Context dependence of the event-related brain potential associated with reward and punishment

The error-related negativity (ERN) is an event-related brain potential elicited by error commission and by presentation of feedback stimuli indicating incorrect performance. In this study, the authors report two experiments in which participants tried to learn to select between response options by trial and error, using feedback stimuli indicating monetary gains and losses. The results demonstrate that the amplitude of the ERN is determined by the value of the eliciting outcome relative to the range of outcomes possible, rather than by the objective value of the outcome. This result is discussed in terms of a recent theory that holds that the ERN reflects a reward prediction error signal associated with a neural system for reinforcement learning.     

Self‐report and behavioral measures of reward sensitivity predict the feedback negativity

Rewards are integral to learning associations that aid in survival. The feedback negativity (FN), an event‐related potential that differentiates outcomes indicating monetary losses versus gains, has recently emerged as a possible neural measure of reward processing. If this view is correct, then the FN should correlate with measures of reward sensitivity in other domains, although few studies have investigated this question. In the current study, 46 participants completed a self‐report measure of reward responsiveness, a signal detection task that generated a behavioral measure of reward sensitivity, and a gambling task that elicited an FN. Consistent with the view that the FN reflects reward‐related neural activity, a larger FN correlated with increased behavioral and self‐report measures of sensitivity to reward.     

The role of outcome expectations in the generation of the feedback‐related negativity

The feedback‐related negativity (FRN) is thought to index activity within the midbrain dopaminergic reward‐learning system, with larger FRN magnitudes observed when outcomes are worse than expected. This view holds that the FRN is an index of neural activity coding for prediction errors, and reflects activity that can be used to adaptively alter future performance. Untested to date, however, is a key prediction of this view: the FRN should not appear in response to negative outcomes when outcome expectations are not allowed to develop. The current study tests this assumption by eliciting FRNs to win and loss feedback in conditions of participant choice, participant observation of computer choice, and, critically, simple presentation of win or loss feedback in the absence of a predictive choice cue. Whereas FRNs were observed in each of the conditions in which there was time for an expectation to develop, no FRN was observed in conditions without sufficient time for the development of an expectation. These results provide empirical support for an untested but central tenet of the reinforcement learning account of the genesis of the FRN.     

How does cognitive reappraisal affect the response to gains and losses?

To investigate the influence of cognitive reappraisal, one important kind of emotion regulation strategy, on psychological and electrophysiological responses to gains and losses, a monetary gambling task was performed in two conditions, that is, spontaneity and regulation. Event‐related potentials (ERP) and self‐rating emotional experiences to outcome feedback were recorded during the task. Cognitive reappraisal reduced self‐rating emotional experience to both gains and losses and the amplitudes of the feedback‐related negativity (FRN) and the P3 of ERPs. According to these results, we suggest that the application of cognitive reappraisal strategy significantly modulated the motivational salience of current outcomes, thus weakening the subjective emotional experience elicited. In addition, cognitive reappraisal might have changed the allocation of cognitive resources during outcome evaluation. This study extends emotion regulation studies by applying monetary outcomes as emotional stimuli, and also implicates the significance of emotion regulation in decision‐making processes.     

Retest reliability of medial frontal negativities during performance monitoring

The error‐related negativity (ERN) and feedback‐related negativity (FRN) have been used as electrophysiological indices of performance monitoring produced in response to internally generated (errors) and externally generated (feedback) activations of the anterior cingulate cortex (ACC). No studies to date have systematically examined the measurement reliability of these components. In this article, we present the retest reliability of the ERN and FRN during response tasks designed to elicit errors or feedback responses on two occasions. Data from four experiments are presented in which participants performed tasks over various periods of time. Results indicate good retest reliability of the ERN and FRN amplitudes and source generation of these components. The present article provides important validation of the ERN and FRN as stable and trait‐like electrophysiological reflections of performance monitoring and ACC functional integrity.     

Time‐frequency approaches to investigating changes in feedback processing during childhood and adolescence

Processing feedback from the environment is an essential function during development to adapt behavior in advantageous ways. One measure of feedback processing, the feedback negativity (FN), is an ERP observed following the presentation of feedback. Findings detailing developmental changes in the FN have been mixed, possibly due to limitations in traditional ERP measurement methods. Recent work shows that both theta and delta frequency activity contribute to the FN; utilizing time‐frequency methods to measure change in power and phase in these frequency bands may provide more accurate representation of feedback processing development in childhood and adolescence. We employ time‐frequency power and intertrial phase synchrony measures, in addition to conventional time‐domain ERP methods, to examine the development of feedback processing in the theta (4–7 Hz) and delta (.1–3 Hz) bands throughout adolescence. A sample of 54 female participants (8–17 years old) completed a gambling task while EEG was recorded. As expected, time‐domain ERP amplitudes showed no association with age. In contrast, significant effects were observed for the time‐frequency measures, with theta power decreasing with age and delta power increasing with age. For intertrial phase synchrony, delta synchrony increased with age, while age‐related changes in theta synchrony differed for gains and losses. Collectively, these findings highlight the importance of considering time‐frequency dynamics when exploring how the processing of feedback develops through late childhood and adolescence. In particular, the role of delta band activity and theta synchrony appear central to understanding age‐related changes in the neural response to feedback.     

Time‐frequency theta and delta measures index separable components of feedback processing in a gambling task

Previous work using gambling tasks indicate that the feedback negativity (FN) reflects primary or salient stimulus attributes (often gain vs. loss), whereas the feedback‐P300 appears sensitive to secondary stimulus information. A recent time‐frequency approach has characterized separable theta (3–7 Hz) and delta (0–3 Hz) feedback processes, independently sensitive to primary feedback attributes, specifically loss and gain outcomes, respectively. The current study extends this time‐frequency work to evaluate both primary and secondary (relative outcome and outcome magnitude) feedback attributes. Consistent with previous reports, theta indexed an initial, lower‐level response sensitive to the primary (most salient) feedback attributes (specifically losses), while delta was sensitive to both primary attributes (specifically gains) and assessed secondary stimulus features.     

Valences of self-evaluation and approach-avoidance tendencies: research based on regulatory focus theory

Four studies were conducted to investigate the relationship between valences of self-evaluation and approach-avoidance tendencies. Based on regulatory focus theory (Higgins, 1997, 1998), we predicted that positivity of self-evaluation is related to the tendency to approach gains, while negativity of self-evaluation is related to the tendency to avoid losses. In Study 1, a self-report measure of behavioral tendencies for approaching gains and avoiding losses was developed. In Studies 2 to 4, correlations between these approach/avoidance tendencies and various kinds of self-evaluations were examined. Overall, the authors' predictions were supported. The results suggest that the self-evaluation system and the self-regulation system work in close cooperation with each other in controlling human behavior.     

Predictive information processing in the brain: errors and response monitoring

The monitoring of one's own actions is essential for adjusting behavior. In particular, response errors are important events that require behavioral adjustments. Correct and incorrect responses, as well as feedback to responses, are followed by brain activity originating mainly in the anterior cingulate, which can be measured with fMRI and event-related potential (ERP) techniques. After each response a small negativity (Nc or CRN) is elicited in the ERP, which is strongly enhanced in incorrect trials (Ne or ERN). Following feedback stimuli that signal a negative outcome of an action, a similar negativity, the feedback-related negativity (FRN) is elicited. Recently it has been shown that these neurophysiological correlates of response monitoring and evaluation can be classified even on the single-trial level in the EEG and thus could be utilized not only to distinguish between correct and erroneous actions, but also can be used online for a wide range of applications such as prediction of clinical outcomes or brain computer interfaces.     

It's worse than you thought: the feedback negativity and violations of reward prediction in gambling tasks

The reinforcement learning theory suggests that the feedback negativity should be larger when feedback is unexpected. Two recent studies found, however, that the feedback negativity was unaffected by outcome probability. To further examine this issue, participants in the present studies made reward predictions on each trial of a gambling task where objective reward probability was indicated by a cue. In Study 1, participants made reward predictions following the cue, but prior to their gambling choice; in Study 2, predictions were made following their gambling choice. Predicted and unpredicted outcomes were associated with equivalent feedback negativities in Study 1. In Study 2, however, the feedback negativity was larger for unpredicted outcomes. These data suggest that the magnitude of the feedback negativity is sensitive to violations of reward prediction, but that this effect may depend on the close coupling of prediction and outcome.     

Differential effects of instructed and objective feedback reliability on feedback‐related brain activity

Feedback reliability refers to the probability that the same decision leads to the same positive or negative feedback in the future. Previous research has shown that unreliable feedback is associated with attenuated feedback‐related brain activity in ERPs, represented by a reduced fronto‐central valence effect (feedback‐related negativity or reward positivity) and a reduced feedback‐related P3. Here, we asked whether these effects reflect top‐down mechanisms or whether they can be explained by implicit feedback‐outcome contingency learning. In two experiments, participants performed a trial‐and‐error learning task while subjective or objective feedback reliability was varied across blocks. In Experiment 1, we manipulated instructed feedback reliability while holding objective feedback reliability constant. Low instructed feedback reliability led to an attenuation of the fronto‐central valence effect and the P3. In Experiment 2, we manipulated objective feedback reliability while holding instructed feedback reliability constant. Here, no modulation of feedback‐related brain activity was observed. These results suggest that effects of feedback reliability are driven by top‐down mechanisms based on explicit knowledge. Specifically, effects on the fronto‐central valence effect could indicate a devaluation of unreliable feedback or a bias on the generation or utilization of reward prediction errors.     

Event‐related potentials in response to feedback following risk‐taking in the hot version of the Columbia Card Task

Given the importance of risk‐taking in individuals’ personal and professional life, several behavioral tasks for measuring the construct have been developed. Recently, a new task was introduced, the Columbia Card Task (CCT). This task measures participants’ risk levels and establishes how sensitive participants are to gains, losses, and probabilities when taking risk. So far, the CCT has been examined in behavioral studies and in combination with several (neuro)biological techniques. However, no electroencephalography (EEG) research has been done on the task. The present study fills this gap and helps to validate this relatively new experimental task. To this end, n = 126 students were asked to complete self‐reports (reward responsiveness, impulsiveness, and sensation‐seeking) and to perform the CCT (and other risk tasks) in an EEG setup. The results show that feedback appraisal after risky decision‐making in the CCT was accompanied by a feedback‐related negativity (FRN) and a P300, which were stronger in response to negative than positive feedback. Correlations between the FRN and P300 difference wave on the one hand and risk‐related self‐reports and behavior on the other were nonsignificant and small, but were mostly in the expected direction. This pattern did not change after excluding participants with psychiatric/neurological disorders and outliers. Excluding participants with reversed (positive > negative) difference waves strengthened FRN correlations. The impact such individuals can have on the data should be taken into account in future studies. Regarding the CCT in particular, future studies should also address its oddball structure and its masking of true values (censoring).     

Expectancy affects the feedback‐related negativity (FRN) for delayed feedback in probabilistic learning

Learning from feedback is a prerequisite for adapting to the environment. Prediction error signals coded by midbrain dopamine (DA) neurons are projected to the basal ganglia and anterior cingulate cortex (ACC). It has been suggested that neuronal activity shifts away from the DA system when feedback is delayed. The feedback‐related negativity (FRN), an ERP that is generated in the ACC and has been shown to be sensitive to feedback valence and prediction error magnitude, was found to be reduced for delayed feedback. It has, however, not yet been investigated if the FRN for delayed feedback reflects a reward prediction error. In this study, effects of feedback delay (1 s vs. 7 s) on the processing of expected and unexpected positive and negative feedback were investigated in a between‐subjects design in healthy human participants conducting a probabilistic feedback learning task. FRN and P300 amplitudes were decreased for subjects learning from delayed compared to immediate feedback. Importantly, the FRN, extracted from the negative‐positive feedback difference wave, was significantly smaller for expected compared to unexpected feedback for both the immediate and delayed feedback conditions. Expectancy effects for the P300 were also seen, but did not interact with feedback valence. These results demonstrate an influence of feedback expectancy, and thus the prediction error, on early feedback processing even for delayed feedback, suggesting that neuronal structures underlying feedback processing are comparable for immediate and delayed feedback, at least to some extent. Modulations of the P300 by feedback delay may be linked to feedback salience.     

The effect of interpersonal competition on monitoring internal and external error feedback

The present study examined the effect of the social context on early emotional appraisal of performance errors and negative feedback reflected by the error‐related negativity (ERN), feedback‐related negativity (FRN), and P300. Participants performed a probabilistic learning task in which they received valid and invalid performance feedback. During one half of the task they were led to believe that they were competing online against another participant. As expected, the ERN following response errors was enhanced in the competition compared to the neutral condition. The FRN was more negative following negative compared to positive feedback and valid compared to invalid feedback, but only during competition. The P300 was larger to false positive than false negative feedback, which was independent of the social context. In conclusion, ERN and FRN, but not P300, may be sensitive to affective distress elicited by expectation violations during social interaction.     

The feedback-related negativity (FRN) in adolescents

This study examined age-related differences in the ERP correlates of external feedback processing (i.e., the feedback-related negativity [FRN]) in adolescent and young adult males, using a simple gambling task involving unpredictable monetary losses and gains of low and high magnitude. The FRN was larger after losses than gains, and was modulated by the magnitude of gains, but not the magnitude of losses, for all participants regardless of age. FRN amplitude was larger in adolescents than adults and also discriminated relatively less strongly between gains and losses in adolescents. In addition, the morphology of the waveform after high losses suggests that feedback in this condition may have been processed less efficiently by adolescents. Our results suggest that, although the FRN in adults and adolescents share some common characteristics, the neural processes that generate the FRN are still developing in midadolescence. These findings are discussed in the context of adolescent risk taking.     

Reliability of the electrocortical response to gains and losses in the doors task

The ability to differentiate between rewards and losses is critical for motivated action, and aberrant reward and loss processing has been associated with psychopathology. The reward positivity (RewP) and feedback negativity (FN) are ERPs elicited by monetary gains and losses, respectively, and are promising individual difference measures. However, few studies have reported on the psychometric properties of the RewP and FN—crucial characteristics necessary for valid individual difference measures. The current study examined the internal consistency and 1‐week test‐retest reliability of the RewP and FN as elicited by the doors task among 59 young adults. The RewP, FN, and their difference score (ΔRewP) all showed significant correlations between Time 1 and Time 2. The RewP and FN also achieved acceptable internal consistency at both time points within 20 trials using both Cronbach's α and a generalizability theory‐derived dependability measure. Internal consistency for ΔRewP was notably weaker at both time points, which is expected from two highly intercorrelated constituent scores. In conclusion, the RewP and FN have strong psychometric properties in a healthy adult sample. Future research is needed to assess the psychometric properties of these ERPs in different age cohorts and in clinical populations.     

Temporal dissociation of salience and prediction error responses to appetitive and aversive taste

The feedback‐related negativity (FRN), a frontocentral ERP occurring 200–350 ms after emotionally valued outcomes, has been posited as the neural correlate of reward prediction error, a key component of associative learning. Recent evidence challenged this interpretation and has led to the suggestion that this ERP expresses salience instead. Here, we distinguish between utility prediction error and salience by delivering or withholding hedonistically matched appetitive and aversive tastes, and measure ERPs to cues signaling each taste. We observed a typical FRN (computed as the loss‐minus‐gain difference wave) to appetitive taste, but a reverse FRN to aversive taste. When tested axiomatically, frontocentral ERPs showed a salience response across tastes, with a particularly early response to outcome delivery, supporting recent propositions of a fast, unsigned, and unspecific response to salient stimuli. ERPs also expressed aversive prediction error peaking at 285 ms, which conformed to the logic of an axiomatic model of prediction error. With stimuli that most resemble those used in animal models, we did not detect any frontocentral ERP signal for utility prediction error, in contrast with dominant views of the functional role of the FRN ERP. We link the animal and human literature and present a challenge for current perspectives on associative learning research using ERPs.     

Contextual valence modulates the neural dynamics of risk processing

A well‐known bias in risky decision making is that most people tend to be risk averse when gains are salient but risk seeking when losses are salient. The present study addressed the neural dynamics of this process by recording ERPs during a gambling task in a gain and a loss context. Behaviorally, participants were found to be risk averse in the gain context but risk neutral in the loss context. During the anticipation stage, an increased stimulus‐preceding negativity was elicited by high‐ versus low‐risk choices in the gain but not the loss context. During the outcome‐appraisal stage, the feedback‐related negativity was larger after high‐ versus low‐risk choices in the gain instead of the loss context. For the P300, an outcome valence effect (a larger P300 for gain vs. loss outcomes) emerged following the high‐ versus low‐risk decisions in the gain but not the loss context. Our findings suggest that risk processing can be modulated by the context of valence during the anticipation stage and by both the contextual valence and the outcome valence during the outcome‐appraisal stage, which may be driven by the motivational salience imposed by the context of valence.     

Electrophysiological correlates of prediction formation in anticipation of reward‐ and punishment‐related feedback signals

Feedback processing during decision making involves comparing anticipated and actual outcome. Although effects on ERPs of valence, magnitude, expectancy, and context during feedback processing have been extensively investigated, the electrophysiological processes underlying prediction formation in anticipation of feedback signals have received little attention. The aim of the present study was to explore these processes of prediction formation and their influence on subsequent feedback signals. Twenty healthy, right‐handed volunteers performed a forced‐choice task in which they had to indicate which of two presented objects was more expensive. After the volunteer's choice, an expert cue, which was accurate in 80% of trials, was presented to manipulate prediction formation about future reward and punishment. ERPs were recorded during presentation of the expert cue and during feedback. Results revealed that prediction formation of future rewards and punishments is accompanied by differences in the P2 component and a subsequent delay period. During feedback processing, the prediction‐related P2 was associated with the processing of valence reflected in the feedback‐related P2. Furthermore, the prediction‐related difference in the delay period was associated with error processing in feedback‐related medial frontal negativity. These findings suggest that prediction signals prior to feedback contain information about whether a prediction is correct or wrong (expectancy) and if the outcome will be a reward or punishment (valence).     

Dorsal anterior cingulate cortex shows fMRI response to internal and external error signals

In our event-related functional magnetic resonance imaging (fMRI) experiment, participants learned to select between two response options by trial-and-error, using feedback stimuli that indicated monetary gains and losses. The results of the experiment indicate that error responses and error feedback activate the same region of dorsal anterior cingulate cortex, suggesting that this region is sensitive to both internal and external sources of error information.