Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Jul 1.
Published in final edited form as: Int J Psychophysiol. 2020 Apr 21;153:45–52. doi: 10.1016/j.ijpsycho.2020.04.015

Intolerance of Uncertainty, Depression and the Error-Related Negativity

Jared R Ruchensky 1, Elizabeth A Bauer 1, Annmarie MacNamara 1
PMCID: PMC7934182  NIHMSID: NIHMS1647210  PMID: 32330538

Abstract

Prior work has yielded contradictory findings regarding the association between depression and the error-related negativity (ERN), an event-related potential thought to reflect individual variation in sensitivity to internal threat (i.e., errors), and the error positivity (Pe), which is thought to reflect more elaborative, conscious processing of errors. One possibility is that variation in transdiagnostic dimensional constructs related to threat processing might help explain inconsistencies in the relationship between depression and the ERN and Pe. Here, we used a large, unselected sample (N = 100) to determine whether variation in intolerance of uncertainty (IU), a transdiagnostic trait dimension that is implicated in both depression and anxiety, might moderate associations between depressive symptomatology and error processing. Results showed that greater levels of depressive symptomatology were associated with larger ΔERNs (error minus correct trials) when IU was low, but were unrelated to ΔERN when IU was high; main effects of depression and IU on ΔERN were not observed. The Pe was not associated with IU or depression. Additionally, IU, but not depression, was associated with faster response times on error and correct trials. Overall, results suggest that error processing may differ for individuals with elevated depression with versus without elevated IU. Moreover, prior failures to observe associations between depression and the ERN might stem from failure to account for related transdiagnostic constructs.

Keywords: ERN, ERP, IUS, event-related potential, flankers, error positivity

1. Introduction

Depression is an impairing disorder that adversely affects functioning even at subthreshold levels (Fergusson, Horwood, & Ridder, 2005). Its clinical presentation varies, and symptoms frequently co-occur with those of other internalizing disorders. Therefore, identification of (1) neurobiological measures that distinguish between depression and its comorbidities (Weinberg, Liu, & Shankman, 2016) and (2) dimensional constructs that cut across depression to explain heterogeneity (Gentes & Ruscio, 2011; Watson, 2005) could lead to a more individualized understanding of depression and its treatment. Prior research has investigated the neurobiological processing of internal threats, such as error commission, as they relate to depression (Olvet & Hajcak, 2008; Weinberg et al., 2016b), but results have been inconsistent. Consideration of transdiagnostic dimensions underpinning depression and its comorbidities may help clarify associations between depression and abnormalities in internal threat processing (Moran, Schroder, Kneip, & Moser, 2017).

Like external threats (e.g., predators), errors are threatening because they can jeopardize wellbeing (e.g., by resulting in injury). One way of measuring individual differences in the significance of error processing/internal threat is using event-related potentials (ERPs), the error-related negativity (ERN) and the error positivity (Pe). The ERN (alternatively referred to as the error negativity [Ne; Gehring, Liu, Orr, & Carp, 2012]) is a frontocentrally maximal, negative-going deflection in the electroencephalographic waveform that peaks between 0–100 ms following response. The Pe is a positive-going component that is largest at central scalp sites and is evident between 200–400 ms after error commission. The ERN and the Pe are larger for error than correct responses. Whereas the ERN is believed to measure the early stages of error-processing, the Pe is thought to reflect the more elaborated, conscious response to errors (Overbeek, Nieuwenhuis, & Ridderinkhof, 2005). Internalizing disorders are thought to be characterized by aberrant threat responding, and in line with this notion, the ERN (and to a lesser extent, the Pe) has been linked to anxiety and depression.

For example, undergraduates with greater self-reported anxiety exhibit larger ERNs (Moser, Moran, Schroder, Donnellan, & Yeung, 2013; Zambrano-Vazquez & Allen, 2014). Additionally, participants diagnosed with internalizing disorders, such as generalized anxiety disorder (GAD; Cavanagh, Meyer, & Hajcak, 2017; Weinberg, Klein, & Hajcak, 2012), and social anxiety disorder (Endrass, Riesel, Kathamnn, & Buhlmann, 2014) have larger ERNs compared to healthy controls. Furthermore, the ERN has been shown to predict the development of anxiety (Meyer, Hajcak-Proudfit, Torpey-Newman, Kujawa, & Klein, 2015). Therefore, the ERN may be both a correlate of anxiety and a biomarker of anxiety risk, although there is debate regarding the particular anxiety phenotype associated with an enhanced ERN (Cavanagh & Shackman, 2015; Riesel, Goldhahn, & Kathmann, 2017; Zambrano-Vazquez & Allen, 2014). Comparatively few studies have focused on the Pe, but it too has been associated with anxiety (Moser, Moran, & Jendrusina, 2012). However, unlike the ERN, the direction of the association between the Pe and anxiety is less well-established. For example, reduced Pe amplitudes have been exhibited in female undergraduates with greater self-reported worry (Moser et al., 2012), and in obsessive-compulsive disorder patients compared to healthy controls (Agam et al., 2014). On the other hand, however, Pe amplitudes were found to be larger for participants with GAD compared to healthy controls (Weinberg, Olvet, & Hajcak, 2010).

In contrast to the well-established anxiety-ERN relationship (Cavanagh & Shackman, 2015), it is unclear how the ERN and Pe relate to depression. Research has found that depression is associated with smaller (Weinberg et al., 2012; Weinberg et al., 2016a) and larger ERNs (Chiu & Deldin, 2007; Holmes & Pizzagalli, 2008). In addition, recent meta-analytic work found no overall association between depression and the ERN (Moran et al., 2017), suggesting instead that results may be complex, and could depend on co-occurring disorders or traits. For example, Weinberg and colleagues found that the ERN was enhanced in individuals with “pure” GAD but not in those with comorbid depression, indicating that depression might suppress the relationship between anxiety and the ERN (Weinberg, Kotov, & Proudfit, 2015). Still, the particular type of comorbid symptomatology may influence results: Hanna and colleagues (2018) found potentiation of the ERN for children and adolescents with obsessive compulsive disorder, both with and without comorbid depression, whereas depression alone did not modulate the ERN. Furthermore, certain types of depression but not others might be associated with altered ERNs: one study found that remitted melancholic depressed individuals showed smaller ERNs than controls, whereas remitted non-melancholic depressed individuals did not (Weinberg, et al., 2016a).

In addition to comorbid psychopathology or depressive subtypes, severity of depression might influence results. For instance, Schrijvers and colleagues (2009) found no evidence for an enhanced ERN in a sample of depressed inpatients and attributed this to higher levels of depression severity in their sample. By contrast, mild to moderate levels of depression have been found to be associated with an increased ERN in other work (e.g., Chiu & Deldin, 2007; Holmes & Pizzagalli, 2008). However, these differences in results might also be attributed to variation in co-occurring, transdiagnostic trait dimensions. That is, because Schrijvers and colleagues’ (2009) sample was substantially more depressed, high levels of certain depressive features, such as anhedonia and apathy, might have attenuated the potentiating effect of affective distress on the ERN, for example, by reducing motivation and/or engagement with the task. In sum, there is a lack of consensus on how depression relates to the ERN, and evidence suggests that this relationship may be qualified by the presence of severity and/or other internalizing constructs. Far fewer studies have examined the Pe in relation to depression; however, results for this component are also mixed: some research has found no evidence of an association between the Pe and depression (Chiu & Deldin, 2007), whereas other work has found a negative association (Olvet, Klein, & Hajcak, 2010).

One promising trait for understanding internalizing and externalizing psychopathology is intolerance of uncertainty (IU), which is a dispositional construct that reflects the extent to which ambiguous situations are perceived as threatening (Freeston, Rhéaume, Letarte, Dugas, & Ladouceur, 1994; Gentes & Ruscio, 2011). IU is present in unselected samples and relates to the development and maintenance of internalizing disorders (Boswell, Thompson-Hollands, Farchione, & Barlow, 2013; Gentes & Ruscio, 2011). IU has been associated with the magnitude of the ERN, though the direction of effects has been found to vary across subscales (Jackson, Nelson, & Hajcak, 2016). In other work, IU was found to mediate the relationship between depression and frontal EEG asymmetry (i.e., reduced reward sensitivity; Nelson, Shankman, & Proudfit, 2014). Therefore, examination of IU might be fruitful in explaining individual differences in the ERN in depression.

In the current study, we set out to determine how IU and depressive symptomatology might relate to the ERN and the Pe in a large, unselected sample of undergraduates. Given evidence that IU may relate more strongly to anxiety than depression (Jensen, Cohen, Mennin, Fresco, & Heimberg, 2016), and that anxiety has been more consistently associated with enhanced ERNs, we hypothesized that IU would relate to larger ERNs, whereas there would be no main effect of depression on the ERN (Hanna et al., 2018; Moran et al., 2017; Weinberg et al., 2015). However, we did hypothesize that depressive symptomatology would be associated with smaller (i.e., more positive) ERNs (Weinberg et al., 2012; Weinberg et al., 2016a), but only among participants with low IU, because participants with high IU would have larger ERNs, regardless of depression (Hanna et al., 2018; but see Weinberg et al., 2015). Given limited research on the Pe, we did not have a directional prediction for an association with IU (Jackson et al., 2016; Tanovic, Gee, & Joorman, 2018) or depression (Olvet et al., 2010).

2. Materials and Methods

2.1. Participants

Participants were 100 undergraduates (57% female) who completed the study for course credit (Mage=18.89 years; SD=2.64). Study procedures were in compliance with the Helsinki Declaration of 1975 (as revised in 1983) and approved by the Texas A&M University institutional review board.

2.2. Self-report measures

After consenting to the study procedures, participants completed the following self-report measures:

The Intolerance of Uncertainty Scale (IUS; short form) is a 12-item self-report measure (derived from the original 27-item version, Freeston et al., 1994) that assesses responses to uncertainty, ambiguous situations, and the future (Carleton, Norton, & Asmundson, 2007). Participants make their responses using a 5-point Likert scale ranging from 1 (not at all characteristic of me) to 5 (entirely characteristic of me), with higher scores indicating greater intolerance of uncertainty. Scores in the current study ranged from 13 to 46 (M=29.35; SD=7.98).

The Beck Depression Inventory – II (BDI-II; Beck, Steer, & Brown, 1996) is a widely-used self-report measure of depressive symptoms. Participants rate each of the 21 items on a 4-point scale between 0 and 3, with higher total scores indicating greater depression. Scores in the current study ranged from 1 to 46 (M=11.68; SD=8.51). Scores for 67 participants fell within the “no depression” range (0—13); scores for 17 participants fell within the “mild depression” range (14—19), and scores for 16 participants fell within the “moderate to severe depression” range (20—63; Beck et al., 1996).

2.3. Flanker task

Participants completed an arrow version of the Eriksen flanker task (Eriksen & Eriksen, 1974), which was displayed using Presentation software (Neurobehavioral Systems Inc., Albany, CA). On each trial, participants viewed five white arrows presented for 200 ms against a black background; they were asked to indicate the direction of the middle arrow within 1800 ms by pressing the left or right mouse button. Half of the trials were congruent (< < < < < or > > > > >), whereas the other half were incongruent (e.g., < < > < < or > > < > >). Before beginning the task, participants completed a practice block containing 10 trials. Trial order was random for each participant and participants completed 11 blocks of 30 trials with self-paced breaks in between each block (330 trials in total for each participant). A white fixation cross was presented following each trial for 600–1000 ms. Reaction time was recorded as time in ms between stimulus onset and participant response.

2.4. EEG recording and processing

Continuous EEG was recorded using an ActiCap and the ActiChamp amplifier system (Brain Products, Gilching Germany). Thirty-two electrode sites were used based on the 10/20 system. The electrooculogram (EOG) was recorded from four facial electrodes: two electrodes were placed approximately 1 cm above and below the right eye, forming a bipolar channel to measure vertical eye movement and blinks and two electrodes were placed approximately 1 cm beyond the outer edges of each eye, forming a bipolar channel to measure horizontal eye movements. EEG data was digitized at a 24-bit resolution with a sampling rate of 1000 Hz.

Offline, data were processed using BrainVision Analyzer 2 software (Brain Products GmbH, Gilching, Germany). The signal from each electrode was referenced offline to the average of the left and right mastoids (TP9/10) and band-pass filtered with high-pass and low-pass filters of .1 and 30 Hz, respectively. For the ERN, CRN, Pe, and correct positivity (Pc), data were segmented for each trial beginning 500 ms prior to response and lasting for 1500 ms (1000 ms beyond response). Eye blink and ocular correction used the method developed by Miller, Gratton and Yee (1988). Artifact analysis was used to identify a voltage step of more than 50.0 μV between sample points, a voltage difference of 300.0 μV within a trial, and a maximum voltage difference of less than 0.50 μV within 100 ms intervals. Trials were also inspected visually for any remaining artifacts, and data from individual channels containing artifacts were rejected on a trial-to-trial basis. The average numbers of trials rejected per participant and condition were as follows: ERN=4.33% (SD=9.10), CRN=1.18% (SD=2.71), Pe=3.90% (SD=8.92), Pc=1.04% (SD=2.42). Importantly, the percent of trials rejected was not associated with our individual difference measures (i.e., depression and IU; all ps>.37).

2.5. Data Analyses

For response-locked ERPs, correct and error trials were averaged separately and baseline correction was performed using a 200 ms window from −500 to −300 ms before response onset. The ERN and correct response negativity, CRN, were scored as the average activity on error and correct trials, respectively, from 0 to 100 ms after response at a pooling of FC1, FC2, Cz and Fz (Jackson et al., 2016). The Pe and Pc were scored as the average activity on error and correct trials, respectively, from 200 to 400 ms after response at a pooling of Cz and Pz (Falkenstein, Willemssen, Hohnsbein, & Hielscher, 2005). Although the main set of analyses focused on the ERN and CRN collapsed across both congruent and incongruent trials, it is possible that congruency affects the amplitude of ΔERN (Yeung, Botvinick, & Cohen, 2004). Therefore, we present results for the ERN and CRN on incongruent trials only in the Supplementary Material.

To examine the unique and interactive associations between IU and depressive symptoms and each ERP and behavioral data, we conducted multiple regression analyses with scores for depression, IU and the interaction term (depression X IU) entered simultaneously. Consistent with standardization guidelines (Cohen, Cohen, West & Aiken, 2003), individual predictors were first Z-scored and the interaction term was created by multiplying these standardized scores. Separate regressions were conducted for each ERP and behavioral data. All analyses were performed using SPSS statistical software version 24.0 (IBM, Armonk, NY). Split-half reliability was also calculated for all dependent variables by correlating even-odd trials and using the Spearman-Brown correction formula (Helmstadetr, 1964).

3.1. Results

Table 1 presents means and standard deviations for the ERN, the Pe and behavioral data. The ERN was more negative than the CRN, F(1,99)=123.40, p<.001, ηp2=.56. Similarly, the Pe was significantly more positive than the Pc, F(1,99)=282.65, p<.001, ηp2=.74. Subsequent analyses used ΔERN and ΔPe, which were created by subtracting correct from error trials. Internal consistency was acceptable to excellent for the ERN (r=.76), CRN (r=.99), and ΔERN (r=.62). The Pe (r=.89), Pc (r=.98), ΔPe (r=.88), accuracy (r=.97) and RT for correct (r=.99) and incorrect (r=.85) trials also had excellent reliability.

Table 1.

Means and standard deviations for ERPs and RT.

Error Correct Δ (Error – Correct)
ERN (μV) Pe (μV) RT (ms) CRN (μV) Pc (μV) RT (ms) ΔERN (μV) ΔPe (μV)
M 1.12 14.36 314.60 7.13 3.36 394.11 −6.01 11.00
SD 6.64 6.61 55.55 6.88 4.24 51.65 5.41 6.54
Open in a new tab

Note: ERN = error-related negativity; Pe = error positivity; RT = reaction time; CRN = correct-related negativity; Pc = correct positivity.

ΔERN

Figure 1 depicts grand average waveforms for error and correct trials, difference waveforms corresponding to error minus correct trials and scalp topographies for error minus correct trials during the time window in which the ERN was scored. Figure 2 presents mean ΔERN values graphed separately for participants with low, medium and high levels of IU and for participants with low and high levels of depression. Figure 3 depicts waveforms and scalp topographies for error minus correct trials shown separately for participants with low, medium and high levels of IU, at low and high levels of depression. Of note, we used continuous IU and depressive symptomatology scores for all analyses; participants were divided into groups for illustrative purposes only (see Table 1 in Supplementary Material for descriptive statistics of groupings).

Figure 1.

Figure 1.

Open in a new tab

Grand-average waveforms at the fronto-central pooling (FC1, FC2, Fz, and Cz) where the ERN and CRN were scored, and scalp distributions of the voltage differences for error minus correct trials between 0 and 100 ms after response.

Figure 2.

Figure 2.

Open in a new tab

The relationship between depression and the ΔERN at low (solid line, −1 SD), medium (dotted line, mean) or high (dashed line, +1 SD) levels of IU. Error bars represent standard error of the mean.

Figure 3.

Figure 3.

Open in a new tab

Response-locked grand-averaged waveforms at the fronto-central pooling (FC1, FC2, Fz, and Cz) where the ERN and CRN were scored and scalp topographies depicting the error minus correct difference (0–100 ms), shown separately for participants low (left) and high (right) in depression, as well as for participants with low IU (top), medium IU (middle) and high IU (bottom). Groupings were based on a median split and terciles, respectively, and are for illustrative purposes only.

Table 2 presents results for regression analyses involving ΔERN. There were no significant results when entering depression into separate regressions predicting ΔERN. There were also no significant results when depression and IU were entered into the regression simultaneously. When depression, IU and the interaction term (depression X IU) were entered into the model simultaneously, depression and IU interacted to predict ΔERN (β=.23, p=.04), with the overall model itself trending towards significance [R2=.08, F(3,96)=2.61, p=.06]. To examine the nature of the depression X IU interaction, we assessed the association between depression symptomatology and the ΔERN separately at low (−1 SD), average (mean), and high (+1 SD) levels of IU. Simple slope analyses revealed that higher levels of depression were significantly associated with larger ΔERNs at low, t(96)=2.06, p=.04, but not medium, t(96)=1.53, p=.13, or high, t(96)=.08, p=.93, levels of IU. Parallel regression results conducted with the ERN and CRN separately are presented in the Supplementary Material1,2.

Table 2.

Standardized beta coefficients (β) for ERP regression analyses

ΔERN
Predictors I II III
Depression −.13 −.07 −.18
IU −.18 −.15 −.14
Depression X IU - - .23*
 
ΔPe
Predictors I II III
Depression .07 .08 .09
IU .01 −.03 −.03
Depression X IU - - −.02
Open in a new tab

Note: ERN = error-related negativity; IU = intolerance of uncertainty; Pe = error positivity.

*

p < .05. Regression analyses for first model (I) involved entering independent variables into separate regressions. Regression analyses for second model (II) involved entering independent variables simultaneously. Regression analyses for third model (III) involved entering individual predictors (depression, IU) and the interaction term simultaneously.

Prior work has found that the ΔERN may differ in magnitude for congruent versus incongruent trials (Yeung et al., 2004). As such, it seemed possible that associations between our individual difference variables and ΔERN might be driven in part by an association between these variables and different proportions of congruent and incongruent trials in error and correct trial waveforms. Therefore, we calculated the difference between the relative proportion of incongruent (versus total) error trials and the relative proportion of incongruent (versus total) correct trials: (#incongruent errors/#total errors) – (#incongruent correct/#total correct) and checked to see if this measure was associated with our individual difference variables. This metric did not correlate significantly with depression (r=.15, p=.15) or IU (r=.14, p=.17). Additionally, when we entered this metric into our final regression equation we found that the results were in line with those originally reported, such that neither depression (β=−.15, p=.21) nor IU (β=− .12, p=.26) significantly predicted ΔERN but the interaction of depression and IU predicted ΔERN, albeit at trend level (β=.21, p=.06).

ΔPe

Table 2 presents results for regression analyses involving ΔPe. There were no significant results when IU and depression were entered into separate regressions nor when they were entered simultaneously. Results were also not significant when individual predictors (depression, IU) and the interaction term (depression X IU) were entered simultaneously.

Performance

Table 3 presents results for regression analyses involving performance DVs. Participants were more accurate for congruent (M=93.01%, SD=0.11) versus incongruent (M=78.97%, SD=0.12) trials, t(87)=15.82, p<.001. Overall, participants performed well on the task (accuracy: M=87.35, SD=10.47). For accuracy, results were not significant when depression and IU were entered into separate regressions. Similarly, results were not significant when both IU and depression were entered simultaneously, or when individual predictors and the interaction term (depression X IU) were entered simultaneously.

Table 3.

Standardized beta coefficients (β) for behavioral regression analyses

Accuracy
Predictors I II III
Depression −.04 .01 .04
IU −.12 −.13 −.13
Depression X IU - - −.06
 
Correct Trials: RT
Predictors I II III
Depression −.13 .01 −.04
IU −.35** −.35** −.35**
Depression X IU - - .10
Incorrect Trials: RT
Predictors I II III
Depression −.08 .02 .02
IU −.26* −.26* −.26*
Depression X IU - - .02
Open in a new tab

Note: IU = intolerance of uncertainty; RT = reaction time.

*

p < .05.

**

p < .01. Regression analyses for first model (I) involved entering independent variables into separate regressions. Regression analyses for second model (II) involved entering independent variables simultaneously. Regression analyses for third model (III) involved entering individual predictors (depression, IU) and the interaction term simultaneously.

Consistent with prior findings (Moser et al., 2013), participants responded faster on error versus correct trials, F(1,99)=291.20, p<.001, ηp2=.75. Participants also responded faster on congruent (M=363.85, SD=46.88) versus incongruent (M=418.12, SD=70.61) correct trials, t(99)=8.98, p<.001. They also responded faster on congruent (M=290.24, SD=66.41) versus incongruent (M=310.83, SD=49.02) error trials, t(87)=3.18, p<.01. When depression scores alone were entered as a predictor of RT, results were not significant for correct RT or incorrect RT. IU alone predicted correct (β=−.35, p<.001) and incorrect RT (β=−.26, p=.01). Next, we entered both IU and depression into the model simultaneously; results for IU, but not depression, were significant for correct (β=−.35, p=.001) and incorrect (β=−.26, p=.02) trials. Associations with IU also held when the interaction term (depression X IU) was included as a predictor for correct (β=−.35, p=.001) and incorrect (β=−.26, p=.02) trials. The first two regression models reached significance [I: R2=.07, F(1,98)=6.80, p=.01; II: R2=.07, F(2,97)=3.39, p=.04], but the last model did not [R2=.07, F(3,96)=2.24, p=.09]. Together, RT results suggest that higher IU is uniquely associated with faster RT on both correct and incorrect trials. Additional results for post-error slowing and post-error accuracy are presented in the Supplementary Material3,4.

4. Discussion

To help clarify mixed findings regarding the ERN-depression relationship, we used a large, unselected sample to examine how IU might moderate the association between depression and the ERN. Results showed that depression and IU interacted to predict variation in ΔERN. At low levels of IU, greater depression was associated with larger ΔERN amplitudes, whereas at high levels of IU, depression was unrelated to ΔERN. On their own, neither depression nor IU were significantly associated with ΔERN. In terms of behavioral data, IU emerged as a significant predictor of RT on correct and incorrect trials, suggesting that those with high IU respond more quickly. In contrast, there were no significant associations between any of the predictors with accuracy or ΔPe.

The finding that depression relates to an enhanced ΔERN when IU is low is broadly in line with work that has examined processing of external threats in depression. For instance, depressed individuals have been found to exhibit larger startle eyeblink when anticipating shock (Grillon, Franco-Chaves, Mateus, Ionescu, & Zarate, 2013). Similarly, there is evidence that depressed individuals display greater neural reactivity to standardized emotional stimuli using imaging techniques (e.g., greater amygdala activity; Hamilton et al., 2012). Whereas this work has primarily used standardized affective stimuli such as pictures of threatening faces, evidence also suggests that more personalized/idiographic stimuli may be especially likely to elicit increased processing in depression (Speed, Nelson, Auerbach, Klein, & Hajcak, 2016). As direct, personal sources of threat, electric shock and errors might therefore be particularly likely to attract increased processing in depression.

Nonetheless, depression was only associated with increased error processing for participants who were also low in IU. Because depression tends to covary with other internalizing constructs and symptom scores, the potentiating effects of depression on the ERN might not be observed without considering moderation by other constructs (like IU). Therefore, greater depression or IU (see Supplementary Material) relates to a larger ERN only when the other is low, at least in our unselected sample. Of note, prior work has found that depression is associated with a reduced ERN, using clinical samples (Weinberg et al. 2016a) and also in childhood/adolescent samples (Meyer, Bress, Hajcak, & Gibb, 2018). Therefore, one possibility is that in an unselected sample, increased ERNs might serve as relatively non-specific correlates of latent psychopathology or psychopathology risk.

The significant results for the ERN, rather than the Pe, suggest that the interaction of depression and IU is specific to early threat processing (ERN), rather than more elaborative threat processing (Pe). Similarly, depression and IU did not interact to predict accuracy. Therefore, although depression (in the context of low IU) was associated with alterations in early error processing, it was not related to prolonged processing of errors or adaptive behavior/improved performance. As an early evaluative response that may signal the need for behavioral adjustments, the ERN is proposed to trigger resultant “downstream” processes such as activation of the dorsolateral prefrontal cortex. Nonetheless, these processes may or may not lead to improved behavior (Weinberg et al., 2016a). As such, future work may wish to examine functional connectivity between the anterior cingulate cortex, which is integral to the genesis of the ERN (Yeung et al., 2004), and brain regions more directly involved in behavioral adjustments, in order to increase knowledge surrounding depression and its association with error processing and behavioral response.

In contrast to depression, IU was associated with altered behavior. That is, participants with heightened IU responded more quickly across both error and correct trials, but nonetheless failed to show a decrement in accuracy. Therefore, participants with higher IU may have been more motivated to perform well on the task and/or may have engaged compensatory resources (e.g., increased attention) to maintain both fast responses and accuracy on par with that of participants with lower IU, who responded more slowly. Interestingly, emerging evidence suggests that when task instructions emphasize speed rather than accuracy, patients with obsessive compulsive disorder – a patient group known to exhibit increased IU (Gentes & Ruscio, 2011) - show increased ERNs, whereas this difference is obscured if participants are asked to focus only on responding accurately (Riesel, in press). Here, we emphasized both speedy responding and accuracy and observed a trend-level association between IU and larger ERNs, in line with the notion that participants with higher IU were responsive to task instructions emphasizing speed.

Our failure to find a main effect of depression on ΔERN is consistent with results from a recent meta-analysis, which found weak evidence for an association between depression and ΔERN (Moran et al., 2017). Overall, the depression-ERN literature is mixed and null results are common. In contrast to the large, albeit inconclusive, literature on depression and the ERN, only one other study has examined - and failed to find evidence of - a relationship between IU and the ERN (Jackson et al., 2016). Although IU is considered a transdiagnostic construct that underpins internalizing disorders (Gentes & Ruscio, 2011), it may only relate to enhanced processing of internal threats when accounting for depression. That is, an alternative interpretation of the interaction observed here would be that depression moderates the IU-ERN relationship such that high IU relates to larger ERNs only when depression is low (see Supplementary Material). This interpretation is in line with prior work that found evidence of larger ERNs in GAD, a disorder associated with elevated levels of IU (Dugas, Buhr, & Ladouceur, 2004; Weinberg et al., 2012). Specifically, the ERN was larger in participants with “pure” GAD, but not those diagnosed with both GAD and MDD (Weinberg et al., 2012). Taken together, results suggest that depression may obscure the association between anxiety and related constructs (IU) with the ERN.

Overall, there is a need for a more nuanced approach to understanding the depression–ERN link (Moran et al., 2017). Prior research may have yielded mixed results in part because of a failure to account for trait dimensions that commonly occur with depressive features. The results of the current study highlight the need to consider the contribution of frequently co-occurring constructs when examining the depression-ERN relationship. Nonetheless, it is worth noting that our overall model was not significant. This indicates that variation in the ERN is also due to additional factors (not examined here), reflecting the likely complex network of biological and psychological factors that contribute to neural response to error commission. As such, future work might benefit from attention to other co-occurring traits that could help explain additional heterogeneity in the ERN. For example, consideration of other dimensional constructs, such as personality, may further clarify under what conditions the ERN is altered (Hill, Samuel, & Foti, 2016). Further, future work may wish to determine whether the current results hold in more severely depressed samples (Schrijvers et al., 2009), as the current study relied on an undergraduate sample with a restricted range of depressive symptoms. Additionally, future work may wish to consider the flanker congruency effect when reporting on the role of individual difference variables in the ΔERN, given that the association between ΔERN and depression X IU did not reach significance when we restricted ΔERN analyses to incongruent trials.

4.1. Conclusions

In sum, results indicate that depression is characterized by increased reactivity to errors at low levels of IU (and vice versa). As such, the ERN might represent a relatively non-specific correlate of affective psychopathology/risk for psychopathology in unselected samples. Importantly, however, the presence of elevated scores on either depression or IU appears to attenuate associations with the other; as such, prior failures to observe associations between depression and the ERN might stem from failure to account for related transdiagnostic dimensions, such as IU.

Supplementary Material

1

Highlights.

  • Examined intolerance of uncertainty (IU), depression, and error processing (ERN)

  • Depressive symptomatology related to a larger ERN at low, but not high, IU

  • Depression or IU is associated with increased ERNs, but effects are not additive

  • ERN associations with depression, IU are attenuated at high levels of the other

  • The ERN might be a relatively non-specific biomarker in unselected samples

Acknowledgments

Annmarie MacNamara was supported by National Institute of Mental Health grant, K23MH105553 during preparation of this manuscript.

Footnotes

1.

No significant results for ΔERN were observed when entering subscales of IU in separate regressions. Specifically: for Prospective IU [depression with β=−.15, p=.20; Prospective IU with β=−.17, p=.11; interaction term with β=.19, p=.10] and Inhibitory IU [depression with β=−.18, p=.11; Inhibitory IU with β=−.09, p=.42; interaction term with β=.21, p=.07]. Similarly, no significant results were observed when including depression, IU subscales, and both interaction terms (depression X Inhibitory IU; depression X Prospective IU), in the same model. Specifically: Inhibitory IU with β=.06, p=.65; Prospective IU with β=−.21, p=.12; depression with β=−.18, p=.15; Inhibitory IU X depression with β=.09, p=.50; Prospective IU X depression with β=.13, p=.34.

2.

When controlling for both correct RT and incorrect RT, results for the final model predicting ΔERN were in line with those presented in the main text (R2=.13, F(5,94)=2.86, p=.02). Specifically: depression with β=−.17, p=.15; IU with β=−.05, p=.63; depression x IU with β=.20, p=.07; correct RT with β=.26, p=.04; incorrect RT with β=−.02, p=.90.

3.

To examine the potential effect of congruency on behavioral data, we conducted separate repeated measures ANOVAs for correct RT and accuracy with congruency as a within-subjects factor and depression, IU, and the depression X IU interaction as covariates of interest. There were no significant interactions between depression [F(1,96)=.71, p=.41, ηp2=.01], IU [F(1,96)=2.43, p=.12, ηp2=.03], or depression X IU [F(1,96)=1.95, p=.12, ηp2=.02] and congruency. There were no significant interactions between depression [F(1,84)=.34, p=.56, ηp2=.00], IU [F(1,84)=3.13, p=.08, ηp2=.04], or depression X IU [F(1,84)=.03, p=.87, ηp2=.00] and congruency. Therefore, the individual difference variables (depression, IU) and their interaction (depression X IU) do not appear to relate to accuracy and speed of RT across congruent versus incongruent trials.

4.

To examine potential interactions between congruency and the individual difference variables (depression, IU) on behavioral data, we conducted separate repeated measures ANOVAs for correct RT and accuracy with congruency as a within-subjects factor and depression, IU, and the depression X IU interaction as covariates of interest. Only correct RT and accuracy were examined in this way because of low trial counts for incongruent versus congruent error RT. For correct RT, there were no significant interactions between congruency and depression [F(1,96)=.71, p=.41, ηp2=.01], IU [F(1,96)=2.43, p=.12, ηp2=.03], or depression X IU [F(1,96)=1.95, p=.12, ηp2=.02]. Likewise, for accuracy, there were no significant interactions between congruency and depression [F(1,84)=.34, p=.56, ηp2=.00], IU [F(1,84)=3.13, p=.08, ηp2=.04], or depression X IU [F(1,84)=.03, p=.87, ηp2=.00]. Therefore, the individual difference variables (depression, IU) and their interaction (depression X IU) do not appear to relate to behavior on congruent versus incongruent trials.

Declaration of interest: none.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Agam Y, Greenberg JL, Isom M, Falkenstein MJ, Jenike E, Wilhelm S, & Manoach DS (2014). Aberrant error processing in relation to symptom severity in obsessive–compulsive disorder: A multimodal neuroimaging study. NeuroImage: Clinical, 5, 141–151. doi: 10.1016/j.nicl.2014.06.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Beck AT, Steer RA, & Brown GK (1996). Beck depression inventory-II. San Antonio, TX: Psychological Corporation. [Google Scholar]
  3. Boswell JF, Thompson-Hollands J, Farchione TJ, & Barlow DH (2013). Intolerance of uncertainty: A common factor in the treatment of emotional disorders. Journal of Clinical Psychology, 69, 630–645. doi: 10.1002/jclp.21965 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Carleton RN, Norton PJ, & Asmundson GJG (2007). Fearing the unknown: A short version of the intolerance of uncertainty scale. Journal of Anxiety Disorders, 21, 105–117. doi: 10.1016/j.janxdis.2006.03.014 [DOI] [PubMed] [Google Scholar]
  5. Cavanagh JF, Meyer A, & Hajcak G. (2017). Error-specific cognitive control alterations in generalized anxiety disorder. Biological Psychiatry: Cognitive Neuroscience and Neuroimaging, 2, 413–420. doi: 10.1016/j.bpsc.2017.01.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Cavanagh JF, & Shackman AJ (2015). Frontal midline theta reflects anxiety and cognitive control: Meta-analytic evidence. Journal of Physiology – Paris, 109, 3–15. doi: 10.1016/j.jphysparis.2014.04.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Cohen J, Cohen P, West SG, & Aiken LS (2003). Applied multiple regression/correlation analysis for the behavioral sciences (3rd ed.). Mahwah, NJ: Erlbaum. [Google Scholar]
  8. Chiu PH, & Deldin PJ (2007). Neural evidence for enhanced error detection in major depressive disorder. The American Journal of Psychiatry, 164, 608–616. [DOI] [PubMed] [Google Scholar]
  9. Dugas MJ, Buhr K, & Ladouceur R. (2004). The role of intolerance of uncertainty in etiology and maintenance. In Heimberg RG, Turk CL, & Mennin DS (Eds.), Generalized anxiety disorder: Advances in research and practice. New York: Guilford Press. [Google Scholar]
  10. Endrass T, Riesel A, Kathmann N, & Buhlmann U. (2014). Performance monitoring in obsessive-compulsive disorder and social anxiety disorder. Journal of Abnormal Psychology, 123, 705–714. doi: 10.1037/abn0000012 [DOI] [PubMed] [Google Scholar]
  11. Eriksen B, & Eriksen C. (1974). Effects of noise letters upon the identification of a target letter in a nonsearch task. Perception and Psychophysics, 16, 143–149. doi: 10.3758/BF03203267 [DOI] [Google Scholar]
  12. Falkenstein M, Willemssen R, Hohnsbein J, & Hielscher H. (2005). Error processing in Parkinson’s Disease. Journal of Psychophysiology, 19, 305–310. [Google Scholar]
  13. Fergusson DM, Horwood J, Ridder EM, & Beautrais AL (2005). Subthreshold depression in adolescence and mental health outcomes in adulthood. Archives of General Psychiatry, 62, 66–72. doi: 10.1001/archpsyc.62.1.66 [DOI] [PubMed] [Google Scholar]
  14. Freeston MH, Rhéaume J, Letarte H, Dugas MJ, & Ladouceur R. (1994). Why do people worry? Personality and Individual Differences, 17, 791–802. doi: 10.1016/0191-8869(94)90048-5 [DOI] [Google Scholar]
  15. Gehring WJ, Liu Y, Orr JM, & Carp J. (2012). The Error-Related Negativity (ERN/Ne). In Luck SJ & Kappenman ES (Eds.), The Oxford Handbook of Event-Related Potential Components (231–291). New York, NY: Oxford University Press, Inc. [Google Scholar]
  16. Gentes EL, & Ruscio AM (2011). A meta-analysis of the relation of intolerance of uncertainty to symptoms of generalized anxiety disorder, major depressive disorder, and obsessive-compulsive disorder. Clinical Psychology Review, 31, 923–933. doi: 10.1016/j.cpr.2011.05.001 [DOI] [PubMed] [Google Scholar]
  17. Grillon C, Franco-Chaves JA, Mateus CF, Ionescu DF, & Zarate CA (2013). Major depression is not associated with blunting of aversive responses; Evidence for enhanced anxious anticipation. PLoS ONE, 8, 1–7. doi: 10.1371/journal.pone.0070969 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Hajcak G, McDonald N, & Simons RF (2004). Error-related psychophysiology and negative affect. Brain and Cognition, 56, 189–197. doi: 10.1016/j.bandc.2003.11.001 [DOI] [PubMed] [Google Scholar]
  19. Hamilton JP, Etkin A, Furman DJ, Lemus MG, Johnson RF, & Gotlib IH (2012). Functional neuroimaging of major depressive disorder: A meta-analysis and new integration of baseline activation and neural response data. American Journal of Psychiatry, 169, 693–703. doi: 10.1176/appi.ajp.2012.11071105 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hanna GL, Liu Y, Isaacs YE, Ayoub AM, Brosius A, Salander Z, Arnold PD, & Gehring WJ (2018). Error-related brain activity in adolescents with obsessive-compulsive disorder and major depressive disorder. Depression and Anxiety, 35, 752–760. doi: 10.1002/da.22767 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Helmstadter GC (1964). Principles of psychological measurement. New York, NY: Appleton-Century-Crofts, Inc. [Google Scholar]
  22. Hill KE, Samuel DB, & Foti D. (2016). Contextualizing individual differences in error monitoring: Links with impulsivity, negative affect, and conscientiousness. Psychophysiology, 53, 1143–1153. doi: 10.1111/psyp.12671 [DOI] [PubMed] [Google Scholar]
  23. Holmes AJ, & Pizzagalli DA (2008). Spatiotemporal dynamics of error processing dysfunctions in major depressive disorder. Archives of General Psychiatry, 65, 179–188. doi: 10.1001/archgenpsychiatry.2007.19 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Jackson F, Nelson BD, & Hajcak G. (2016). The uncertainty of errors: Intolerance of uncertainty is associated with error-related brain activity. Biological Psychology, 113, 52–58. doi: 10.1016/j.biopsycho.2015.11.007 [DOI] [PubMed] [Google Scholar]
  25. Jensen D, Cohen JN, Mennin DS, Fresco DM, & Heimberg RG (2016). Clarifying the unique associations among intolerance of uncertainty, anxiety, and depression. Cognitive Behaviour Therapy, 45, 431–444. doi: 10.1080/16506073.2016.1197308 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. MacNamara A, Kotov R, & Hajcak G. (2016). Diagnostic and symptom-based predictors of emotional processing in generalized anxiety disorder and major depressive disorder: An event-related potential study. Cognitive Therapy and Research, 40, 275–289. doi: 10.1007/s10608-015-9717-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Meyer A, Hajcak Proudfit G, Torpey-Newman DC, Kujawa A, & Klein DN (2015). Enhanced error-related brain activity in children predicts the onset of anxiety disorders between the ages of 6 and 9. Journal of Abnormal Psychology, 124, 266–274. doi: 10.1037/abn0000044 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Meyer A, Bress JN, Hajcak G, & Gibb BE (2018). Maternal Depression Is Related to Reduced Error-Related Brain Activity in Child and Adolescent Offspring. Journal of Clinical Child & Adolescent Psychology, 47(2), 324–335. doi: 10.1080/15374416.2016.1138405 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Miller GA, Gration G, & Yee CM (1988). Generalized implementation of an eye movement correction procedure. Psychophysiology, 25, 241–243. doi: 10.1111/j.1469-8986.1988.tb00999.x [DOI] [Google Scholar]
  30. Moran TP, Schroder HS, Kneip C, & Moser JS (2017). Meta-analysis and psychophysiology: A tutorial using depression and action-monitoring event-related potentials. International Journal of Psychophysiology, 111, 17–32. doi: 10.1016/j.ijpsycho.2016.07.001 [DOI] [PubMed] [Google Scholar]
  31. Moser JS, Moran TP, & Jendrusina AA (2012). Parsing relationships between dimensions of anxiety and action monitoring brain potentials in female undergraduates. Psychophysiology, 49, 3–10. doi: 10.1111/j.1469-8986.2011.01279.x [DOI] [PubMed] [Google Scholar]
  32. Moser JS, Moran TP, Schroder HS, Donnellan MB, & Yeung N. (2013). On the relationship between anxiety and error monitoring: A meta-analysis and conceptual framework. Frontiers of Human Neuroscience, 7, 1–19. doi: 10.3389/fnhum.2013.00466 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Nelson BD, Shankman SA, & Produfit GH (2014). Intolerance of uncertainty mediates reduced reward anticipation in major depressive disorder. Journal of Affective Disorders, 158, 108–113. doi: 10.1016/j.jad.2014.02.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Nelson BD, Jackson J, Amir N, & Hajcak G. (2017). Attention bias modification reduces neural correlates of response monitoring. Biological Psychology, 129, 103–110. doi: 10.1016/j.biopsycho.2017.08.059 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Olvet DM, & Hajcak G. (2008). The error-related negativity (ERN) and psychopathology: Toward an endophenotype. Clinical Psychology Review, 28, 1343–1354. doi: 10.1016/j.cpr.2008.07.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Olvet DM, Klein DN, & Hajcak G. (2010). Depression symptom severity and error-related brain activity. Psychiatry Research, 179, 30–37. doi: 10.1016/j.psychres.2010.06.008 [DOI] [PubMed] [Google Scholar]
  37. Overbeek TJM, Nieuwenhuis S, Ridderinkhof KR (2005). Dissociable components of error processing: On the functional significance of the Pe vis-à-vis the ERN/Ne. Journal of Psychophysiology, 19, 319–329. doi: 10.1027/0269-8803.19.4.319 [DOI] [Google Scholar]
  38. Riesel A. (in press). The erring brain: Error-related brain activity as a risk-marker for psychopathology. [Google Scholar]
  39. Riesel A, Goldhahn S, & Kathmann N. (2017). Hyperactive performance monitoring as a transdiagnostic marker: Results from the health anxiety in comparison to obsessive-compulsive disorder. Neuropsychologia, 96, 1–8. doi: 10.1016/j.neuropsychologia.2016.12.029 [DOI] [PubMed] [Google Scholar]
  40. Schrijvers D, de Bruijn ERA, Maas YJ, Vancoillie P, Hulstijn W, & Sabbe BGC (2009). Action monitoring and depressive symptom reduction in major depressive disorder. International Journal of Psychophysiology, 71, 218–224. doi: 10.1016/j.ijpsycho.2008.09.005 [DOI] [PubMed] [Google Scholar]
  41. Speed BC, Nelson BD, Auerbach RP, Klein DN, & Hajcak G. (2016). Depression risk and electrocortical reactivity during self-referential emotional processing in 8 to 14 year-old girls. Journal of Abnormal Psychology, 125, 607–619. doi: 10.1037/abn0000173 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Tanovic E, Gee DG, & Joormann J. (2018). Intolerance of uncertainty: Neural and psychophysiological correlates of the perception of uncertainty as threatening. Clinical Psychology Review, 60, 87–99. doi: 10.1016/j.cpr.2018.01.001 [DOI] [PubMed] [Google Scholar]
  43. Watson D. (2005). Rethinking the mood and anxiety disorders: A quantitative hierarchical model for DSM-V. Journal of Abnormal Psychology, 114, 522–536. doi: 10.1037/0021-843X.114.4.522 [DOI] [PubMed] [Google Scholar]
  44. Weinberg A, Olvet DM, & Hajcak G. (2010). Increased error-related brain activity in generalized anxiety disorder. Biological Psychology, 85(3), 472–480. doi: 10.1016/j.biopsycho.2010.09.011 [DOI] [PubMed] [Google Scholar]
  45. Weinberg A, Klein DN, Hajcak G. (2012). Increased error-related brain activity distinguished generalized anxiety disorder with and without comorbid major depressive disorder. Journal of Abnormal Psychology, 121, 885–896. doi: 10.1037/a0028270 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Weinberg A, Kotov R, & Proudfit GH (2015). Neural indicators of error processing in generalized anxiety disorder. Journal of Abnormal Psychology, 124, 172–185. doi: 10.1037/abn0000019 [DOI] [PubMed] [Google Scholar]
  47. Weinberg A, Liu H, & Shankman SA (2016a). Blunted neural response to errors as a trait marker of melancholic depression. Biological Psychology, 113, 100–107. doi: 10.1016/j.cpr.2008.07.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Weinberg A, Meyer A, Hale-Rude E, Perlman G, Kotov R, Klein DN, & Hajcak G. (2016b). Error-related negativity (ERN) and sustained threat: Conceptual framework and empirical evaluation in an adolescent sample. Psychophysiology, 53, 372–385. doi: 10.1111/psyp.12538 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Yeung N, Botvinick MM, & Cohen JD (2004). The neural basis of error detection: Conflict monitoring and the error-related negativity. Psychological Review, 111, 931–959. doi: 10.1037/0033-295X.111.4.931 [DOI] [PubMed] [Google Scholar]
  50. Yook K, Kim K, Suh SY, & Lee KS (2010). Intolerance of uncertainty, worry, and rumination in major depressive disorder and generalized anxiety disorder. Journal of Anxiety Disorders, 24, 623–628. [DOI] [PubMed] [Google Scholar]
  51. Zambrano-Vazquez L, & Allen JJB (2014). Differential contributions of worry, anxiety, and obsessive compulsive symptoms to ERN amplitudes in response monitoring and reinforcement learning tasks. Neuropsychologia, 61, 197–209. doi: 10.1016/j.neuropsychologia.2014.06.023 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1647210-supplement-1.docx (661.2KB, docx)

RESOURCES