Major Depressive Disorder and Emotion-Related Impulsivity: Is
Cognitive Inhibition the Mechanism?
By Maria Roos Dekker (10551301)
Master thesis, Research master Psychology at the University of Amsterdam (UvA)
Under supervision of Dr. S. L. Johnson (University of California, Berkeley) and Dr. B. J. Verschuere (UvA).
Date: July 17th 2015
Abstract
Previous studies indicate that depressed individuals show mood-related impulsive tendencies that have been associated with increased suicide risk. However, little is known about the underlying mechanisms. This study investigated inhibition deficits as a possible construct underlying emotion-related impulsivity. It was hypothesized that after a negative mood induction, depressed participants would show more inhibition deficits than controls, and that these deficits would mediate the link between depression and emotion-related impulsivity. To test these hypotheses, 47 participants diagnosed with major depressive disorder and 63 non-depressed controls underwent a negative mood induction and then completed tasks that assessed components of inhibition: the ability to suppress pre-potent responses (Antisaccade task) and the ability to resist interference (Word-naming task). Participants also completed self-report measures of current depressive symptoms and impulsivity (Three-Factor Impulsivity scale and Positive Urgency Measure). Depressed individuals reported more impulsivity in both emotional and non-emotional states compared to controls. In addition, depressed individuals did not differ from controls in the ability to inhibit pre-potent responses and the ability to resist interference. Moreover, individuals that were impulsive in negative emotional states were characterized by lower pre-potent response capacities and were less capable to resist interference. Thus, this study indicates that impulsivity is an important characteristic of major depression, which could inform psychological interventions for depression.
Introduction
Major depressive disorder (MDD) is one of the most common psychological disorders, with a lifetime prevalence of 16.6 percent (Kessler, Chiu, Demler & Walters, 2005). Major depression may increase the risk of premature death (Carney et al., 2003) and has been associated with impairment in both interpersonal and occupational functioning (Reddy, 2010). Despite the available psychological and pharmacological treatments for depression, relapse rates remain high (Steinert, Hofmann, Kruse & Leichsenring, 2014). It is therefore important to understand the underlying factors that contribute to the development, maintenance and recurrence of depression.
One of the core problems related to depression is poor cognitive control over emotions. A recent model suggests that poor control over emotions might reflect a more general tendency towards a reflexive, automatic and impulsive response to emotion (Carver, Johnson & Joormann, 2008). This broader lack of control has been described as emotion-relevant impulsivity. A key issue in understanding emotion-relevant impulsivity is that this lack of control will manifest in different types of behavior, depending on the emotion that is being reacted to. That is, emotion-related impulsivity has been emotion-related to both externalizing (e.g. impulsive violence) and internalizing symptoms (e.g. depression) (Johnson, Carver & Joormann, 2013). The focus of this study is on understanding emotion-relevant impulsivity in depression.
Two different facets of impulsivity, positive and negative urgency (the tendency to act rashly following a positive or negative mood, respectively), have been described in the literature (Cyders et al., 2007). Nevertheless, the scope of emotion-relevant impulsivity is broader than urgency. For instance, the tendency to reflexively generalize from negative events, to passively respond to emotions and to let emotions influence the way of looking at life, have been described in the literature as other emotion-related facets of impulsivity (Johnson, Joormann, Kim & Nam, 2011). Since depressed individuals tend to experience more negative (i.e. sad), rather than positive emotions, the over-reactivity to these emotions will often trigger behavior that follows from sadness (e.g. inaction, lethargy, fatigue and negative generalization). Moreover, when depressed individuals experience anger, the impulsive lack of control may lead to more dramatic expressions of aggression (Dutton & Karakanta, 2013). Depression is also tied to an increased risk of self-destructive behavior (i.e. self-harm and suicide) during negative mood states (Oquendo et al., 2004), and this has been tied to emotion-relevant impulsivity (Bresin, Carter & Gordon, 2012). However, it is of importance to note that depressed individuals show a general impulsivity during both positive and negative emotions (Carver et al., 2008). Impulsive tendencies in depression during positive emotion states may relate to tendencies to pursue immediate gratification (e.g. excessive alcohol use and binge eating) that have been found to be
elevated during depression (Karyadi & King, 2011; Smith, Guller & Zapolski, 2013).
The association between depression and impulsivity is well supported. For instance, several studies found an association between depressive symptoms and self-report measures of impulsivity (Miller, Flory, Lynam & Leukefeld, 2003; d’Acremont & Van der Linden, 2007; Clarke, 2012). In addition, self-reported impulsivity has been related to depression in clinical samples (Corruble, Benyamina, Bayle, Falissard & Hardy, 2003; Peluso et al., 2007; Ekinci, Albayrak & Caykoylu, 2011; Perroud, Baud, Mouthon, Courtet & Malafosse, 2011; Henna et al., 2013) and may be a factor that contributes to the onset of depression (Granö et al., 2007). Furthermore, depressive symptoms have been associated with different types of impulsive behavior, such as pathological gambling (Becona, Lorenzo & Fuentes, 1996; Clarke, 2006; Dussault, Brendgen, Vitaro, Wanner & Tremblay, 2011) and risky sexual behavior (Khan et al., 2009).
Nonetheless, only a few studies have investigated emotion-related impulsivity in relation to depression. Research has shown, for instance, that emotion-reactive impulsivity is related to depressive symptom severity (d’Acremont & Van der Linden, 2007; Johnson et al., 2013; Miller et al., 2003; Pang, Farrahi, Glazier, Sussman & Leventhal, 2014; see Berg, Latzman, Bliwise & Lilienfeld, 2015 for a meta-analysis). This effect does not appear to be just due to the severity of negative emotions experienced during depression, in that depressive symptoms are specifically tied to impulsivity in positive mood states (positive urgency) (Johnson et al., 2013; Marmorstein, 2013; Karyadi & King, 2011). In clinical samples, it has been found that individuals diagnosed with past MDD have a general tendency to be over-reactive towards both negative and positive emotions, as assessed by the Three-Factor Impulsivity self-report scale (Carver, Johnson & Joormann, 2013). Findings suggest that negative urgency is elevated even after remission among those diagnosed with MDD (Carver, Johnson, & Joorman, 2013; Grall-Bronnec et al., 2012).
Little is known about the underlying mechanisms of emotion-related impulsive tendencies. It has been proposed that deficits in cognitive control minimize the likelihood that the reflexive, impulsive system will be overridden (Carver et al., 2008). Since cognitive control is thought to be superordinate to automatic tendencies, sufficient effortful control could either dampen or enhance automatic behavior. For instance, in individuals with a high approach temperament, cognitive control can restrain the automatic tendency of approach-related behavior (e.g. by reducing sensation seeking or aggressive tendencies). Likewise, sufficient cognitive abilities can reduce the tendency towards inaction in individuals with a non-reactive or blunted approach system (i.e. force the formation of an action that the individual reflexively does not want to take). When sufficient cognitive control abilities are not available, impulsive responses to
emotional states will occur. In case of depression, deficits in cognitive control may reduce the chance that automatic tendencies following negative (e.g. apathy, fatigue and passivity) or positive emotions (e.g. excessive alcohol use) are overcome, which can maintain or enhance these types of behaviors (Carver et al., 2008; Karyadi & King, 2011; Snyder, 2012). Thus, the same cognitive control deficits are thought to underlie impulsive tendencies towards negative as well as positive emotions, but can manifest in different kinds of behaviors (e.g. inaction or action). More specifically, deficits in one aspect of executive functioning, cognitive inhibition, might underlie emotion-related impulsive tendencies. This idea can be supported by research showing that urgency is related to difficulties in inhibiting pre-potent responses, most typically assessed using measures of response inhibition (Gay, Rochat, Billieux, d’Acremont & van der Linden, 2008; Gay et al., 2010; Rochat, Beni, Annoni, Vuadens & van der Linden, 2013; Wilbertz et al., 2014; Gunn & Finn, 2015).
Empirical research suggests, however, that inhibition is not a unitary construct, but consists of several components that are differently related to cognitive task performance. The current study was designed to assess individual differences in two key facets of inhibition the ability to suppress automatic, dominant or pre-potent responses (pre-potent response inhibition) and the ability to control interference from irrelevant external information (resistance to distractor interference (DI)) (Friedman & Miyake, 2004).
Previous literature provides only a rough estimate of the strength of effect of the different aspects of inhibition in depression. Findings of meta-analyses indicate that MDD is related to deficits in pre-potent response inhibition, as assessed by the Color-Word Stroop and the Hayling task (Snyder, 2012), and with small-to-medium deficits on a go/no go task (Wright et al., 2014). The Antisaccade task, another variant of pre-potent response inhibition tasks, has been found to differentiate dysphoric and depressed individuals from controls in some (Sweeney, Strojwas, Mann & Thase, 1998; Derakshan, Salt & Koster, 2009; Carvalho et al., 2014) but not all studies (De Lissnyder, Derakshan, De Raedt & Koster, 2011). Moreover, findings may vary by task; research using the Stop Signal task found no significant correlation with depressive symptoms (Vergara-Lopez, Lopez-Vergara & Colder, 2013) or depressive diagnoses on this task (Sjoerds, van den Brink, Beekman, Penninx, & Veltman, 2014).
Beyond these studies on cognitive inhibition, other work has focused more specifically on inhibition of negatively valenced information. Individuals with major depressive disorder have been shown to have difficulties inhibiting automatic responses towards negative emotional stimuli on the emotional Color-Word Stroop task as compared to controls (Mitterschiffthaler et al., 2008; Dai & Feng, 2011). These findings indicate that depressed individuals have difficulty
disengaging from and inhibiting attention toward negatively valenced information. This has been conceptualized as a more specific form of deficit in pre-potent response inhibition (Dai & Feng, 2011).
Beyond difficulty with inhibiting pre-potent responses to negatively valenced information, depressive symptoms have found to be associated with difficulty inhibiting task-irrelevant negative information (poor resistance to DI), as assessed by a negative affective priming task (Joormann & Gotlib, 2010; Frings, Wentura, & Holtz, 2007; Zetsche & Joormann, 2011). Parallel findings have emerged in clinical samples as compared to healthy controls, with depression relating to interference of irrelevant negative information (poor resistance to DI) on the Prose Distraction Task (Lau, Christensen, Hawley, Gemar & Segal, 2007) and the Emotional Flanker Task (Zetsche, d’Avanzato & Joormann, 2012). However, Joormann, Nee, Berman, Jonides and Gotlib (2010) did not find support for depression-related deficits in resistance to DI.
Thus, results from a wide array of tasks showed that inhibitory dysfunction is present in depression. However, the literature is rather inconsistent about which specific elements of inhibition are impaired, and findings have been a bit more consistent for pre-potent response inhibition, than resistance to DI, suggesting that certain facets of inhibition may be more impaired than others. If one form of inhibition is more distinctly impaired for those with depression, this will be an important area to target for future studies of cognitive mediation. Moreover, researchers tend to find that inhibition is impaired specifically for negative information in depression (e.g. Lau et al., 2007; Zetsche & Joormann, 2011). Although this is typically attributed to the role of negative content as a prime for negative mood state (Zetsche & Joormann, 2011), inhibition has not yet been examined in negative mood states. The current study will therefore directly activate the negative mood state rather than relying on negative words. Furthermore, it remains unexamined whether emotion-related inhibitory deficits can help explain why depressed individuals tend to be impulsive during affective states.
Aims and hypotheses
This study investigated emotion-related impulsivity and two types of inhibition (i.e. pre-potent response inhibition and resistance to DI) in participants diagnosed with partially or fully remitted or current major depressive disorder (depression group) and non-depressed controls. Based on previous literature, it was hypothesized that after a negative mood induction, the depression group would show (1) more pre-potent inhibition deficits (less accuracy on the Antisaccade task) (Derakshan, Salt & Koster, 2009; Carvalho et al., 2014) and (2) more deficits from distractor interference (slower responses on the distractor compared to no-distractor primes on the
Word-naming task) (Lau et al., 2007; Zetsche et al., 2012) than controls. In addition, (3) it was expected that the relationship between depression and emotion-relevant impulsivity (Three-Factor Impulsivity scale and Positive Urgency Measure (Carver, Johnson, Joormann, Kim & Nam, 2011; Carver et al., 2013; Cyders et al., 2007) was mediated by inhibition deficits on the Antisaccade and the Word-naming task (see Figure 1) (see measures below).
Methods
Participants
Participants (N = 110) were undergraduate students (69% female, mean age = 21.89, SD = 4.37), of which 84.6% majored in Psychology. Participants described their ethnicity as 32.4% European American/White, 13.5% Asian American/Asian, 11.7% Chinese, 9% Indian, and 33.3% other. The participants had a mean GPA of 3.45 (SD = .37), rated on a four-point scale. Forty-seven participants were included in the depression group and 63 participants were included as controls. Demographic and clinical characteristics of the sample are summarized in Table 2. Eighteen participants scored within the range of severe depressive symptoms on the Inventory of Depressive-Symptomatology Self-Report (IDS-SR30, see below) (> 48) (Rush et al., 2003). Only 4 of the participants in the depression group were diagnosed with current major depressive disorder by the DSM-IV criteria (APA, 2000); 8 met criteria for partial remission, and 35 met criteria for full remission. All participants received partial credits toward their psychology class research participation requirements as part of their study.
Procedure
Potential participants from undergraduate psychology classes at a large public university were screened on the Inventory to Diagnose Depression-Lifetime (IDD-L, see below). Those who met the IDD-L criteria for past depression were invited to take part in the study by e-mail. In addition, to recruit control participants, potential participants were invited to take part in the study if they endorsed scores ≤ 4 symptoms on the IDD-L. All participants completed questionnaires online (described below) and were then invited to take part in a face-to-face session. During the session, participants first completed informed consent procedures, after which the Vocabulary subscale of the Shipley Institute of Living Scale (SHIPLEY, see below) was administered. Then participants completed the Implicit Positive and Negative Affect Test (IPANAT, see below) to determine baseline affect (IPANAT time 1). Subsequently, participants randomly watched one out of three film clips (Table 1) before each task (Antisaccade or Word-naming task) to induce a negative mood. The tasks were administered in counterbalanced order. After each film clip the IPANAT
was administered to assess the effectiveness of the mood induction (IPANAT time 2 and 3). Participants were seated about 35 cm from the computer screen for the film clips and tasks. After completing the above measures and tasks, participants watched a film clip designed to induce a positive mood (see Table 1) (for procedure order, see also Figure 2). Then, a diagnostic interview was administered (see below). The tasks and interview took place in a quiet, private testing room, and all participants were fully debriefed at the end of the study.
Measures
The Structured Clinical Interview for the DSM-IV Axis I disorders, past and current
MDD module (SCID; First, Spitzer, Gibbon & Williams, 1997) was administered by trained
clinical graduate students. The SCID has good psychometric properties, with good inter-rater reliability (Cohen Kappa for MDD =.66) (Lobbestael, Leurgans & Arntz, 2011). Inter-rated reliability for this study was excellent (13 tapes rated by 2 judges, Cohen Kappa for past MDD =.84; for current MDD=1.00). Participants were included in depression group if they met the criteria for major depressive episode at some point in their life (individuals that are remitted, partially remitted or currently depressed). Controls were individuals that did not meet criteria for past or current MDD.
Emotion-eliciting film clips (Rottenberg, Ray & Gross, 2007; Joormann, Talbot & Gotlib,
2007) have been widely used to induce a negative mood. Film clips used in the current study have been previously validated as an effective way to elicit emotions in depressed samples (Rottenberg, Gross & Gotlib, 2005). Three film clips were used, to allow for analyses of individual differences in responses to specific film clips. Furthermore, because the affective response following an emotion-eliciting film does not last for the duration of both behavioral tasks (see below), two mood inductions were used (Rottenberg et al., 2007). In addition, one film clip was used to induce a positive mood at the end of the study (see Table 1) (Gruber, Johnson, Oveis & Keltner, 2008).
The Implicit Positive and Negative Affect Test (IPANAT; Quirin, Kazén, & Kuhl, 2009) was used to assess the effectiveness of negative mood inductions. Participants are asked to rate the extent to which six artificial words (safme, vikes, tunba, talep, belni and sukov) express a certain mood (Positive Affect (PA): happy, cheerful, energetic; Negative Affect (NA): helpless,
tense, inhibited). Each artificial word is presented along with the six emotions words. Words are
rated on a scale ranging from (1“Doesn’t fit at all” to 4 “Fits very well”). The IPANAT has demonstrated to measure both trait and state variance. The outcome measure is the average score of the three negative mood adjectives, ranging from 1 to 4. The IPANAT has good psychometric
properties with high internal consistency for both subscales (PA α = .87; NA α = .89; current sample NA α = .84) and good construct validity (Quirin et al., 2009).
The Inventory to Diagnose Depression-Lifetime (IDD-L; Zimmerman & Coryell, 1987) is 45-item self-report measure developed to measure lifetime depressive symptoms and diagnoses of major depressive disorder. Participants are asked to rate 22 depressive symptoms on a 5-point scale (e.g., 1 “I did not feel sad or depressed” to 5 “I was so sad or unhappy that I couldn’t stand
it”). Weight loss is coded only if participants report that they were deliberately trying to lose
weight. For each symptom endorsed, participants are asked whether it persisted for more or less than two weeks. Symptom severity is scored as the total number of DSM-IV depressive symptoms surpassing a severity threshold and lasting at least two weeks. A dichotomous score is also generated for diagnosis of past depression, according to DSM-IV criteria (APA, 2000). The IDD-L has good psychometric properties with high internal consistency (α = .92; current sample
α = .88) and strong concordance with other measures of lifetime depression (Diagnostic
Interview Schedule (DIS): K = .60) (Zimmerman & Coryell, 1987).
The Inventory of Depressive-Symptomatology Self-Report (IDS-SR30; Rush, Gullion, Basco, Jarrett & Trivedi, 1996) was designed to assess current depressive symptoms. The IDS-SR30 consists of 30 items that assess depressive symptoms during the past seven days. Each item is rated on a Likert scale (0 “I do not feel sad” to 3 “I feel sad nearly all of the time”). The total score ranges from 0 to 84. The IDS-SR30 has strong psychometric properties with good internal consistency (α = .80; current sample α = .86) and high concurrent validity with the depression factor of the Symptom Checklist-90 items, revised version (SCL-90R) (r = .84) (Derogatis, 1977; Corruble, Legrand, Duret, Charles & Guelfi, 1999).
The Three-Factor Impulsivity scale (Carver et al., 2011) is a 97-item self-report measure designed to assess both emotion-relevant and non-emotion-relevant items. The scale draws on previously validated measures of impulsivity, including Urgency and the Positive Urgency Measure, as well as a novel set of items designed to broaden coverage of ways that emotion might influence cognition and behavior. Two of the factor-analytically based scales are emotion-related: Factor 1 = Pervasive Influence of Feelings (e.g. ‘When even one thing goes wrong I begin to
wonder if I can do well at anything at all’), and Factor 3 = Feelings Trigger Action (e.g. ‘When I have an emotional reaction to something, I often act without thinking’). Factor 2 = (Lack of)
Follow-Through (e.g. ‘I tend to give up easily’) does not refer to emotions. 77 items are rated on a Likert scale (1 “I agree a lot” to 5 “I disagree a lot”) and 19 items are rated on a Likert scale ranging from (1 ”never” to 5 “very often”). This scale has good psychometric properties with high internal consistency for all subscales (α = .90; current sample α = .90) and has been found to
correlate with a broad range of psychopathologies including depression, as well as risk factors for impulsivity (Carver et al., 2011; Carver et al., 2013).
The Positive Urgency Measure (Cyders et al., 2007) is a 14-item self-report measure designed to assess rashly behavior following positive mood states (e.g. ‘When overjoyed, I feel
like I can’t stop myself from going overboard’). Items are rated on a Likert scale (1 ”agree
strongly” to 4 ”disagree strongly”). Seven items overlap with the Three-Factor Impulsivity scale. The Positive Urgency Measure has good psychometric properties with high reliability (α = .94; current sample α = .95) and good discriminant validity among five impulsivity-like traits (Del Boca, Darkes, Greenbaum & Goldman, 2004; Cyders et al., 2007).
The Vocabulary subscale of the Shipley Institute of Living Scale (SHIPLEY; Shipley, 1940) is designed to provide an estimated IQ score. For each of the 40 items, participants are presented with a target word and asked to choose one out of three words that is closest to the meaning of the target word. The total raw score ranges from 0 to 40. Raw scores can be converted to estimate Wechsler Adult Intelligence Scale (WAIS) scores. In this study, the Shipley was added to the study only for the second half of the participants. The Vocabulary subscale has good psychometric properties with good reliability (test-retest correlation = .77) and has been found to strongly correlate with the WAIS Full Scale IQ (Goodman, Streiner & Woodward, 1974; Zachary, 1986).
Behavioral measures
The Antisaccade task (Kane, Bleckley, Conway & Engle, 2001) is a well-validated measure of pre-potent response inhibition. In this computerized task, participants are asked to actively suppress a reflexive saccade towards a cue and instead look to the other side of the screen to correctly identify a target letter (i.e. B, P, R) by pressing the keys 4, 5 or 6 on the right side of the number keypad, which had been labeled as B, P and R, respectively. The task consisted of four blocks: a response-mapping practice block (10 trials), a pro-saccade practice block (10 trials), an antisaccade practice block (10 trials) and an antisaccade experimental block (40 trials). In the response-mapping trials, participants were required to look at the middle of the screen and react as accurately and quickly as possible. A ‘READY’ signal appeared on the screen at the start of the block. This signal stayed on the screen until the subject pressed the spacebar, which was followed by a black screen for 400 ms. Then, a fixation signal (***) was shown on the screen for a duration of 200, 600, 1000, 1400 or 1800 ms. The fixation durations varied randomly over trials. A second black screen appeared for 100 ms, followed by an equal sign (=), which was centrally presented for 150 ms. Then, a third black screen was shown for 50 ms after which a
target letter appeared at the center of the screen. The target letter only appeared very briefly (100 ms), after which it was masked by ‘H’ (50 ms), followed by the number ‘8’. The ‘8’ stayed on the screen until the response was made (but only up to 10 seconds). The prosaccade trials were parallel with the response-mapping trials, with the exception that the target letter and the masks (‘H’ and ‘8’) appeared either on the right or left side of the fixation (***). The location of the target letter was cued by a ‘=’ symbol that flashed for 150 ms. Participants were instructed to direct their eyes towards the flashing cue. The antisaccade practice trials, differed from prosaccade practice trials only in that the cue ‘=’ appeared on the opposite side of the screen from the upcoming target letter (see Figure 3). The antisaccade experimental trials were identical to the antisaccade practice block. The outcome measure was accuracy (proportion of trials correct) on the task. This task has good psychometric properties with good reliability for the proportion of errors (split-half correlation = .87; current sample split-half correlation = .451) and good convergent validity (Friedman & Miyake, 2004).
The Word-naming task (Kane, Hasher, Stoltzfus, Zacks & Connely, 1994) was designed to assess resistance to DI. In this computerized task, participants were asked to name a green target word that was presented either alone (no distractor prime) or with a red distractor word (distractor prime). The task included 96 experimental trials, of which 64 trials included distractor primes and 32 trials with no distractor primes. Participants completed 6 practice trials that consisted of both distractor primes (4) and no-distractor primes (2). The target words that had to be named out loud were cat, pot, jar, tie, cup, fun, gin, bag and rod. The targets were presented above and below a fixation point at the center of the screen and appeared equally often in each position. At the start of the experiment, a ‘READY’ cue appeared. This cue stayed on the screen until the subject pressed ‘spacebar’, which was followed by a blank screen (1100 ms) and a fixation point (500 ms). Then, the prime target appeared on the screen for 225 ms while a beep was presented simultaneously via audio. Then, the prime target was masked by colored dots for 100 ms. The screen was blank until the participant responded (see Figure 4). The experimenter recorded the verbal responses and pressed a spacebar as soon as the participant responded. The spacebar was used as a backup for reaction time (RT) recording, but was not used for primary analyses. RTs were calculated from the time of the beep (when the word was visually presented) until the participant made a verbal response. Trials on which an error was made were eliminated from RT analyses. An error was scored when participants failed to make a verbal response, named the incorrect target word (the distractor), partially named the incorrect target word or
1 Reliability is lower for the current sample due to practice effects; block 1 (M = .61, SD = .21), block 2 (M = .69, SD = .17); t (107) = -4.3, p = 0.00.
stutter while saying correct target names (Kane et al., 1994). The outcome measure was the difference in RT between no-distractor and distractor primes (interference score). This task has good psychometric properties with good reliability (split-half correlation = .76; current sample split-half correlation = .68), good ecological validity and has found to be correlated to other measures of resistance to DI, such as the Eriksen Flanker task (Friedman & Miyake, 2004; Jones & Estes, 2012).
Operationalization and Data Analysis
Statistical analyses were performed using the Statistical Package for Social Sciences (SPSS) for Windows, version 22 (IBM, 2013).
As a manipulation check, a repeated-measures Analysis of Variance (ANOVA) with comparisons was conducted (with Time as within factor) to investigate whether implicit negative affect (scores IPANAT) changed significantly from baseline over time.
Hypothesis one was that the depression group would be less accurate on the Antisaccade task, and formally tested with an Analysis of Covariance (ANCOVA) with diagnostic group (depression vs. controls) as the independent variable. The outcome measure was the accuracy on the antisaccade trials and the covariate was the accuracy on the prosaccade trials. As a secondary analysis, an ANCOVA was performed to assess whether the groups differed on RT on the antisaccade trials. The covariate that was controlled for in the analysis was the RT on prosaccade trials.
Hypothesis two was that the depression group would show slower responses on the distractor compared to no-distractor primes on the Word-naming task, as compared to controls. This was tested with an ANOVA, with diagnostic group as the independent variable. The outcome measure was the difference in RT between no-distractor and distractor primes (no-distractor minus (no-distractor trials = interference score) on the Word-naming task. Effect sizes for each of the two tests were calculated to provide preliminary data on the strength of effect for the two forms of inhibition.
Hypothesis three was that the relationship between depression and impulsivity would be mediated by inhibition deficits on the Antisaccade and the Word-naming task. This was tested with a mediation analysis. Differences between the diagnostic groups on the inhibition measures were tested by hypotheses one and two. A Multivariate Analysis of Variance (MANOVA) was used to assess whether the diagnostic groups differed on the impulsivity measures (Three-Factor Impulsivity scale and Positive Urgency Measure). Correlational analyses (two-tailed) were used to assess the relationship between inhibition deficits (Antisaccade and Word-naming outcome
measures) and impulsivity. Parallel analyses were conducted to consider whether current depression (IDS-SR30 score) influenced findings.
Data Trimming and Outlier Analyses
For the Antisaccade task, data was missing for two persons due to equipment failure, one participant was excluded because of poor cooperation during the experiment combined with a score lower than the sample mean by 3 SD on the antisaccade trials, and one participant was excluded because prosaccade results indicated that the participant did not understand the task or was not motivated. Subsequently, two participants were considered statistical outliers because their scores deviated more than 3 SD from the mean RT scores on the prosaccade and antisaccade trials. These scores were replaced by the mean plus 3 SD. After exclusions, 106 participants were included in analyses.
For the Word-Naming task, one data file was missing due to equipment failure. Records of 18 participants did not contain the beep, so these were scored based on the time from the spacebar pressed by the experimenter, using a linear regression model to estimate RT values of these participants. Since participants became faster in verbal responding over time, trial number was included in the regression function. Excluding these participants from analyses did not yield significantly different results.
Errors on the Word-naming task were removed before computing individual scores, resulting in an average of 1 trial per individual (SD = 1.13). In accordance with Friedman and Miyake (2004), RTs were cleaned at the trial level in the following manner: First, RTs < 200 ms or > 1500 ms were removed. Then, RTs that were smaller than 300 ms or greater than 1000 ms were replaced with values of 300 ms and 1000 ms, respectively. Finally, mean RTs scores that deviated more than 3SD from the mean for that participant were replaced by the mean plus or minus 3 SD. In addition, the interference scores of two participants deviated more than 3SD from the mean and were replaced by the mean score plus or minus 3 SD.
Results
Imputation
Missing values on the Three-Factor Impulsivity scale, the Positive Urgency Measure and the IDS-SR30 were imputed using multiple regression. Missing values on the implicit negative mood scale (IPANAT) were imputed using the Maximization Expectation algorithm (EM).
Mood Induction
A one-way repeated measures ANOVA was conducted to assess whether implicit negative mood changed over time. The outcome measure of the IPANAT did not meet the assumption of sphericity (x2(5) = 194.99, p < 0.01). Therefore, the degrees of freedom were corrected using Greenhouse-Geisser estimates of sphericity (ε = 0.77). Results showed a significant time effect (N = 110, F(2.30, 1007.48) = 23.17, p < 0.01, η2 = .05). Further examination of this effect with
planned comparisons indicated that implicit negative mood scores were significantly higher at both time points 2 and 3 compared to baseline (time point 1), p’s < 0.01.
Group Differences in Inhibition and Impulsivity
A one-way ANCOVA was conducted to assess whether the diagnostic groups (depression versus control) differed on antisaccade accuracy and a one-way ANOVA was used to assess group differences on the Word-naming task. As shown in Table 3, the depression group did not differ significantly from the control group on antisaccade accuracy (while controlling for prosaccade accuracy) or word-naming performance.
As a secondary analysis, a one-way ANCOVA was performed to assess diagnostic group differences on antisaccade RT, controlling for RT on prosaccade trials. Individuals in the depression group did not differ from controls on antisaccade RT (see Table 3).
A MANOVA was conducted to assess whether the diagnostic groups differed on the impulsivity measures. Using Wilks’s Lambda statistic, results showed a significant effect of diagnostic group on the four impulsivity measures (N = 110,
∧
= .85, F(4,105) = 4.52, p = .00, ηp= .15). As shown in Table 3, separate univariate ANOVAs revealed that the depression group significantly differed from controls on Pervasive Influence of Feelings, (Lack of) Follow Through and the Positive Urgency Measure. In addition, there was a non-significant trend towards a significant group difference on Feelings Trigger Action.
Associations Among Components of Inhibition and Impulsivity
Correlational analyses were used to assess the relationship between the inhibition measures (Antisaccade accuracy and RT, and Word-naming interference score) and impulsivity (Three-Factor Impulsivity scale and the Positive Urgency Measure). The bivariate relationship between the Word-naming data and the impulsivity measures was non-linear; therefore nonparametric correlations (Spearman’s rho) were used. As shown in Table 4, no significant relationship was found between accuracy on antisaccade trials (controlling for prosaccade accuracy) and Pervasive Influence of Feelings, (Lack of) Follow Through and the Positive Urgency Measure. Individuals
that obtained higher scores on Feelings Trigger Action performed significantly more poorly on the Antisaccade task; this relationship was curvilinear (antisaccade accuracy diminished as Feelings Trigger Action reached a higher range). In addition, there was a non-significant trend towards a significant association between accuracy on antisaccade trials (controlling for prosaccade accuracy) and the Positive Urgency Measure (p = .06). RT on antisaccade trials was not significantly related to any of the four impulsivity self-report scales controlling for prosaccade trials. Interference scores on the Word-naming were related to significantly higher scores on Feelings Trigger Action, but interference scores were not significantly related to Pervasive Influence of Feelings, (Lack of) Follow Through or the Positive Urgency Measure (see Table 4).
Current Depressive Symptoms as a Potential Confound
As noted above, only four participants met diagnostic criteria for current major depressive episodes. As would be expected, the IDS-SR30 score was significantly correlated with diagnostic group (N = 110, rs = .46, p = .00). To examine the role of subsyndromal depressive symptoms, correlational analyses were conducted to examine whether current depressive symptoms (IDS-SR30) relate to inhibition measures, including the accuracy on antisaccade trials controlling for prosaccade accuracy (N = 106, partial rs = -.05, p = .64), the RT on antisaccade trials controlling for prosaccade RT (N = 106, partial rs = .13, p = .19) or the interference score on the Word-naming task (N = 110, r = -.01, p = .96).
The data did not meet the assumption of homogeneity of regression slopes, therefore three separate ANCOVAs were used to test whether IDS-SR30 scores influenced the diagnostic group differences on the three impulsivity measures. Results showed no significant effect of diagnostic group on Pervasive Influence of Feelings (N = 110, F(1,106) = 1.63, p = .20, ηp = .02),
(Lack of) Follow Through (N = 110, F(1,106) = .03, p = .86, ηp = .00) or the Positive Urgency
Measure (N = 110, F(1,106) = .01, p = .91, ηp = .00), while controlling for IDS-SR30. Thus, the difference between the depression group and controls on the three impulsivity measures was influenced by current depressive symptoms.
Exploratory Analyses
Exploratory analyses were conducted to assess whether inhibition deficits were tied to the severity of depression. Only individuals that experienced two or more depressive episodes were included in the depression group.
Two one-way ANCOVAs were used to assess whether the diagnostic groups differed on antisaccade accuracy and RT, controlling for prosaccade accuracy and RT, respectively. Results showed no significant group difference on antisaccade accuracy (N depression group = 20, N controls = 62; F(1, 79) = .21, p = .65, ηp = .00) or antisaccade RT (F(1, 79) = 1.83, p = .18, ηp =
.02).
A one-way ANOVA was conducted to assess whether the diagnostic groups differed on the interference score on the Word-naming task. Results showed a significant difference between the depression group and controls (N depression group = 21, N controls = 63; F(1, 82) = 4.11, p < .05, ηp = .05). That is, the depression group (M = -47.53, SD = 26.95) showed significantly more
interference on the Word-naming task than controls (M = -34.13, SD = 25.97).
Discussion
A growing body of research indicates that depression is associated with impulsive responses to negative and positive affective states (e.g. d’Acremont & Van der Linden, 2007; Carver et al., 2013). Despite the emergent evidence, the underlying mechanisms driving emotion-relevant impulsivity are not yet well understood. The goal of the present study was to investigate whether impairments in cognitive control, particularly the inhibition of automatic responses and irrelevant information, could help explain why depressed individuals tend to be impulsive in emotional states. For this purpose, we administered the Antisaccade task, the Word-naming task and self-report measures of impulsivity in a sample of undergraduate students diagnosed with major depressive disorder (current or in partial or full remission) and nondepressed control participants.
Findings indicated that individuals with depression were more impulsive in both emotional and non-emotional states, compared to never-depressed individuals. In addition, emotion-related impulsivity was associated with difficulties inhibiting pre-potent responses and inhibiting irrelevant information. Furthermore, depressed individuals did not differ from controls in the ability to inhibit pre-potent responses and the ability to resist interference from distracting information.
Our results partially replicate previous literature, by showing that depressed individuals are more impulsive in negative (Pervasive Influence of Feelings) and positive (Positive Urgency) mood states (e.g. Carver et al., 2013), than non-depressed individuals. That is, depressed individuals tend to respond to negative emotions with fatigue and passivity, tend to generalize from a single negative event and let emotions influence general worldviews (Pervasive Influence of Feelings). When either negative or positive emotions are experienced, depressed individuals
tend to act rashly, without taking the consequences of their behavior into account (Carver et al., 2011). The results could not support the association between depression and the other emotion-related facet of impulsivity, Feelings Trigger Action (Carver et al., 2013; Johnson et al., 2013). This is surprising since Feelings Trigger Action explicitly captures impulsive reactivity to emotion. The trend towards significance indicates that this might be due to a slightly underpowered sample. Thus, these findings indicate that depressed individuals are characterized by a general over-responsiveness to emotion that is not only specific towards negativity and sadness, which has important implications. Impulsivity is currently not stated in the DSM-IV as one of the symptoms of major depressive disorder (APA, 2000). The current findings, however, suggest that impulsivity is an important characteristic of depression, which could be taken into account in diagnosis. In addition, attention is tailored towards negative affect in depression. This is important because an over-reactivity towards sad emotions has been related to increased suicide risk (Oquendo et al., 2004), and passive behavior following impulsivity can increase the chance of relapse (Bockting et al., 2006). Nevertheless, impulsivity following positive affect can also have destructive consequences (e.g. excessive alcohol use) (Karyadi & King, 2011). It is therefore important to target impulsivity towards both negative and positive mood states.
Impulsive tendencies in depression were not only bound to mood states. That is, compared to never-depressed controls, depressed individuals scored higher on Lack of Follow Through, which captures the tendency to be easily distracted, the inability to stay focused on difficult tasks and the lack of self-control. This is in line with studies showing that depressive symptoms are related to an inability to stay focused on difficult tasks (lack of perseverance) (Miller et al., 2003; d’Acremont et al., 2007) but in contrast with other studies revealing no association between depression and Lack of Follow Through (Carver et al., 2013; Johnson et al., 2013). One of the symptoms of major depressive disorder, as stated in the DSM-IV, is a diminished ability to think or concentrate (APA, 2000). It is therefore not surprising that depression was related to a Lack of Follow Through. In addition, in contrast with other studies (Carver et al., 2013; Grall-Bronnec et al., 2012), results regarding all impulsivity facets appeared tied to the level of current depressive symptoms, suggesting that impulsivity is a state marker, rather than a trait-like feature of depression.
The emotion-related facet of impulsivity, Feelings Trigger Action, but not the other impulsivity facets (Pervasive Influence of Feelings, Lack of Follow Through and Positive Urgency), was associated with inhibition deficits. That is, individuals that showed more impulsive tendencies when in a negative mood state performed more poorly on the Antisaccade and Word-naming task, an index of pre-potent response inhibition and resistance to DI,
respectively. Only individuals that scored relatively high on the impulsivity measure showed this profile, indicating that deficits in inhibition have to be severe enough to result in impulsive responses to emotion. These findings fit well with the theory of Carver et al., (2008), who argued that deficits in cognitive control are involved in emotion-related impulsivity. Current findings are also in line with previous research demonstrating that urgency is related to difficulties in inhibiting pre-potent responses (Gay et al., 2008; Gay et al., 2010; Rochat et al., 2013; Wilbertz et al., 2014; Gunn & Finn, 2015). In addition, previous literature is extended by showing that a different facet of inhibition, resistance to DI, is related to emotion-relevant impulsivity. Thus, these findings indicate that individuals that are impulsive in negative emotional states are characterized by lower inhibition capacities, which may make them less able to suppress rash actions.
Although the link between emotion-related impulsivity and inhibition deficits was supported, findings showed no specific inhibition deficits in depression. That is, individuals with depression did not differ from controls in their ability to inhibit pre-potent responses or the ability to resist interference from distracting information. This suggests that depressed individuals are capable of inhibiting pre-potent responses and are as effective as non-depressed individuals in blocking the entry of irrelevant material into short-term memory. These findings are in contrast with previous studies, which were able to differentiate dysphoric and depressed individuals from controls on the Antisaccade task (Sweeney, Strojwas, Mann & Thase, 1998; Derakshan et al., 2009; Carvalho et al., 2014) and several resistance to DI measures (Lau et al., 2007; Zetsche et al., 2012). The paradigms used in the current study to measure inhibition might form an explanation of why no group-differences were found. Specifically, in previous studies, impairments in pre-potent response inhibition (Derakshan et al., 2009; Mitterschiffthaler et al., 2008; Dai & Feng, 2011) and resistance to DI (Lau et al., 2007; Zetsche et al., 2012) have mainly been observed when processing emotional material, whereas the current study used a mood induction to prime (at a mild level) the negative mood state of a stressful life event, rather than relying on negative task stimuli. Including emotional-stimuli (instead of a mood induction) might yield different results. The effect of emotional-stimuli can be explained by differences in mood-congruent attentional processes between depressed and non-depressed individuals. Depression is characterized by an attentional bias towards stimuli relevant for depression (e.g. sad, negative stimuli), whereas healthy individuals tend to selectively avoid sad stimuli (Joormann & Gotlib, 2007; Hankin, Gibb, Abela, & Flory, 2010). Therefore, the lack of inclusion of emotion stimuli in our study may have obscured important attentional processes related to depression, which might form an explanation of why no group differences in inhibition were found.
Another reason of why no deficits in resistance to DI were found could be the type of task used to measure this aspect of inhibition. Previous research has documented depressive deficits in resistance to DI using a broad range of tasks (e.g. the Emotional Flanker Task, Prose Distraction Task and ‘suppress/ignore’ task), whereas the Word-naming task specifically has not been previously investigated in depressive samples. Although these tasks involve the same type of challenge, of resisting interference from distracting information, there are some differences between these tasks. For instance, in the Prose Distraction Task participants have to read a story aloud, while ignoring distracting (i.e. italicized) information in the story. In the Word-naming task participants only have to ignore a single word, while reading another word. Therefore, it might be the case that the Word-naming task was too easy, and that the use of a different task to assess resistance to DI yields different results.
The use of a sample consisting mainly of remitted depressed individuals could form another explanation of why no inhibition deficits were found in the current sample. Studies showing deficits in pre-potent response inhibition and resistance to DI are mainly conducted in currently depressed samples (e.g. Mitterschiffthaler et al., 2008; Snyders, 2012; Zetsche et al., 2012; Carvalho et al., 2014), and executive function deficits are found to be greater in individuals with more severe depressive symptoms (McClintock, Husain, Greer & Cullum, 2010). In addition, research with both remitted and currently depressed individuals found that only currently depressed individuals show inhibition deficits towards emotional material (Goeleven, De Raedt, Baert & Koster, 2006; Joormann & Gotlib, 2010). The current study found preliminary support for this by showing that individuals who experienced two or more depressive episodes, revealed more inhibition deficits from distractor interference than never-depressed controls. Therefore, it might be the case that inhibition deficits are specifically tied to the severity of depression.
In sum, depressed individuals show impulsive tendencies that seem tied to the level of subsyndromal depression. In addition, inhibition deficits (in both depressed and non-depressed individuals) are related to impulsivity in negative mood states, but the relation between depression and deficits in inhibition could not be supported. Together, these findings support the relation between depression and impulsivity, but further research is needed on the mechanisms underlying these impulsive tendencies.
Limitations
Several limitations of this study must be taken into account. The results of this study are based on a cross-sectional design; therefore no statements regarding causality can be drawn. As mentioned
above, this study was slightly underpowered, which could also form an explanation of why group differences in inhibition abilities were not found. Furthermore, although this study used clinical diagnoses to differentiate between depressed and non-depressed individuals, this study is conducted in a sample of undergraduate students taking psychology classes at a highly competitive school; this highly accomplished sample may experience fewer cognitive deficits and impulsive tendencies than would have been observed in a community sample. Therefore, findings of this study should be interpreted with caution when generalizing to the general population. In addition, it should be noted that pre-potent response inhibition and resistance to DI have been assessed in previous literature with a large number of different tests. This might provide an explanation for the inconsistent findings across studies, but also makes direct comparison of the results of this study to other literature more difficult. Finally, co-morbidity and the severity of psychopathology were not taken into account, which could have influenced the results. For instance, individuals with both depressive and manic episodes (bipolar disorder), show more executive impairments than unipolar depressed individuals (Stoddart, Craddock & Jones, 2007).
Recommendations for Future Research
Some recommendations for future research can be made. Friedman and Miyake (2004) argued that there is a third type of inhibition: resistance to pro-active interference (PI), also named ‘updating’ (i.e. resisting memory interference from information that was previously, but no longer relevant). Assessing three forms of inhibition in the current study would have been overly demanding, and could compromise the quality of the results. However, it is important for future research to investigate resistance to PI in relation to depression. That is, findings of several studies indicated that depression was associated with impairment in the suppression of previously memorized negative material (‘updating’) (resistance to PI) (Joormann & Gotlib, 2008; Joormann, Yoon & Zetsche, 2007; Joormann et al., 2010; Snyder, 2012), whereas several other studies did not find support for this finding (De Lissnyder, Koster, Derakshan & De Raedt, 2010; Zetsche et al., 2012). Although less direct, other research suggests the promise of continuing to examine resistance to PI. For instance, depressed individuals tend to experience a higher frequency of unwanted intrusive thoughts (Wenzlaff, Wegner & Roper, 1988) that can negatively influence task performance (Seibert & Ellis, 1991), and the efficacy of thought suppression has been related to resistance to PI (Friedman & Miyake, 2004). Taken together, these findings provide support for the idea that depression is related to resistance to PI, which could be an interesting avenue for future research.
In addition to investigating resistance to PI, it would be interesting to focus on rumination. Rumination has been related cross-sectionally to depression, but has also been shown to predict the onset, severity, and course of depression (Nolen-Hoeksema, Wisco & Lyubomirsky, 2008). Furthermore, specific inhibition deficits in resistance to PI have been found to relate to rumination even after controlling for depression (Zetsche et al., 2012). In addition, De Lissnyder et al., (2011) found that rumination, but not depression, is predictive of impairments in pre-potent response inhibition. It would therefore be interesting to investigate whether rumination plays a distinct role in the relation between depression and inhibition deficits.
Next to further investigating inhibition deficits as a potential underlying construct of emotion-related impulsivity, it would also be worthwhile to focus on emotion-regulation processes that could play a role. For instance, reappraisal (the reinterpretation of an emotional situation) has been described in the literature as an appropriate way to change the intensity of an emotional response (Gross, 1998). Emotion-regulation strategies (i.e. rumination and insufficient reappraisal) have been found to mediate the association between depression and urgency (d’Acremont & Van der linden, 2007). It therefore forms a promising avenue for future research to further investigate the influence of emotion-regulation strategies on emotion-related impulsivity in depression.
In addition, emotion-relevant impulsivity is mainly assessed by self-report measures in previous research (e.g. Johnson et al., 2013). Further research is needed that use behavioral measures of impulsivity, in the context of experimentally induced states of heightened affect, which would make findings more generalizable to daily situations. Related to this, little is known about specific behaviors that follow from urgency in depression. A recent meta-analysis showed that urgency is related to different types of impulsive behaviors, such as substance use, aggression, delinquency and gambling behavior (Sharma, Markon & Clark, 2014), but only excessive alcohol use and binge eating have been related to positive urgency in depressive samples so far (Karyadi & King, 2011; Smith et al., 2013). Therefore, further research should be conducted in depressive samples on the specific behaviors that follow impulsive tendencies in both positive and negative mood states.
It is also under-investigated how positive urgency relates to motivational systems in depression. As noted above, Carver et al., (2008) argued that deficits in executive processes combined with a reactive approach system can result in impulsive, rashly behavior, whereas the same poor executive oversight combined with a non-reactive approach system can result in inaction. There is emerging evidence that depression is associated with a blunted approach system, indicated by a low incentive sensitivity on behavioral tasks (Henriques & Davidson,
2000) and self-report measures (Campbell-Sills, Liverant, & Brown, 2004). However, it is not clear how a blunted approach system relates to positive urgency in depression, which would be an interesting avenue for future research.
Conclusion
This is the first study that investigates inhibition deficits as a possible underlying construct of impulsivity in a clinically depressed sample. Although the mediating role of inhibition deficits could not be confirmed, the current study replicated and extended previous research by showing that clinically depressed individuals are more impulsive in both emotional and non-emotional states, than non-depressed individuals. Furthermore, individuals that are impulsive in negative emotional states are characterized by lower pre-potent response capacities and show deficits in resistance to DI, which could help explain difficulties suppressing impulsive responses to emotions.
These findings have important implications for clinical settings. As noted above, when depressed individuals experience impulsivity in negative affective states that is tied to aggression, there is an increased risk for self-destructive behavior (i.e. self-harm and suicide) (Oquendo et al., 2004). Therefore, it might be valuable to integrate elements of interventions in therapy that specifically target these reflexive reactions towards negative emotions. For instance, implementation interventions, in which individuals develop a specific if-then plan of how to deal with an affective state, can be helpful in diminishing reflexive behavioral responses to emotion (Gollwitzer, 1999; Webb et al., 2012). Thus, this study indicates that impulsivity is an important characteristic of major depression.
References
American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed. tekst rev.). Washington, DC: Author.
Becona, E., Lorenzo, M. D. C., & Fuentes, M. J. (1996). Pathological gambling and depression. Psychological Reports, 78, 635-640.
Berg, J. M., Latzman, R. D., Bliwise, N. G., & Lilienfeld, S. O. (2015). Parsing the heterogeneity of impulsivity: A meta-analytic review of the behavioral implications of the UPPS for Psychopathology. Psychological Assessment. Advance online publication.
Bresin, K., Carter, D. L., & Gordon, K. H. (2012). The relationship between trait impulsivity, negative affective states, and urge for nonsuicidal self-injury: a daily diary study.
Psychiatry Research, 205, 227-231.
Bockting, C. L. H., Spinhoven, P., Koeter, M. W. J., Wouters, L. F., Visser, I., & Schene, A. H. (2006). Differential predictors of response to preventive cognitive therapy in recurrent depression: a 2-year prospective study. Psychotherapy and Psychosomatics, 75, 229-236. Campbell-Sills, L., Liverant, G. I., & Brown, T. A. (2004). Psychometric evaluation of the
behavioral inhibition/behavioral activation scales in a large sample of outpatients with anxiety and mood disorders. Psychological Assessment, 16, 244–254.
Carney, R. M., Blumenthal, J. A., Catellier, D., Freedland, K. E., Berkman, L. F., Watkins, L. L., Czajkowski, S. M., Hayano, J. & Jaffe, A. S. (2003). Depression as a risk factor for mortality after myocardinal infarction. American Journal of Cardiology, 92, 1277–1281. Carvalho, N., Noiret, N., Vandel, P., Monnin, J., Chopard, G., & Laurent, E. (2014). Saccadic eye
movements in depressed elderly patients. PloS ONE, 9, e105355.
Carver, C. S., Johnson, S. L., & Joormann, J. (2008). Serotonergic function, two-mode models of self regulation, and vulnerability to depression: what depression has in common with impulsive aggression. Psychological Bulletin, 134, 912–943.
Carver, C. S., Johnson, S. L., & Joormann, J. (2013). Major depressive disorder and impulsive reactivity to emotion: towards a dual-process view of depression. British Journal of
Clinical Psychology, 52, 285-299.
Carver, C. S., Johnson, S. L., Joormann, J., Kim, Y., & Nam, J. Y. (2011). Serotonin transporter polymorphism interacts with childhood adversity to predict aspects of impulsivity.
Psychological Science, 22, 589-595.
Clarke, D. (2006). Impulsivity as a mediator in the relationship between depression and problem gambling. Personality and Individual Differences, 40, 5-15.
Clarke, D. (2012). The moderating effect of impulsivity on the relationship between stressful life events and depression among college women. International Journal of Mental Health
and Addiction, 10, 152–161.
Corruble, E., Benyamina, A., Bayle, B., Falissard, B., & Hardy, P. (2003). Understanding impulsivity in severe depression? A psychometrical contribution. Progress in
Neuro-Psychopharmacology & Biological Psychiatry, 27, 829–833
Corruble, E., Legrand, J. M., Duret, C., Charles, G., & Guelfi, J. D. (1999). IDS-C and IDS-SR: psychometric properties in depressed in-patients. Journal of Affective Disorders, 56, 95-101.
Cyders, M. A., Smith, G. T., Spillane, N. S., Fischer, S., Annus, A. M., & Peterson, C. (2007). Integration of impulsivity and positive mood to predict risky behavior: development and validation of a measure of positive urgency. Psychological Assessment, 19, 107–118. Dai, Q., & Feng, Z. (2011). Deficient interference inhibition for negative stimuli in depression: an
event-related potential study. Clinical Neurophysiology, 122, 52-61.
d’Acremont, M., & Van der Linden, M. (2007). How is impulsivity related to depression in adolescence? Evidence from a French validation of the cognitive emotion regulation questionnaire. Journal of Adolescence, 30, 271–282.
De Lissnyder, E., Derakshan, N., De Raedt, R., & Koster, E. H. W. (2011). Depressive symptoms and cognitive control in a mixed antisaccade task: specific effects of depressive rumination. Cognition Emotion, 25, 886–897.
De Lissnyder, E., Koster, E. H., Derakshan, N., & De Raedt, R. (2010). The association between depressive symptoms and executive control impairments in response to emotional and non-emotional information. Cognition and Emotion, 24, 264-280.
Del Boca, F. K., Darkes, J., Greenbaum, P. E., & Goldman, M. S. (2004). Up close and personal: Temporal variability in the drinking of individual college students during their first year.
Journal of Consulting and Clinical Psychology, 72, 155–164.
Derakshan, N., Salt, M., & Koster, E. H. W. (2009). Attentional control in dysphoria: an investigation using the antisaccade task. Biological Psychology, 80, 251–255.
Derogatis, L. R. (1977). SCL-90R Version Manual I, Clinical Psychometrics Research Unit, John Hopkins University School of Medicine, Baltimore.
Dussault, F., Brendgen, M., Vitaro, F., Wanner, B., & Tremblay, R. E. (2011). Longitudinal links between impulsivity, gambling problems and depressive symptoms: a transactional model from adolescence to early adulthood. Comprehensive Psychiatry, 52, 520-524.
Dutton, D. G., & Karakanta, C. (2013). Depression as a risk marker for aggression: a critical review. Aggression and Violent Behavior, 18, 310-319.
Ekinci, O., Albayrak, Y., & Caykoylu, A. (2011). Impulsivity in euthymic patients with major depressive disorder: the relation to sociodemographic and clinical properties. The Journal
of Nervous and Mental Disease, 199, 454-458.
First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (1997). Structured clinical
interview for DSM-IV axis I disorders (SCID-I). New York: Biometric Research
Department.
Friedman, N. P., & Miyake, A. (2004). The relations among inhibition and interference control functions: A latent-variable analysis. Journal of Experimental Psychology: General, 133, 101-135.
Frings, C., Wentura, D., & Holtz, M. (2007). Dysphorics cannot ignore unpleasant information.
Cognition and Emotion, 21, 1525–1534.
Gay, P., Rochat, L., Billieux, J., d’Acremont, M., & Van der Linden, M. (2008). Heterogeneous inhibition processes involved in different facets of self-reported impulsivity: Evidence from a community sample. Acta Psychologica, 129, 332–339.
Gay, P., Courvoisier, D. S., Billieux, J., Rochat, L., Schmidt, R. E., & van der Linden, M. (2010). Can the distinction between intentional and unintentional interference control help differentiate varieties of impulsivity? Journal of Research in Personality, 44, 46-52. Goeleven, E., De Raedt, R., Baert, S., & Koster, E. H. W. (2006). Deficient inhibition of
emotional information in depression. Journal of Affective Disorders, 93, 149–157. Gollwitzer, P. M. (1999). Implementation intentions: strong effects of simple plans. American
Psychologist, 54, 493–503.
Goodman, J. T., Streiner, D. L., & Woodward, C. A. (1974). Test-retest reliability of the shipley institute of living scale: practice effects or random variation. Psychological Reports, 35, 351-354.
Grall-Bronnec, M., Wainstein, L., Feuillet, F., Bouju, G., Rocher, B., Vénisse, J. L., & Sébille-Rivain, V. (2012). Clinical profiles as a function of level and type of impulsivity in a sample group of at-risk and pathological gamblers seeking treatment. Journal of
Gambling Studies, 28, 239-252.
Granö, N., Keltikangas‐järvinen, L., Kouvonen, A., Virtanen, M., Elovainio, M., Vahtera, J., & Kivimäki, M. (2007). Impulsivity as a predictor for newly diagnosed depression. Health
and Disability, 48, 173-179.
consequences for experience, expression, and physiology. Journal of Personality and
Social Psychology, 74, 224-237.
Gruber, J., Johnson, S., Oveis, C., & Keltner, D. (2008). Risk for mania and positive emotional responding: too much of a good thing?. Emotion, 8, 23-33.
Gunn, R. L. & Finn, P. R. (2015). Applying a dual process model of self-regulation: The association between executive working memory capacity, negative urgency, and negative mood induction on pre-potent response inhibition. Personality and Individual
Differences, 75, 210-215.
Hankin, B. L., Gibb, B. E., Abela, J. R. Z., & Flory, K. (2010). Selective attention to affective stimuli and clinical depression among youth: Role of anxiety and specificity of emotion.
Journal of Abnormal Psychology, 119, 491-501.
Henna, E., Hatch, J. P., Nicoletti, M., Swann, A. C., Zunta-Soares, G., & Soares, J. C. (2013). Is impulsivity a common trait in bipolar and unipolar disorders? Bipolar Disorders, 15, 223-227.
Henriques, J. B., & Davidson, R. J. (2000). Decreased responsiveness toreward in depression.
Cognition and Emotion, 14, 711–724.
IBM Company (2013). Statistical Package for the Social Sciences, version 20.0. Armonk, NY:IBM Corporations.
Joormann, J., & Gotlib, I. H. (2007). Selective attention to emotional faces following recovery from depression. Journal of Abnormal Psychology, 116, 80-85.
Joormann, J., & Gotlib, I. H. (2008). Updating the contents of working memory in depression: Interference from irrelevant negative material. Journal of Abnormal
Psychology, 117, 182-192.
Joormann, j., & Gotlib, I. H. (2010). Emotion regulation in depression: relation to cognitive inhibition. Cognition and Emotion, 24, 281-298.
Joormann, J., Talbot, L., & Gotlib, I. H. (2007). Biased processing of emotional information in girls at risk for depression. Journal of Abnormal Psychology, 116, 135-143.
Joormann, J., Yoon, K. L., & Zetsche, U. (2007). Cognitive inhibition in depression. Applied and
Preventive Psychology, 12, 128–139.
Joormann, J., Nee, D. E., Berman, M. G., Jonides, J., & Gotlib, I. H. (2010). Interference resolution in major depression. Cognitive, Affective & Behavioral Neuroscience, 10, 21-33.
