reflects the degree of error awareness (Endrass et al., 2007, 2005; Murphy et al., 2012; Nieuwenhuis et al., 2001; Overbeek et al., 2005; Shalgi et al., 2009). Therefore we can attribute the linkage between HCT score and Pe to the degree of conscious processing (i.e. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Although self-monitoring is an important process for adaptive behaviors in multiple domains, the exact relationship among different internal monitoring systems is unclear. Here, we aimed to determine whether and how physiological monitoring (interoception) and behavioral monitoring (error processing) are related to each other. To this end we examined within-subject correlations among measures representing each function. Score on the heartbeat counting task (HCT) was used as a measure of interoceptive awareness. The amplitude of two event-related potentials (error-related negativity [ERN] and error-positivity [Pe]) elicited in error trials of a choice-reaction task (Simon task) were used as measures of error processing. The Simon task presented three types of stimuli (objects, faces showing disgust, and happy faces) to further examine how emotional context might affect inter-domain associations. Results showed that HCT score was robustly correlated with Pe amplitude (the later portion of error-related neural activity), irrespective of stimulus condition. In contrast, HCT score was correlated with ERN amplitude (the early component) only when participants were presented with disgust-faces as stimuli, which may have automatically elicited a physiological response. Behavioral data showed that HCT score was associated with the degree to which reaction times slowed after committing errors in the object condition. Cardiac activity measures indicated that vigilance level would not explain these correlations. These results suggest a relationship between physiological and behavioral monitoring. Furthermore, the degree to which behavioral monitoring relies on physiological monitoring appears to be flexible and depend on the situation.
    Full-text · Article · Oct 2014 · International Journal of Psychophysiology
  • Source
    • "As expected, the analysis reveals an increase in ERN amplitude over time, in consonance with Müller et al. (2005), as well as an increase in the amplitude of the P E . This is in line with the first-indicator hypothesis (Holroyd & Coles, 2002; Stahl, 2010), which would posit greater reliance on internal information along the course of the experimental block (Müller et al., 2005), and also concurs with the ERN reflecting a violation of the expected performance outcome (Holroyd & Coles, 2002), and the P E correlating with error awareness (Endrass et al., 2007; Hughes & Yeung, 2011; Nieuwenhuis et al., 2001; Shalgi et al., 2009), since both would be greater as learning becomes more secure. In contrast, no effect of learning is observed on the FRN, as opposed to the report of Müller et al. (2005). "
    [Show abstract] [Hide abstract]
    ABSTRACT: Neural responses to performance errors and external feedback have been suggested to be altered in obsessive-compulsive disorder. In the current study, an associative learning task was used in healthy participants assessed for obsessive-compulsive symptoms by the OCI-R questionnaire. The task included a condition with equivocal feedback that did not inform about the participants' performance. Following incorrect responses, an error-related negativity and an error positivity were observed. In the feedback phase, the largest feedback-related negativity was observed following equivocal feedback. Theta and beta oscillatory components were found following incorrect and correct responses, respectively, and an increase in theta power was associated with negative and equivocal feedback. Changes over time were also explored as an indicator for possible learning effects. Finally, event-related potentials and oscillatory components were found to be uncorrelated with OCI-R scores in the current non-clinical sample.
    Full-text · Article · Jul 2014 · Biological Psychology
Show more