What are the influences of orthogonally-manipulated valence and arousal on performance monitoring processes? The effects of affective state

https://doi.org/10.1016/j.ijpsycho.2013.01.005Get rights and content

Abstract

Studies of the influence of affective state on the cognitive control process of performance monitoring are mixed and few studies have orthogonally manipulated affective valence and arousal. Performance monitoring can be measured using behaviors (e.g., response times and error rates) and components of the event-related potentials (ERPs), such as the error-related negativity (ERN), correct-related negativity (CRN), and post-error positivity (Pe). We used a pre/post design and standard mood induction paradigm in 121 healthy participants randomly assigned to orthogonal valence (positive or negative) and arousal (high or low) conditions (i.e., happy, calm, anxious, or sad mood states). Following mood induction, valence and arousal ratings differed between groups. Behavioral findings showed decreased accuracy in participants with high arousal and negative valence (i.e., anxious condition), but no additional response time (RT), post-error slowing, or accuracy effects. Amplitude of the CRN differentiated high and low valence, but was not related to arousal. Positive valence was associated with decreased CRN amplitude even when baseline affect and demographic variables were controlled. Valence and arousal did not significantly differentiate the amplitude of the ERN, although the ERN minus CRN difference was related to arousal but not valence ratings in multiple regression analyses. Affect-related differences were not shown for the Pe. Findings provide a context to understand how dimensional aspects of emotional valence and arousal influence performance-monitoring processes and suggest a need for further research on the functional role of the CRN and its relation to affective valence.

Highlights

► Examined orthogonal contributions of valence and arousal to performance monitoring. ► Negative valence with high arousal (anxiety) showed decreased accuracy. ► CRN amplitude was related to positive valence, but not arousal. ► ERN minus CRN amplitude difference related to arousal, but not valence. ► Overall ERN and Pe amplitudes did not significantly differ by valence or arousal.

Introduction

Mood states can affect multiple cognitive processes, including memory (e.g., Chepenik et al., 2007), executive functioning (e.g., Phillips et al., 2002), and attentional control (e.g., Jefferies et al., 2008). It is unclear, however, whether cognitive control and, more specifically, performance monitoring abilities are influenced by mood states. Performance monitoring refers to the cognitive control process of monitoring activities to detect conflict or inaccuracies that may require increased cognitive resources (Carter et al., 1998). The neural bases of performance monitoring, particularly with regard to response monitoring, can be measured with millisecond accuracy using the error-related negativity (ERN), correct-response negativity (CRN), and post-error positivity (Pe) components of the scalp-recorded event-related potential (ERP).

The ERN is a negative potential that peaks approximately 50 ms after an erroneous response. The ERN is larger (i.e., more negative) in amplitude on error trials relative to correct trials and consistently manifests across multiple task situations and response modalities (Falkenstein et al., 1991, Gehring et al., 1993, Nieuwenhuis et al., 2001). There are several theories regarding the functional significance of the ERN, with the most commonly accepted theories indicating that the ERN represents the detection of response-related conflict, the detection of a mismatch between error and correct trials, a reinforcement learning indication of incorrect performance prediction, or an affective response to mistakes (see Hoffman and Falkenstein, 2012, Olvet and Hajcak, 2008, van Veen and Carter, 2002).

The CRN occurs in the same time frame as the ERN, but on correct trials. The functional significance of the CRN remains a matter of debate, although some suggest that the CRN represents an accurate comparison of correct and error representations, an index of level of response certainty, an evaluation of whether a response strategy is adaptive, or an index of the anticipated probability of a stimulus presentation (Bartholow et al., 2005, Endrass et al., 2012a, Ford, 1999, Pailing and Segalowitz, 2004, Scheffers and Coles, 2000, Vidal et al., 2000). Source localization studies indicate that the ERN and CRN have a similar frontal-medial scalp distribution and originate broadly from areas near the anterior cingulate cortex (ACC; Brazdil et al., 2005, Falkenstein et al., 2000, Roger et al., 2010, Stemmer et al., 2004, van Veen and Carter, 2002).

The Pe has a posterior scalp distribution and occurs between 200 and 500 ms after an error (Falkenstein et al., 1991, Falkenstein et al., 2000, Nieuwenhuis et al., 2001). The Pe typically follows the ERN and is thought to functionally reflect conscious error processing (Endrass et al., 2007, Endrass et al., 2012b, Hughes and Yeung, 2011, Larson and Perlstein, 2009, Nieuwenhuis et al., 2001, Shalgi et al., 2009) or continued processing of the affective significance of an error (Falkenstein et al., 2000, Overbeek et al., 2005). Source localization studies also implicate the ACC, along with potential additional areas such as the insula, as the neural generators of the Pe (e.g., Herrmann et al., 2004, Overbeek et al., 2005, Ullsperger et al., 2010).

Research investigating the influence of induced positive and negative affect on reflections of performance monitoring such as the ERN and Pe is mixed (Clayson et al., 2012). Several studies suggest that the ERN is not affected by changes in affective state (see Olvet and Hajcak, 2008). For example, ERN amplitude was not enhanced in individuals with a spider phobia under conditions of symptom provocation, when a live tarantula was held or placed in an aquarium next to the participant, but Pe amplitude was decreased — potentially suggesting decreased awareness or orientation to errors (Moser et al., 2005). Individuals in a sad mood condition did not show differences in ERN or Pe amplitudes relative to a neutral group, unless they showed high trait levels of neuroticism (Olvet and Hajcak, 2012). Our group recently showed nonsignificant differences in ERN and Pe amplitudes between groups of individuals receiving either encouraging or derogatory feedback on task performance (Clayson et al., 2012). These findings are consistent with research that showed no differences in ERN amplitude between conditions with and without trial-by-trial performance feedback (Olvet and Hajcak, 2009a) and between affective states such as tension, anger, and depressed mood state (Tops et al., 2006).

In contrast, a growing body of research suggests that affective states may indeed alter neural indices of performance monitoring such as the ERN and Pe. For example, Inzlicht and Al-Khindi (2012) cleverly used the misattribution of arousal paradigm to show decreased-amplitude ERN when participants could attribute their arousal to a placebo concoction rather than personal performance. Similarly, religious primes can either decrease ERN amplitude in people who believe in God or increase ERN amplitude in those who do not believe, suggesting that the state emotion has person-specific effects on ERN amplitude (Inzlicht and Tullett, 2010). Other studies that show such state-related changes in ERN and Pe amplitudes have utilized affective pictures, faces, and movie clips. For example, Larson et al. (2006) showed more negative ERN amplitudes when errors were made during presentation of pleasant pictures relative to unpleasant or neutral pictures. van Wouwe et al. (2011) reported decreased-amplitude ERN following positive video clips relative to following neutral video clips. In contrast, Wiswede et al. (2009) showed increased ERN amplitude following unpleasant pictures compared to neutral and pleasant pictures and Boksem et al. (2011) reported increased ERN amplitudes during a Simon task using emotionally-valent faces in the context of faces that showed an evaluative (i.e., disgusted) expression relative to those that showed a pleasant (i.e., happy) expression. There is considerable research that suggests that affective information can modulate both early and late ERP components (e.g., Foti et al., 2009), but the relationship between induced affective-state and ERN amplitude is not completely clear, although it appears that increased arousal or an absence of congruence with performance and mood states contributes to increased-amplitude ERN.

The current literature on affective state and performance monitoring, including the ERN, CRN, and Pe, is often confounded by a relative absence of empirical quantification of changes in mood state. For example, studies of picture viewing as a mood manipulation in studies of the ERN and Pe (e.g., Larson et al., 2006, Wiswede et al., 2009) imply that changes in mood occurred during the performance monitoring task, but these changes were not rigorously quantified (see Olvet and Hajcak, 2012). Further, studies of state mood and ERN amplitudes have tended to focus primarily on positive affect versus negative affect, without attention to the possible interactions of mood-related valence (i.e., negative versus positive) and mood-related arousal (i.e., low versus high). Indeed, we were unable to find studies that examined error-related performance monitoring along with orthogonal dimensions of valence and arousal. Valence and arousal represent independent dimensions of emotion, with valence ranging from pleasant to unpleasant and affect describing the energy or level of activation associated with the affect, tending to range from low to high energy (Barrett and Russell, 1998). Studies, such as the current investigation, orthogonally investigating valence and arousal effects will inform the field by providing a way to systematically examine the unique contributions of these dimensions as well as their potential interactive effects. These studies have implication for performance monitoring in both healthy individuals and those with affective dysregulation such as depression or anxiety, particularly since the impact of affect on conflict-related performance monitoring processes has not been fully established.

To address these shortcomings in the extant literature, the current examination quantified changes in mood-related arousal and valence to assess the relationship between changes in mood state and performance-monitoring indices (i.e., ERN, CRN, Pe) using a previously verified mood-induction procedure that involves music and mood-related rumination across a range of valence and arousal combinations (Eich et al., 2007, Jefferies et al., 2008). Specific mood induction conditions included sad (negative valence, low arousal), calm (positive valence, low arousal), anxious (negative valence, high arousal), and happy (positive valence, high arousal). We utilized a pre-mood induction to post-mood induction design so that each participant could serve as his or her own control. That is, each participant completed a neutral baseline for comparison in order to make the strongest possible inferences without the added error that would come from including another between-groups neutral condition.

The mixed findings in the state-related performance monitoring literature make determining hypotheses regarding the distinctive contributions of emotional valence and arousal difficult; however, there are some intimations for expectations in the current literature. Individuals with chronically-high levels of negative valence and arousal (e.g., those with clinical elevations in anxiety or obsessive–compulsive disorder) tend to show increased ERN amplitudes relative to individuals with more neutral valence and arousal levels (e.g., Aarts and Pourtois, 2010, Endrass et al., 2008, Gehring et al., 2000, Olvet and Hajcak, 2008). The literature on affective traits and performance monitoring is more variable for individuals with chronic levels of negative valence and low arousal (e.g., those with chronic sadness or depression), but seems to suggest slightly increased ERN amplitudes (or error minus correct differences) relative to psychiatrically-healthy controls (Georgiadi et al., 2011, Olvet and Hajcak, 2008, Weinberg et al., 2010). The research on positive valence (regardless of arousal) is still in its infancy and does not provide clear suggestions, with some studies showing an increased ERN amplitudes associated with increased levels of happiness and calmness (West and Travers, 2008), whereas others show decreased (i.e., more positive) ERN amplitudes in individuals with higher satisfaction with life and no relationship with other positive personality traits including positive affect and optimism (Larson et al., 2010c).

Based on this literature and the state mood induction literature to date, we hypothesized that ERN amplitude would increase during a negative-valence, high-arousal mood state, namely, anxiety. We also hypothesized that ERN amplitude would increase during a negative-valence, low-arousal mood state, namely, sadness. We did not have specific hypotheses about the CRN or Pe based on a relative lack of information in the current literature.

Section snippets

Participants

All participants provided written informed consent as approved by the local Institutional Review Board. Participants were recruited from undergraduate psychology courses. Exclusion criteria, assessed via participant self-report, included current or previous diagnosis of a psychiatric disorder, current psychoactive medication use, current substance abuse or dependence, neurological disorders, head injury, left-handedness, uncorrected visual impairment. These exclusion criteria were chosen to

Mood induction manipulation check

We first examined whether the mood induction procedure induced the intended mood states. See Table 1 for mean arousal and valence ratings for each group across the six ratings. Consistent with previous studies using the same mood induction process (e.g., Jefferies et al., 2008), participants at baseline were in a generally positive valence (mean valence = 2.2 [1.5]) with neutral arousal (mean arousal = .9 [1.8]) prior to beginning the mood induction procedure.

A 4-Group × 6-Time repeated measures

Discussion

We aimed to examine the influences of affective valence and arousal on electrophysiological and behavioral indices of response-related performance monitoring (i.e., ERN, CRN, Pe) using a standard mood induction paradigm. To address some of the inconsistencies in the literature we ensured that the mood induction procedure altered participant self-reported mood and specifically examined both the interactive (i.e., group-based) and orthogonal roles of valence and arousal. Participant ratings of

Acknowledgment

We gratefully acknowledge the assistance of Justin Hoskin, Kevin Voisin, and Jordan Davis in data collection. The Brigham Young University Office of Research and Creative Activities provided funding for this study. The authors report no conflicts of interest.

References (91)

  • M.J. Herrmann et al.

    Source localization (LORETA) of the error-related-negativity (ERN/Ne) and positivity (Pe)

    Brain Research. Cognitive Brain Research

    (2004)
  • G. Hughes et al.

    Dissociable correlates of response conflict and error awareness in error-related brain activity

    Neuropsychologia

    (2011)
  • M.J. Larson et al.

    The relationship between performance monitoring, satisfaction with life, and positive personality traits

    Biological Psychology

    (2010)
  • M.J. Larson et al.

    Performance monitoring following conflict: internal adjustments in cognitive control?

    Neuropsychologia

    (2012)
  • D.M. Olvet et al.

    The error-related negativity (ERN) and psychopathology: toward an endophenotype

    Clinical Psychology Review

    (2008)
  • D.M. Olvet et al.

    Depression symptom severity and error-related brain activity

    Psychiatry Research

    (2010)
  • P.E. Pailing et al.

    The effects of uncertainty in error monitoring on associated ERPs

    Brain and Cognition

    (2004)
  • M. Pietschmann et al.

    Aging, probabilistic learning and performance monitoring

    Biological Psychology

    (2011)
  • C. Roger et al.

    Rostral Cingulate Zone and correct response monitoring: ICA and source localization evidences for the unicity of correct- and error-negativities

    NeuroImage

    (2010)
  • D. Schrijvers et al.

    Action monitoring in major depressive disorder with psychomotor retardation

    Cortex

    (2008)
  • B. Stemmer et al.

    Error detection in patients with lesions to the medial prefrontal cortex: an ERP study

    Neuropsychologia

    (2004)
  • M. Tops et al.

    Task engagement and the relationships between the error-related negativity, agreeableness, behavioral shame proneness and cortisol

    Psychoneuroendocrinology

    (2006)
  • V. van Veen et al.

    The anterior cingulate as a conflict monitor: fMRI and ERP studies

    Physiology and Behavior

    (2002)
  • F. Vidal et al.

    Is the ‘error negativity’ specific to errors?

    Biological Psychology

    (2000)
  • A. Weinberg et al.

    Increased error-related brain activity in generalized anxiety disorder

    Biological Psychology

    (2010)
  • D. Wiswede et al.

    Modulation of the error-related negativity by induction of short-term negative affect

    Neuropsychologia

    (2009)
  • K. Aarts et al.

    Anxiety not only increases, but also alters early error-monitoring functions

    Cognitive, Affective, & Behavioral Neuroscience

    (2010)
  • L.F. Barrett et al.

    Independence and bipolarity in the structure of current affect

    Journal of Personality and Social Psychology

    (1998)
  • B.D. Bartholow et al.

    Strategic control and medial frontal negativity: beyond errors and response conflict

    Psychophysiology

    (2005)
  • M.A. Boksem et al.

    Facing disapproval: performance monitoring in a social context

    Social Neuroscience

    (2011)
  • M. Brazdil et al.

    Intracerebral error-related negativity in a simple go/no-go task

    Journal of Psychophysiology

    (2005)
  • Bruyneel, L., van Steenbergen, H., Hommel, B., Band, G.P., De Raedt, R., Koster, E.H.W., in press. Happy but still...
  • C.S. Carter et al.

    Anterior cingulate cortex, error detection, and the online monitoring of performance

    Science

    (1998)
  • L.G. Chepenik et al.

    The influence of sad mood on cognition

    Emotion

    (2007)
  • P.H. Chiu et al.

    Neural evidence for enhanced error detection in major depressive disorder

    The American Journal of Psychiatry

    (2007)
  • P.E. Clayson et al.

    The effects of induced state negative affect on performance monitoring processes

    Social Cognitive and Affective Neuroscience

    (2012)
  • J. Cohen et al.

    Applied multiple regression/correlation analysis for the behavioral sciences

    (2003)
  • R.J. Compton et al.

    Error detection and posterror behavior in depressed undergraduates

    Emotion

    (2008)
  • E. Eich et al.

    Combining music with thought to change mood

  • T. Endrass et al.

    ERP correlates of conscious error recognition: aware and unaware errors in an antisaccade task

    European Journal of Neuroscience

    (2007)
  • T. Endrass et al.

    Response-related negativities following correct and incorrect responses: evidence from a temporospatial principal component analysis

    Psychophysiology

    (2012)
  • T. Endrass et al.

    Temporospatial dissociation of Pe subcomponents for perceived and unperceived errors

    Frontiers in Human Neuroscience

    (2012)
  • B.A. Eriksen et al.

    Effects of noise letters upon the identification of a target letter in a non-search task

    Perception & Psychophysics

    (1974)
  • J.M. Ford

    Schizophrenia: the broken P300 and beyond

    Psychophysiology

    (1999)
  • D. Foti et al.

    Differentiating neural responses to emotional pictures: evidence from temporal-spatial PCA

    Psychophysiology

    (2009)
  • Cited by (28)

    • Coaching older adults discharged home from the emergency department: The role of competence and emotion in following up with outpatient clinicians

      2022, Patient Education and Counseling
      Citation Excerpt :

      In other words, not only do emotions affect intrinsic motivation to engage in a pleasant task, but they make people continue to work longer on less pleasant tasks and even self-regulate behaviors to complete an uninteresting task that has to be done in a responsible manner [20]. This study focused on four emotion exemplars selected from a list of well-studied discrete emotions [24], operationalizing them in terms of valence and arousal level as consistent with prior research [25–27]: happiness (positive valence, high arousal), calm (positive valence, low arousal), anxiety (negative valence, high arousal), and sadness (negative valence, low arousal). Emotions with positive valence have been found to increase active engagement with positive goals that, in turn, lead to desirable outcomes [28], while negative emotions might decrease the likelihood of engaging in a target behavior [29].

    • Intense positive affect without arousal is possible: Subjective and physiological reactivity during a partnered sexual meditative experience

      2022, International Journal of Psychophysiology
      Citation Excerpt :

      Other studies find more evidence of dependence with increases in experimentally induced happiness also yielding increased arousal, which was not intentionally manipulated (Russell and Alden, 1987 - uncited, as reported in (Russell et al., 1989)). The potential orthogonality of valence and arousal is more evident in studies showing they can have independent influences on objective third variables, such as event-related potentials (Larson et al., 2013), and in studies which use more “real-world” stimuli, such as naturalistic variation in affect and arousal across days (McFarlane et al., 1988). Thus, to document experimentally-induced variation in affect and arousal, we appealed to methods that moved 1) beyond simple self-report of affect and 2) which use probes likely to induce more intense affective states than pictures.

    View all citing articles on Scopus
    View full text