- Email: [email protected]
Bias to negative emotions: A depression state-dependent marker in adolescent major depressive disorder
Psychiatry Research 198 (2012) 28–33
Contents lists available at SciVerse ScienceDirect
Psychiatry Research journal homepage: www.elsevier.com/locate/psychres
Bias to negative emotions: A depression state-dependent marker in adolescent major depressive disorder Fadi T. Maalouf a, b, c,⁎, Luke Clark d, e, Lucy Tavitian a, Barbara J. Sahakian e, f, David Brent b, c, Mary L. Phillips b, c a
Department of Psychiatry, American University of Beirut Medical Center, Lebanon Department of Psychiatry, University of Pittsburgh School of Medicine, Pittsburgh, PA, United States Western Psychiatric Institute and Clinic, Pittsburgh, PA, United States d Department of Experimental Psychology, University of Cambridge, United Kingdom e MRC/Wellcome Trust Behavioural and Clinical Neuroscience Institute, University of Cambridge, United Kingdom f Department of Psychiatry, University of Cambridge, United Kingdom b c
a r t i c l e
i n f o
Article history: Received 1 November 2011 Received in revised form 22 January 2012 Accepted 26 January 2012 Keywords: Emotion processing Adolescent Depression
a b s t r a c t The aim of the current research was to examine for the first time the extent to which bias to negative emotions in an inhibitory control paradigm is a state or trait marker in major depressive disorder (MDD) in adolescents. We administered the affective go/no go task which measures the ability to switch attention to or away from positive or negative emotional stimuli to 40 adolescents with MDD (20 in acute episode (MDDa) and 20 in remission (MDDr)) and 17 healthy controls (HC). MDDa were significantly faster on the shift to negative target blocks as compared to shift to positive target blocks while HC and MDDr displayed the opposite pattern as measured by an “emotional bias index” (EBI = latency (shift to negative targets)− latency (shift to positive targets)). There was also a trend for an effect of group on commission errors, suggesting more impulsive responding by MDDa than both MDDr and HC independently of stimulus valence throughout the task. Negative bias was not associated with depression severity or medication status. In conclusion, bias to negative emotional stimuli appears to be present in the acute stage of MDD and absent in remission suggesting that it is a depression state-specific marker of MDD in adolescents. Latency emerges as a better proxy of negative bias than commission errors and accuracy on this inhibitory control task in adolescents with MDD. © 2012 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Major depressive disorder (MDD) is a prevalent and recurrent condition in children and adolescents, and it is associated with substantial impairment as seen in difficulties in school, interpersonal relationships, tobacco and substance abuse, suicide attempts, and a 30-fold increased risk of completed suicide (Lewinsohn et al., 1998). There is strong evidence that MDD is associated with impaired emotion regulation, of which voluntary and involuntary (automatic) cognitive processes are key components (Mayberg, 2007; Phillips et al., 2008). Inhibitory control, which includes selective attention (i.e. to direct or redirect attention toward goal-related stimuli), and inhibition (i.e. to distract from goal-irrelevant stimuli) play significant roles in voluntary emotion regulation and are impaired in mood disorders, particularly in depression. Inhibitory control is involved in modulating the effect of ⁎ Corresponding author at: Department of Psychiatry, American University of Beirut Medical Center, PO Box: 11-0236, Riad El Solh 11072020, Lebanon. Tel.: + 961 1350000x5664; fax: + 961 1749209. E-mail address: [email protected] (F.T. Maalouf). 0165-1781/$ – see front matter © 2012 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.psychres.2012.01.030
emotional stimuli on behavior by redirecting attention to goaloriented information and inhibiting irrelevant information. These strategies are at the core of emotion regulation in MDD (Joormann and Gotlib, 2010). The go/no go paradigm has been traditionally used as a measure of inhibitory control. The addition of an affective component to this paradigm makes it relevant to mood disorders (Murphy et al., 1999; Ladouceur et al., 2005). The affective go/no go task specifically measures bias to emotional stimuli in an inhibitory control paradigm, a paradigm that is clinically relevant when working with depressed patients. Numerous studies have used the affective go/no go paradigm as a measure of bias to emotional stimuli in depressed adults and adolescents through contrasting reaction times for happy stimuli to reaction times for sad stimuli (Murphy et al., 1999; Erickson et al., 2005; Kyte et al., 2005; Kaplan et al., 2006; Rubinsztein et al., 2006). For example, depressed adults have been found to respond more rapidly to sad than happy word targets and omit more happy word targets than sad targets on this task (Murphy et al., 1999; Erickson et al., 2005), a negative bias that was independent of medication status (Erickson et al., 2005). In addition, acutely depressed and
F.T. Maalouf et al. / Psychiatry Research 198 (2012) 28–33
remitted adolescents, when studied as one group, showed a bias towards negatively valenced word stimuli as they were more accurate in their responses to sad targets than controls (Kyte et al., 2005). In the present study we aimed to use the affective go/no go task to determine the extent to which bias to negative emotional stimuli was present in adolescent MDD, and whether it was only present when subjects were asked to shift attention to a new emotional valence (i.e. when inhibitory control is mostly needed). Discriminating between state and trait markers of MDD may be the first step to help identify markers of response and/or guide choice of interventions (Mayberg et al., 1997; Gorlyn et al., 2008), in addition to identifying vulnerability for the illness. State markers, present only during the acute episodes of MDD, are more likely to help predict treatment response and guide treatment choice (Mayberg et al., 1997; Gorlyn et al., 2008). On the other hand, trait markers may represent vulnerability to, or lasting effect of the illness. Previous studies in pediatric MDD were not specifically designed to examine whether bias to negative emotions was a state or trait marker of the illness; to do so in this study, we included adolescent participants with MDD in acute and remitted states in separate groups and studied three groups of participants: acutely depressed (MDDa), remitted (MDDr) and healthy controls (HC). We hypothesized that MDDa would show a bias towards negative stimuli (i.e. they would be slower in responding to positive stimuli as compared to negative stimuli). Existing findings did not allow us to hypothesize whether this bias would only be present when subjects were asked to shift attention to a new emotional valence or whether it would be observed throughout the task. We also hypothesized that this impairment would not be observed in MDDr and HC. 2. Methods 2.1. Participants and measures The study protocol was approved by the University of Pittsburgh Institutional Review Board. Fifty-seven adolescents in total were recruited; these included 40 participants meeting criteria for major depressive disorder (MDD), current or past, according to Diagnostic and Statistical Manual for Mental Disorders (DSM-IV) diagnosed using the Kiddie Schedule for Affective Disorders Present and Lifetime version
29
(K-SADS-PL) (Kaufman et al., 1997) and 17 healthy control participants (HC) with no previous psychiatric history or psychiatric history in either biological parent. Out of the 40 participants with MDD, 20 were in an acute depressive episode (MDDa), defined by a Children Depression Rating Scale (CDRS) (Poznanski et al., 1984) score ≥ 40; and 20 were in remission (MDDr), defined by a CDRS ≤ 28 at the time of testing, a commonly used remission criterion for pediatric depression (Emslie et al., 2002). The groups were balanced for age, gender-ratio, pubertal development and IQ (Table 1). A full prorated Intelligence Quotient (IQ) score was obtained through two verbal and performance subscales (vocabulary and matrix respectively) of the Wechsler Intelligence Scale for Children—IV (WISC-IV). Participants also completed the Petersen Puberty Development Scale (PDS) (Petersen et al., 1988). Exclusion criteria included a history of head injury, neurological disorder (epilepsy, developmental disorder, loss of consciousness for more than 10 min), premorbid IQ estimateb 80, current psychotic symptoms, current history of alcohol and illicit substance abuse or dependence, and current or past history of attention deficit hyperactivity disorder (ADHD). All participants and their parents were made aware of the purpose of the study and signed informed consent to participate. 2.2. Affective go/no go task The affective go/no go task involves the rapid presentation of a series of words in the center of a computer screen. Each word is displayed for 300 ms, and there is a 900-ms interval between the words. These words have a “happy/positive” or “sad/ negative” valence. Participants are given a target valence and are asked to push the button on a press pad when they see a word that matches this valence, with the other valence acting as a distracter. There are 10 blocks of 18 stimuli each, with nine positive words (e.g., joyful, warmth) and nine negative words (e.g., mistake, burden) for a total of 45 negative and 45 positive words, presented randomly (Murphy et al., 1999). These are arranged in sequences of either happy or sad target valences as follows: happy– happy–sad–sad–happy–happy–sad–sad etc.… The first two blocks are considered practice blocks. These are followed by “four” shift conditions, where participants are required to inhibit responding to the previous target, and four non-shift conditions in which participants continue to respond to the same target valence as the previous block. Participants were informed of the shift as they were to begin a set-shift block. There are three dependent measures of interest: i) correct response latency or Response Time (RT), ii) total commission errors (responses to a distracter valence) and iii) total omission errors (non-responses to the target valence). For each commission error, a 500-ms/450-Hz tone sounds; participants do not get any feedback after an omission. Omission errors on this task may indicate a difficulty disengaging from the emotion at hand. For instance, if participants commit more omission errors on happy trials, it may mean that they are experiencing difficulty inhibiting the sad affect and attending to the positive stimulus (Erickson et al., 2005). While response accuracy may be a better gauge of automatic biases, response latency may be tapping into voluntary processes of emotion regulation (Lawson and MacLeod, 1999; Sears et al., 2011), which are the main focus of the current study. Commission errors, on the other hand, indicate impulsivity in responding. These are usually more relevant to manic states and are more automatic in nature (Fleck et al., 2011).
Table 1 Demographic and clinical data.
Age mean (S.D.) Female:male IQ mean (S.D.) PDS mean (S.D.) CDRS mean (S.D.)
Total MDD episode duration mean (S.D.)a Total number of MDD episodes mean (S.D.) Age of illness onset mean (S.D.)a Age of onset of current MDD mean (S.D.)a Age of offset of past MDD mean (S.D.)a Presence of lifetime anxiety disorders N (% Yes) Receiving SSRI/SNRI N (% Yes) a
In years.
HC N = 17
MDDr N = 20
MDDa N = 20
Statistics
15.2 (1.8)
15.4 (1.3)
15.3 (1.6)
F(2,56) = 0.095, p = 0.910
9:8 112 (11)
15:5 113 (12)
17:3 105 (11)
Chi-square (1,57) = 4.82, p = 0.090 F(2,56) = 2.938, p = 0.061
2.7 (0.8)
2.8 (0.5)
2.7 (0.4)
F(2,56) = 0.311, p = 0.734
19.1 (2.3)
23.7 (3.4)
58.5 (10.9)
1.6 (1.8)
2.1 (1.8)
F(2,56) = 189, p = 0.00 HC vs. MDDr, p = 0.139 HC vs. MDDa, p = 0.00 MDD vs. MDDr, p = 0.00 T (37) = − 0.789, p = 0.435
1.2 (0.5)
1.4 (0.6)
T (37) = − 1.366, d.f. = 37, p = 0.180
13.3 (2.0)
11.7 (3.0)
T (37) = 1.989, p = 0.054
13.6 (1.61) 15 (1.4) 10 (50%)
15 (75%)
Chi-square (1, 57) = 2.67, p = 0.102
13 (68%)
13 (68%)
Chi-square (1,57) = 0, p = 1.0
30
F.T. Maalouf et al. / Psychiatry Research 198 (2012) 28–33
2.3. Statistical analyses
3. Results
(F(2,53) = 6.11, p = 0.004) but not for the non-shift condition (F(2,53) = 0.71, p = 0.495). Here, while HC's and MDDr's latencies to negative stimuli were larger than their latencies to positive stimuli under the shift condition (Fig. 1), MDDa was the only group whose latency to negative stimuli was smaller than their latency to positive stimuli (Table 2.). Paired t-tests comparing latency shift-to-positive to latency shift-to-negative showed that this difference was statistically significant only in MDDa (t = 2.87, degree of freedom (d.f.) = 19, p = 0.01 for MDDa, t = − 0.91, d.f. = 19, p = 0.372 for MDDr and t = −1.62, d.f. = 16, p = 0.125 for HC). In order to explore this further, we calculated an “emotional bias index” (EBI) defined as (latency (shift to negative targets) − latency (shift to positive targets)). Positive EBI (i.e. latency to negative targets is larger than latency to positive targets) would suggest a positive bias and a negative EBI would suggest a negative bias. Three-group comparison revealed a significant effect of group on EBI (F(2,56) = 6.1, p = 0.004). Post-hoc pairwise comparisons showed that MDDa had a smaller EBI than both HC and MDDr, and there were no significant between -group differences on EBI for MDDr vs. HC (HC vs. MDDa: F(1,36) = 10.1, p = 0.003; HC vs. MDDr: F(1,36) = 0.89, p = 0.352 and MDDa vs. MDDr: F(1,39) = 6.64, p = 0.014) (Fig. 1).
Demographic and clinical data are reported in Table 1. All behavioral outcome measures are reported in Table 2.
3.2. Commission errors
We have applied 90% Winsorization (5% from each tail) on all outcome measures and calculated Winsorized means accordingly. Each dependent variable (omissions, commissions, latency) was then analyzed in a 2 × 2 × 3 repeated-measures analysis of covariance (ANCOVA) of ‘switch’ (switch, non-switch) by target valence (positive, negative), with group (MDDa, MDDr, HC) as a between-subjects variable and IQ as a covariate. To explore the triple interaction of valence × shift × group upon latency, we then conducted a repeated measure analysis under the shift and non-shift conditions separately and included IQ as a covariate followed by pairwise t-tests to compare latencies in each group to negative and positive stimuli. We performed exploratory analyses using Pearson's correlations between dependent variables that were significantly different among the three groups and relevant clinical data (CDRS, age of illness onset, total duration of depressive episodes, and total number of depressive episodes) in MDDa. Additional analyses were run to explore the potential confounding effects of medication, suicidality and comorbidities. T-tests were used to compare dependent variables that were abnormal in MDDa between those taking, vs. those not taking, selective serotonin reuptake inhibitors (SSRIs), those with vs. those without history of suicidality (separate comparisons for suicide attempts and ideation) and those with vs. those without comorbid anxiety disorders. For the three-group comparisons and exploratory correlation analyses in MDDa and MDDr, statistical significance was set at 0.05, and for post-hoc pair-wise comparisons, we have used a cut-off alpha of 0.016 to account for multiple comparisons.
3.1. Latency There was a statistically significant three-way interaction (shift × valence × group) upon latency F(2,53) = 3.31, p = 0.044. There was, however, no significant main effect of target valence (happy vs. sad) or shift condition (shift vs. non-shift) or group (HC vs. MDDa vs. MDDr). No significant interaction between valence and group or valence and shift condition upon latency (p > 0.05 for all) was found either. To further examine the above three-way interaction, we performed repeated measures analyses under shift and non-shift conditions separately with valence as a within-subjects variable and group as a between-subjects variable. There was a statistically significant valence × group interaction upon latency for the shift condition
There was a main effect of shift condition on commission errors (F(1,53) = 5.3, p = 0.025). Further analysis showed that all participants had more commission errors under the shift condition (mean: 7.58, standard deviation (S.D.): 5.17) as compared to the non-shift condition (mean: 7.33, S.D.: 6.20). This difference, however, was not statistically significant (p = 0.568). There was also a trend for an effect of group on commission errors (F(2,53) = 3.05, p = 0.056) indicating that MDDa had more commission errors overall than both HC and MDDr (mean: 11.5, S.D.:9.3 for HC; mean: 12.6, S.D. = 9.9 for MDDr; and mean: 201, S.D. = 11.6 for MDDa). There was, however, no significant effect of valence, or significant interaction between valence and shift condition, or valence and group, or valence and group and shift condition upon commission errors.
Table 2 Affective go/no go outcome measures by group.
Latency (shift to positive) mean (S.D.) Latency (non-shift positive) mean (S.D.) Latency (shift to negative) mean (S.D.) Latency (non-shift negative) mean (S.D.) Commission (shift to positive) mean (S.D.) Commission (non-shift positive) mean (S.D.) Commission (shift to negative) mean (S.D.) Commission (non-shift negative) mean (S.D.) Omission (shift to positive) mean (S.D.) Omission (non-shift positive) mean (S.D.) Omission (shift to negative) mean (S.D.) Omission (non-shift negative) mean (S.D.)
HC N = 15
MDDr N = 19
MDDa N = 19
462 (106)
478 (66)
488 (93)
470 (107)
485 (77)
473 (82)
486 (98)
486 (73)
451 (77)
488 (86)
481 (80)
483 (86)
2.8 (2.0)
3.2 (2.8)
4.8 (2.9)
2.4 (2.8)
3.4 (3.0)
4.8 (3.8)
3.1 (2.6)
2.9 (2.8)
5.7 (3.0)
3.2 (2.8)
3.0 (3.0)
4.8 (3.6)
4.2 (4.6)
2.6 (3.3)
3.2 (1.9)
4.5 (5.0)
2.9 (3.3)
2.7 (2.6.)
3.8 (3.7)
2.4 (2.7)
2.4 (2.4)
2.8 (3.1)
2.6 (3.2)
2.1 (2.5)
Fig. 1. Latency under the shift condition to negative and positive stimuli per group. Emotional bias index (EBI = latency (shift to negative targets) − latency (shift to positive targets)) across groups: HC (mean = 24.0, S.D.= 61.0), MDDr (mean = 8.2, S.D.= 40.0), MDDa (mean = −37.4, S.D.= 58.3). Three-group comparison: (F(2,56) = 6.1, p = 0.004). Post-hoc pairwise comparisons: MDDab HC (F(1,36)= 10.1, p = 0.003), MDDab MDDr (F(1,39) = 6.64, p = 0.014), and MDDr vs. HC (p > 0.05).
F.T. Maalouf et al. / Psychiatry Research 198 (2012) 28–33
3.3. Omission errors There was no significant effect of group, shift condition, valence or significant interaction between valence and shift condition, or valence and group, or valence and group and shift condition upon omissions. 3.4. Exploratory analysis There was no statistically significant correlation between EBI and depression severity as measured by CDRS scores, age of illness onset, total duration of depressive episodes, or number of depressive episodes in MDDa (p > 0.05 for all). EBI and total commission errors were not statistically different in MDDa between individuals with and without a history of suicide attempt, those with and without a history of suicidal ideation, those on and those off SSRIs, and those with and without a history of anxiety disorders (p > 0.05 for all). 4. Discussion The aim of this study was to examine the extent to which bias towards negative emotions was a state or trait marker in major depressive disorder (MDD) in adolescents. To do so, we assessed the performance of adolescent participants in a current MDD episode (MDDa), participants with MDD in remission (MDDr) and healthy controls (HC) on the affective go/no go task. We showed that bias towards emotionally salient stimuli with negative valence is a state marker of adolescent depression, present in MDDa and not in MDDr. Specifically, under the shift condition, three-group comparison showed that MDDa participants, when compared to the pattern observed in controls and MDDr, were faster on the blocks where target valence shifted to negative as compared to positive shifts. The fact that this difference was only observed under the shift condition may be due to the higher demand shift tasks place on dorsolateral prefrontal cortical regions to inhibit incongruent responses (Bermpohl et al., 2005). While non-shift blocks also demand that a previous incongruent response be inhibited, this process is more pronounced for shift blocks where the carryover effect of a previous congruent stimulus is not present. These results point to impaired inhibitory control processes in acutely depressed adolescents. Inhibitory control is a particularly effective regulatory process that helps depressed individuals to stop focusing on their mood-related thoughts, which maintain their negative mood (Joormann and Gotlib, 2007). Indeed, reduced inhibition when processing negative material has been proposed to underlie self-reported rumination in depression (Joormann and Gotlib, 2010), which has been associated with a poor course for major depressive episodes (Kuehner and Weber, 1999). Recent evidence also shows that impaired inhibitory control predicts the maintenance of depressive symptoms and rumination over a period of 6 months in healthy individuals (Zetsche and Joormann, 2011). It has also been linked to suicidal thinking in individuals in this population, since it is believed that individuals think of suicide as a solution to their problems, because they cannot think of any alternative course of action (Miranda and Nolen-Hoeksema, 2007). Other studies of MDD in adolescents have shown more errors on happy targets than sad targets in a depressed cohort with recent diagnosis of depression, and fewer errors on sad targets as compared to controls, but no group differences were detected on response latency (Kyte et al., 2005). The differences between the findings observed on the dependent measures of response latency in our study and response accuracy in Kyte et al.'s (2005) study may be due to a number of methodological factors. For example, in our study, depressed and remitted depressed adolescents were tested as separate groups and compared to controls. In Kyte et al.'s (2005) study, the case group included both depressed and remitted patients. Another difference was the inclusion of only first episode depressed adolescents by
31
Kyte et al. (2005), while our study did not select the sample based on number of depressive episodes. Finally, while response accuracy is a measure of automatic biases, response latency, especially at longer inter stimulus interval presentations, may be a better measure of voluntary processes of emotion regulation (Lawson and MacLeod, 1999; Sears et al., 2011), which was the main focus of the current study. Findings regarding commission rates indicated that, relative to the HC and MDDr groups, MDDa were more likely to respond to false targets regardless of the target's emotional valence, although this difference fell short of statistical significance. This suggests that the high commission rate in MDDa relative to HC and MDDr may be driven by impulsive responding rather than a bias towards certain emotions. This finding also replicates our previous findings of impulsivity emerging as a state-dependent marker for depression in adolescents (Maalouf et al., 2011), which is also consistent with other studies exploring impulsivity in children with MDD (Palladino et al., 1997; Kyte et al., 2005). These findings also suggest that when examining biases towards emotional stimuli using behavioral tasks in pediatric MDD, latency, rather than commissions, may be a more reliable outcome measure to differentiate between response styles to positively and negatively valenced stimuli as commission errors tap into more automatic processes of emotion regulation. Our data also provide some evidence for a lack of impairment in the processing of emotional stimuli in remitted adolescents with MDD. The lack of negative bias in our remitted group is in contrast to most adult studies that showed residual impairment in the period of remission in processing of emotional stimuli (Leppanen et al., 2004; Atchley et al., 2007; LeMoult et al., 2009) and bias towards sad stimuli (Joormann and Gotlib, 2007). This is consistent, however, with a few adult studies showing that remitted depressed participants do not differ from healthy controls in their inhibition of negative material (Joormann and Gotlib, 2010), and with other pediatric studies showing that on a recall task, remitted children showed no attentional bias to negative information (Timbremont and Braet, 2004). The discrepancy between our findings and the adult findings may be due to the fact that our paradigm did not include a priming strategy such as negative mood induction preceding the affective go/no go task. While some adult studies have been able to detect negative bias in remitted individuals without a priming manipulation (Joormann and Gotlib, 2007), most studies suggest that negative mood induction, simulating real life stressful events, is needed in order to detect the negative bias in remitted individuals (Scher et al., 2005; Ramel et al., 2007). Future studies that utilize mood priming paradigms are therefore necessary to either replicate or refute our findings under these settings. Alternatively, the trait-related negative bias in depression in adults may be due to the cumulative effect of recurrent depressive episodes, a potential “scar effect”, which our adolescent sample did not have, as the average number of episodes was 1.6 in MDDr (Robinson and Sahakian, 2008). This also highlights the importance of early intervention in order to prevent recurrence of episodes and potential development of a persistent negative bias. One of the strengths of our study is that our sample was well categorized into acute and remitted MDD subgroups, and we excluded adolescents with history of ADHD which, if present, could have impacted response style on our chosen behavioral task. This could, however, limit the generalizability of our findings, since ADHD is a prevalent comorbid disorder in adolescents with MDD in clinical practice (Biederman et al., 1995). A limitation is the cross-sectional design of the present study, which precludes us from drawing strong conclusions regarding the subsidence lack of negative bias with remission of depressive symptoms. Future studies should therefore aim at examining emotional bias during acute episode and in remission in the same sample of adolescents with MDD. It is also to be noted that the effect sizes that we observed for the differences between MDDa and HC and MDDr and MDDa on EBI were large (Cohen's d = 1) and small (Cohen d = 0.3) respectively. When
32
F.T. Maalouf et al. / Psychiatry Research 198 (2012) 28–33
considering our null findings for MDDr vs. HC on the EBI, we did also have adequate statistical power to detect any large effect size between MDDr vs. HC on EBI given our sample size. The statistical power for this comparison given a threshold of p b 0.05 and an effect size comparable to the effect size observed in our main finding – MDDa vs. HC – (Cohen's d = 1) would be 80%. A larger sample size would, however, be needed to detect smaller effect sizes between MDDr and HC with adequate statistical power. In summary, this is the first study to examine the extent to which bias towards negative emotional stimuli in an inhibitory control paradigm represents a state vs. a trait marker of depression in adolescents. We showed that bias towards negative emotions, as measured by the affective go/no go task, appeared to be present in the acute stage of MDD and absent in remission, suggesting that it is a state-specific marker in adolescent depression. Studies have shown that antidepressants modulate emotion processing much earlier than the effect on mood appears (Harmer, 2008). Given that the behavioral task that we used in this study can be easily employed in different clinical settings, it shows promise as a task to identify state-specific markers in MDD that may serve as early predictors of treatment response in MDD. Disclosure of interest Dr. Maalouf is on the speaker bureau of Eli Lilly. Dr. Brent receives research support from the National Institute of Mental Health, receives royalties from Guilford Press and is UpToDate Psychiatry Editor. Dr. Luke Clark consults for Cambridge Cognition plc. Dr. Sahakian consults for Cambridge Cognition. She has consulted for Novartis, Shire, GlaxoSmithKline, Lilly Boehringer-Ingelheim and F Hoffman-La Roche Ltd. She has also received honoraria for Grand Rounds in Psychiatry at Massachusetts General Hospital (CME credits) (Boston, 27 April 2007) and for speaking at the International Conference on Cognitive Dysfunction in Schizophrenia and Mood Disorders: Clinical Aspects, Mechanisms and Therapy (Brescia, 17–19 January 2007). She was on the Medical Research Council Neurosciences and Mental Health Board (2010) and on the Science Coordination Team for the Foresight Project on Mental Capital and Wellbeing, 2008 (Office of Science, The Department of Innovation, Universities and Skills,). She is currently on Panel LS5 for the European Research Council. As an Associate Editor, she also receives an honorarium from Psychological Medicine. Dr. Phillips and Ms. Tavitian report no biomedical financial interests or potential conflicts of interest. Acknowledgment This study was funded by the American Academy of Child and Adolescent Psychiatry — Eli Lilly Pilot Research Award and the Junior Faculty Scholar Program at the University of Pittsburgh/Western Psychiatric Institute and Clinic (R25 MH060473-10 Pilkonis (PI)). References Atchley, R.A., Stringer, R., Mathias, E., Ilardi, S.S., Diane Minatrea, A., 2007. The right hemisphere's contribution to emotional word processing in currently depressed, remitted depressed, and never-depressed individuals. Journal of Neurolinguistics 20 (2), 145–160. Bermpohl, F., Fregni, F., Boggio, P.S., Thut, G., Northoff, G., Otachi, P.T.M., Rigonatti, S.P., Marcolin, M.A., Pascual-Leone, A., 2005. Left prefrontal repetitive transcranial magnetic stimulation impairs performance in affective go/no-go task. NeuroReport: For Rapid Communication of Neuroscience Research 16 (6), 615–619. Biederman, J., Faraone, S., Mick, E., Lelon, E., 1995. Psychiatric comorbidity among referred juveniles with major depression: fact or artifact? Journal of the American Academy of Child and Adolescent Psychiatry 34 (5), 579–590. Emslie, G.J., Heiligenstein, J.H., Wagner, K.D., Hoog, S.L., Ernest, D.E., Brown, E., Nilsson, M., Jacobsen, J.G., 2002. Fluoxetine for acute treatment of depression in children and adolescents: a placebo-controlled randomized clinical trial. Journal of the American Academy of Child and Adolescent Psychiatry 41 (10), 1205–1214. Erickson, K., Drevets, W.C., Clark, L., Cannon, D.M., Bain, E.E., Zarate Jr., C.A., Charney, D.S., Sahakian, B.J., 2005. Mood-congruent bias in affective go/no-go performance
of unmedicated patients with major depressive disorder. The American Journal of Psychiatry 162 (11), 2171–2173. Fleck, D.E., Kotwal, R., Eliassen, J.C., Lamy, M., Delbello, M.P., Adler, C.M., Durling, M., Cerullo, M.A., Strakowski, S.M., 2011. Preliminary evidence for increased frontosubcortical activation on a motor impulsivity task in mixed episode bipolar disorder. Journal of Affective Disorders 133 (1–2), 333–339. Gorlyn, M., Keilp, J., Grunebaum, M., Taylor, B., Oquendo, M., Bruder, G., Stewart, J., Zalsman, G., Mann, J., 2008. Neuropsychological characteristics as predictors of SSRI treatment response in depressed subjects. Journal of Neural Transmission 115 (8), 1213–1219. Harmer, C.J., 2008. Serotonin and emotional processing: does it help explain antidepressant drug action? Neuropharmacology 55 (6), 1023–1028. Joormann, J., Gotlib, I.H., 2007. Selective attention to emotional faces following recovery from depression. Journal of Abnormal Psychology 116 (1), 80–85. Joormann, J., Gotlib, I.H., 2010. Emotion regulation in depression: relation to cognitive inhibition. Cognitive Emotion 24 (2), 281–298. Kaplan, J.S., Erickson, K., Luckenbaugh, D.A., Weiland-Fiedler, P., Geraci, M., Sahakian, B.J., Charney, D., Drevets, W.C., Neumeister, A., 2006. Differential performance on tasks of affective processing and decision-making in patients with panic disorder and panic disorder with comorbid major depressive disorder. Journal of Affective Disorders 95 (1–3), 165–171. Kaufman, J., Birmaher, B., Brent, D., Rao, U., Flynn, C., Moreci, P., Williamson, D., Ryan, N., 1997. Schedule for Affective Disorders and Schizophrenia for School-Age Children—Present and Lifetime Version (K-SADS-PL): initial reliability and validity data. Journal of the American Academy of Child and Adolescent Psychiatry 36 (7), 980–988. Kuehner, C., Weber, I., 1999. Responses to depression in unipolar depressed patients: an investigation of Nolen–Hoeksema's response styles theory. Psychological Medicine: A Journal of Research in Psychiatry and the Allied Sciences 29 (6), 1323–1333. Kyte, Z.A., Goodyer, I.M., Sahakian, B.J., 2005. Selected executive skills in adolescents with recent first episode major depression. Journal of Child Psychology and Psychiatry 46 (9), 995–1005. Ladouceur, C.D., Dahl, R.E., Williamson, D.E., Birmaher, B., Ryan, N.D., Casey, B.J., 2005. Altered emotional processing in pediatric anxiety, depression, and comorbid anxiety–depression. Journal of Abnormal Child Psychology 33 (2), 165–177. Lawson, C., MacLeod, C., 1999. Depression and the interpretation of ambiguity. Behaviour Research and Therapy 37 (5), 463–474. LeMoult, J., Joormann, J., Sherdell, L., Wright, Y., Gotlib, I.H., 2009. Identification of emotional facial expressions following recovery from depression. Journal of Abnormal Psychology 118 (4), 828–833. Leppanen, J.M., Milders, M., Bell, J.S., Terriere, E., Hietanen, J.K., 2004. Depression biases the recognition of emotionally neutral faces. Psychiatry Research 128 (2), 123–133. Lewinsohn, P.M., Rohde, P., Seeley, J.R., 1998. Major depressive disorder in older adolescents: prevalence, risk factors, and clinical implications. Clinical Psychology Review 18 (7), 765–794. Maalouf, F.T., Brent, D., Clark, L., Tavitian, L., McHugh, R.M., Sahakian, B.J., Phillips, M.L., 2011. Neurocognitive impairment in adolescent major depressive disorder: state vs. trait illness markers. Journal of Affective Disorders 133 (3), 625–632. Mayberg, H.S., 2007. Defining the neural circuitry of depression: toward a new nosology with therapeutic implications. Biological Psychiatry 61 (6), 729–730. Mayberg, H.S., Brannan, S.K., Mahurin, R.K., Jerabek, P.A., Brickman, J.S., Tekell, J.L., Silva, J.A., McGinnis, S., Glass, T.G., Martin, C.C., Fox, P.T., 1997. Cingulate function in depression: a potential predictor of treatment response. Neuroreport 8 (4), 1057–1061. Miranda, R., Nolen-Hoeksema, S., 2007. Brooding and reflection: rumination predicts suicidal ideation at 1-year follow-up in a community sample. Behaviour Research and Therapy 45 (12), 3088–3095. Murphy, F.C., Sahakian, B.J., Rubinsztein, J.S., Michael, A., Rogers, R.D., Robbins, T.W., Paykel, E.S., 1999. Emotional bias and inhibitory control processes in mania and depression. Psychological Medicine: A Journal of Research in Psychiatry and the Allied Sciences 29 (6), 1307–1321. Palladino, P., Poli, P., Masi, G., Marcheschi, M., 1997. Impulsive–reflective cognitive style, metacognition, and emotion in adolescence. Perceptual and Motor Skills 84 (1), 47–57. Petersen, A.C., Crockett, L., Richards, M., Boxer, A., 1988. A self-report measure of pubertal status: reliability, validity, and initial norms. Journal of Youth and Adolescence 17 (2), 117–133. Phillips, M.L., Ladouceur, C.D., Drevets, W.C., 2008. A neural model of voluntary and automatic emotion regulation: implications for understanding the pathophysiology and neurodevelopment of bipolar disorder. Molecular Psychiatry 13 (9), 833–857 (829). Poznanski, E.O., Grossman, J.A., Buchsbaum, Y., Banegas, M., Freeman, L., Gibbons, R., 1984. Preliminary studies of the reliability and validity of the children's depression rating scale. Journal of the American Academy of Child Psychiatry 23 (2), 191–197. Ramel, W., Goldin, P.R., Eyler, L.T., Brown, G.G., Gotlib, I.H., McQuaid, J.R., 2007. Amygdala reactivity and mood-congruent memory in individuals at risk for depressive relapse. Biological Psychiatry 61 (2), 231–239. Robinson, O.J., Sahakian, B.J., 2008. Recurrence in major depressive disorder: a neurocognrtive perspective. Psychological Medicine: A Journal of Research in Psychiatry and the Allied Sciences 38 (3), 315–318. Rubinsztein, J.S., Michael, A., Underwood, B.R., Tempest, M., Sahakian, B.J., 2006. Impaired cognition and decision-making in bipolar depression but no ‘affective bias’ evident. Psychological Medicine: A Journal of Research in Psychiatry and the Allied Sciences 36 (5), 629–639.
F.T. Maalouf et al. / Psychiatry Research 198 (2012) 28–33 Scher, C.D., Ingram, R.E., Segal, Z.V., 2005. Cognitive reactivity and vulnerability: empirical evaluation of construct activation and cognitive diatheses in unipolar depression. Clinical Psychology Review 25 (4), 487–510. Sears, C.R., Suzie Bisson, M.A., Nielsen, K.E., 2011. Dysphoria and the immediate interpretation of ambiguity: evidence for a negative interpretive bias in error rates but not response latencies. Cognitive Therapy and Research 35 (5), 469–476.
33
Timbremont, B., Braet, C., 2004. Cognitive vulnerability in remitted depressed children and adolescents. Behaviour Research and Therapy 42 (4), 423–437. Zetsche, U., Joormann, J., 2011. Components of interference control predict depressive symptoms and rumination cross-sectionally and at six months follow-up. Journal of Behavior Therapy and Experimental Psychiatry 42 (1), 65–73.
Contents lists available at SciVerse ScienceDirect
Psychiatry Research journal homepage: www.elsevier.com/locate/psychres
Bias to negative emotions: A depression state-dependent marker in adolescent major depressive disorder Fadi T. Maalouf a, b, c,⁎, Luke Clark d, e, Lucy Tavitian a, Barbara J. Sahakian e, f, David Brent b, c, Mary L. Phillips b, c a
Department of Psychiatry, American University of Beirut Medical Center, Lebanon Department of Psychiatry, University of Pittsburgh School of Medicine, Pittsburgh, PA, United States Western Psychiatric Institute and Clinic, Pittsburgh, PA, United States d Department of Experimental Psychology, University of Cambridge, United Kingdom e MRC/Wellcome Trust Behavioural and Clinical Neuroscience Institute, University of Cambridge, United Kingdom f Department of Psychiatry, University of Cambridge, United Kingdom b c
a r t i c l e
i n f o
Article history: Received 1 November 2011 Received in revised form 22 January 2012 Accepted 26 January 2012 Keywords: Emotion processing Adolescent Depression
a b s t r a c t The aim of the current research was to examine for the first time the extent to which bias to negative emotions in an inhibitory control paradigm is a state or trait marker in major depressive disorder (MDD) in adolescents. We administered the affective go/no go task which measures the ability to switch attention to or away from positive or negative emotional stimuli to 40 adolescents with MDD (20 in acute episode (MDDa) and 20 in remission (MDDr)) and 17 healthy controls (HC). MDDa were significantly faster on the shift to negative target blocks as compared to shift to positive target blocks while HC and MDDr displayed the opposite pattern as measured by an “emotional bias index” (EBI = latency (shift to negative targets)− latency (shift to positive targets)). There was also a trend for an effect of group on commission errors, suggesting more impulsive responding by MDDa than both MDDr and HC independently of stimulus valence throughout the task. Negative bias was not associated with depression severity or medication status. In conclusion, bias to negative emotional stimuli appears to be present in the acute stage of MDD and absent in remission suggesting that it is a depression state-specific marker of MDD in adolescents. Latency emerges as a better proxy of negative bias than commission errors and accuracy on this inhibitory control task in adolescents with MDD. © 2012 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Major depressive disorder (MDD) is a prevalent and recurrent condition in children and adolescents, and it is associated with substantial impairment as seen in difficulties in school, interpersonal relationships, tobacco and substance abuse, suicide attempts, and a 30-fold increased risk of completed suicide (Lewinsohn et al., 1998). There is strong evidence that MDD is associated with impaired emotion regulation, of which voluntary and involuntary (automatic) cognitive processes are key components (Mayberg, 2007; Phillips et al., 2008). Inhibitory control, which includes selective attention (i.e. to direct or redirect attention toward goal-related stimuli), and inhibition (i.e. to distract from goal-irrelevant stimuli) play significant roles in voluntary emotion regulation and are impaired in mood disorders, particularly in depression. Inhibitory control is involved in modulating the effect of ⁎ Corresponding author at: Department of Psychiatry, American University of Beirut Medical Center, PO Box: 11-0236, Riad El Solh 11072020, Lebanon. Tel.: + 961 1350000x5664; fax: + 961 1749209. E-mail address: [email protected] (F.T. Maalouf). 0165-1781/$ – see front matter © 2012 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.psychres.2012.01.030
emotional stimuli on behavior by redirecting attention to goaloriented information and inhibiting irrelevant information. These strategies are at the core of emotion regulation in MDD (Joormann and Gotlib, 2010). The go/no go paradigm has been traditionally used as a measure of inhibitory control. The addition of an affective component to this paradigm makes it relevant to mood disorders (Murphy et al., 1999; Ladouceur et al., 2005). The affective go/no go task specifically measures bias to emotional stimuli in an inhibitory control paradigm, a paradigm that is clinically relevant when working with depressed patients. Numerous studies have used the affective go/no go paradigm as a measure of bias to emotional stimuli in depressed adults and adolescents through contrasting reaction times for happy stimuli to reaction times for sad stimuli (Murphy et al., 1999; Erickson et al., 2005; Kyte et al., 2005; Kaplan et al., 2006; Rubinsztein et al., 2006). For example, depressed adults have been found to respond more rapidly to sad than happy word targets and omit more happy word targets than sad targets on this task (Murphy et al., 1999; Erickson et al., 2005), a negative bias that was independent of medication status (Erickson et al., 2005). In addition, acutely depressed and
F.T. Maalouf et al. / Psychiatry Research 198 (2012) 28–33
remitted adolescents, when studied as one group, showed a bias towards negatively valenced word stimuli as they were more accurate in their responses to sad targets than controls (Kyte et al., 2005). In the present study we aimed to use the affective go/no go task to determine the extent to which bias to negative emotional stimuli was present in adolescent MDD, and whether it was only present when subjects were asked to shift attention to a new emotional valence (i.e. when inhibitory control is mostly needed). Discriminating between state and trait markers of MDD may be the first step to help identify markers of response and/or guide choice of interventions (Mayberg et al., 1997; Gorlyn et al., 2008), in addition to identifying vulnerability for the illness. State markers, present only during the acute episodes of MDD, are more likely to help predict treatment response and guide treatment choice (Mayberg et al., 1997; Gorlyn et al., 2008). On the other hand, trait markers may represent vulnerability to, or lasting effect of the illness. Previous studies in pediatric MDD were not specifically designed to examine whether bias to negative emotions was a state or trait marker of the illness; to do so in this study, we included adolescent participants with MDD in acute and remitted states in separate groups and studied three groups of participants: acutely depressed (MDDa), remitted (MDDr) and healthy controls (HC). We hypothesized that MDDa would show a bias towards negative stimuli (i.e. they would be slower in responding to positive stimuli as compared to negative stimuli). Existing findings did not allow us to hypothesize whether this bias would only be present when subjects were asked to shift attention to a new emotional valence or whether it would be observed throughout the task. We also hypothesized that this impairment would not be observed in MDDr and HC. 2. Methods 2.1. Participants and measures The study protocol was approved by the University of Pittsburgh Institutional Review Board. Fifty-seven adolescents in total were recruited; these included 40 participants meeting criteria for major depressive disorder (MDD), current or past, according to Diagnostic and Statistical Manual for Mental Disorders (DSM-IV) diagnosed using the Kiddie Schedule for Affective Disorders Present and Lifetime version
29
(K-SADS-PL) (Kaufman et al., 1997) and 17 healthy control participants (HC) with no previous psychiatric history or psychiatric history in either biological parent. Out of the 40 participants with MDD, 20 were in an acute depressive episode (MDDa), defined by a Children Depression Rating Scale (CDRS) (Poznanski et al., 1984) score ≥ 40; and 20 were in remission (MDDr), defined by a CDRS ≤ 28 at the time of testing, a commonly used remission criterion for pediatric depression (Emslie et al., 2002). The groups were balanced for age, gender-ratio, pubertal development and IQ (Table 1). A full prorated Intelligence Quotient (IQ) score was obtained through two verbal and performance subscales (vocabulary and matrix respectively) of the Wechsler Intelligence Scale for Children—IV (WISC-IV). Participants also completed the Petersen Puberty Development Scale (PDS) (Petersen et al., 1988). Exclusion criteria included a history of head injury, neurological disorder (epilepsy, developmental disorder, loss of consciousness for more than 10 min), premorbid IQ estimateb 80, current psychotic symptoms, current history of alcohol and illicit substance abuse or dependence, and current or past history of attention deficit hyperactivity disorder (ADHD). All participants and their parents were made aware of the purpose of the study and signed informed consent to participate. 2.2. Affective go/no go task The affective go/no go task involves the rapid presentation of a series of words in the center of a computer screen. Each word is displayed for 300 ms, and there is a 900-ms interval between the words. These words have a “happy/positive” or “sad/ negative” valence. Participants are given a target valence and are asked to push the button on a press pad when they see a word that matches this valence, with the other valence acting as a distracter. There are 10 blocks of 18 stimuli each, with nine positive words (e.g., joyful, warmth) and nine negative words (e.g., mistake, burden) for a total of 45 negative and 45 positive words, presented randomly (Murphy et al., 1999). These are arranged in sequences of either happy or sad target valences as follows: happy– happy–sad–sad–happy–happy–sad–sad etc.… The first two blocks are considered practice blocks. These are followed by “four” shift conditions, where participants are required to inhibit responding to the previous target, and four non-shift conditions in which participants continue to respond to the same target valence as the previous block. Participants were informed of the shift as they were to begin a set-shift block. There are three dependent measures of interest: i) correct response latency or Response Time (RT), ii) total commission errors (responses to a distracter valence) and iii) total omission errors (non-responses to the target valence). For each commission error, a 500-ms/450-Hz tone sounds; participants do not get any feedback after an omission. Omission errors on this task may indicate a difficulty disengaging from the emotion at hand. For instance, if participants commit more omission errors on happy trials, it may mean that they are experiencing difficulty inhibiting the sad affect and attending to the positive stimulus (Erickson et al., 2005). While response accuracy may be a better gauge of automatic biases, response latency may be tapping into voluntary processes of emotion regulation (Lawson and MacLeod, 1999; Sears et al., 2011), which are the main focus of the current study. Commission errors, on the other hand, indicate impulsivity in responding. These are usually more relevant to manic states and are more automatic in nature (Fleck et al., 2011).
Table 1 Demographic and clinical data.
Age mean (S.D.) Female:male IQ mean (S.D.) PDS mean (S.D.) CDRS mean (S.D.)
Total MDD episode duration mean (S.D.)a Total number of MDD episodes mean (S.D.) Age of illness onset mean (S.D.)a Age of onset of current MDD mean (S.D.)a Age of offset of past MDD mean (S.D.)a Presence of lifetime anxiety disorders N (% Yes) Receiving SSRI/SNRI N (% Yes) a
In years.
HC N = 17
MDDr N = 20
MDDa N = 20
Statistics
15.2 (1.8)
15.4 (1.3)
15.3 (1.6)
F(2,56) = 0.095, p = 0.910
9:8 112 (11)
15:5 113 (12)
17:3 105 (11)
Chi-square (1,57) = 4.82, p = 0.090 F(2,56) = 2.938, p = 0.061
2.7 (0.8)
2.8 (0.5)
2.7 (0.4)
F(2,56) = 0.311, p = 0.734
19.1 (2.3)
23.7 (3.4)
58.5 (10.9)
1.6 (1.8)
2.1 (1.8)
F(2,56) = 189, p = 0.00 HC vs. MDDr, p = 0.139 HC vs. MDDa, p = 0.00 MDD vs. MDDr, p = 0.00 T (37) = − 0.789, p = 0.435
1.2 (0.5)
1.4 (0.6)
T (37) = − 1.366, d.f. = 37, p = 0.180
13.3 (2.0)
11.7 (3.0)
T (37) = 1.989, p = 0.054
13.6 (1.61) 15 (1.4) 10 (50%)
15 (75%)
Chi-square (1, 57) = 2.67, p = 0.102
13 (68%)
13 (68%)
Chi-square (1,57) = 0, p = 1.0
30
F.T. Maalouf et al. / Psychiatry Research 198 (2012) 28–33
2.3. Statistical analyses
3. Results
(F(2,53) = 6.11, p = 0.004) but not for the non-shift condition (F(2,53) = 0.71, p = 0.495). Here, while HC's and MDDr's latencies to negative stimuli were larger than their latencies to positive stimuli under the shift condition (Fig. 1), MDDa was the only group whose latency to negative stimuli was smaller than their latency to positive stimuli (Table 2.). Paired t-tests comparing latency shift-to-positive to latency shift-to-negative showed that this difference was statistically significant only in MDDa (t = 2.87, degree of freedom (d.f.) = 19, p = 0.01 for MDDa, t = − 0.91, d.f. = 19, p = 0.372 for MDDr and t = −1.62, d.f. = 16, p = 0.125 for HC). In order to explore this further, we calculated an “emotional bias index” (EBI) defined as (latency (shift to negative targets) − latency (shift to positive targets)). Positive EBI (i.e. latency to negative targets is larger than latency to positive targets) would suggest a positive bias and a negative EBI would suggest a negative bias. Three-group comparison revealed a significant effect of group on EBI (F(2,56) = 6.1, p = 0.004). Post-hoc pairwise comparisons showed that MDDa had a smaller EBI than both HC and MDDr, and there were no significant between -group differences on EBI for MDDr vs. HC (HC vs. MDDa: F(1,36) = 10.1, p = 0.003; HC vs. MDDr: F(1,36) = 0.89, p = 0.352 and MDDa vs. MDDr: F(1,39) = 6.64, p = 0.014) (Fig. 1).
Demographic and clinical data are reported in Table 1. All behavioral outcome measures are reported in Table 2.
3.2. Commission errors
We have applied 90% Winsorization (5% from each tail) on all outcome measures and calculated Winsorized means accordingly. Each dependent variable (omissions, commissions, latency) was then analyzed in a 2 × 2 × 3 repeated-measures analysis of covariance (ANCOVA) of ‘switch’ (switch, non-switch) by target valence (positive, negative), with group (MDDa, MDDr, HC) as a between-subjects variable and IQ as a covariate. To explore the triple interaction of valence × shift × group upon latency, we then conducted a repeated measure analysis under the shift and non-shift conditions separately and included IQ as a covariate followed by pairwise t-tests to compare latencies in each group to negative and positive stimuli. We performed exploratory analyses using Pearson's correlations between dependent variables that were significantly different among the three groups and relevant clinical data (CDRS, age of illness onset, total duration of depressive episodes, and total number of depressive episodes) in MDDa. Additional analyses were run to explore the potential confounding effects of medication, suicidality and comorbidities. T-tests were used to compare dependent variables that were abnormal in MDDa between those taking, vs. those not taking, selective serotonin reuptake inhibitors (SSRIs), those with vs. those without history of suicidality (separate comparisons for suicide attempts and ideation) and those with vs. those without comorbid anxiety disorders. For the three-group comparisons and exploratory correlation analyses in MDDa and MDDr, statistical significance was set at 0.05, and for post-hoc pair-wise comparisons, we have used a cut-off alpha of 0.016 to account for multiple comparisons.
3.1. Latency There was a statistically significant three-way interaction (shift × valence × group) upon latency F(2,53) = 3.31, p = 0.044. There was, however, no significant main effect of target valence (happy vs. sad) or shift condition (shift vs. non-shift) or group (HC vs. MDDa vs. MDDr). No significant interaction between valence and group or valence and shift condition upon latency (p > 0.05 for all) was found either. To further examine the above three-way interaction, we performed repeated measures analyses under shift and non-shift conditions separately with valence as a within-subjects variable and group as a between-subjects variable. There was a statistically significant valence × group interaction upon latency for the shift condition
There was a main effect of shift condition on commission errors (F(1,53) = 5.3, p = 0.025). Further analysis showed that all participants had more commission errors under the shift condition (mean: 7.58, standard deviation (S.D.): 5.17) as compared to the non-shift condition (mean: 7.33, S.D.: 6.20). This difference, however, was not statistically significant (p = 0.568). There was also a trend for an effect of group on commission errors (F(2,53) = 3.05, p = 0.056) indicating that MDDa had more commission errors overall than both HC and MDDr (mean: 11.5, S.D.:9.3 for HC; mean: 12.6, S.D. = 9.9 for MDDr; and mean: 201, S.D. = 11.6 for MDDa). There was, however, no significant effect of valence, or significant interaction between valence and shift condition, or valence and group, or valence and group and shift condition upon commission errors.
Table 2 Affective go/no go outcome measures by group.
Latency (shift to positive) mean (S.D.) Latency (non-shift positive) mean (S.D.) Latency (shift to negative) mean (S.D.) Latency (non-shift negative) mean (S.D.) Commission (shift to positive) mean (S.D.) Commission (non-shift positive) mean (S.D.) Commission (shift to negative) mean (S.D.) Commission (non-shift negative) mean (S.D.) Omission (shift to positive) mean (S.D.) Omission (non-shift positive) mean (S.D.) Omission (shift to negative) mean (S.D.) Omission (non-shift negative) mean (S.D.)
HC N = 15
MDDr N = 19
MDDa N = 19
462 (106)
478 (66)
488 (93)
470 (107)
485 (77)
473 (82)
486 (98)
486 (73)
451 (77)
488 (86)
481 (80)
483 (86)
2.8 (2.0)
3.2 (2.8)
4.8 (2.9)
2.4 (2.8)
3.4 (3.0)
4.8 (3.8)
3.1 (2.6)
2.9 (2.8)
5.7 (3.0)
3.2 (2.8)
3.0 (3.0)
4.8 (3.6)
4.2 (4.6)
2.6 (3.3)
3.2 (1.9)
4.5 (5.0)
2.9 (3.3)
2.7 (2.6.)
3.8 (3.7)
2.4 (2.7)
2.4 (2.4)
2.8 (3.1)
2.6 (3.2)
2.1 (2.5)
Fig. 1. Latency under the shift condition to negative and positive stimuli per group. Emotional bias index (EBI = latency (shift to negative targets) − latency (shift to positive targets)) across groups: HC (mean = 24.0, S.D.= 61.0), MDDr (mean = 8.2, S.D.= 40.0), MDDa (mean = −37.4, S.D.= 58.3). Three-group comparison: (F(2,56) = 6.1, p = 0.004). Post-hoc pairwise comparisons: MDDab HC (F(1,36)= 10.1, p = 0.003), MDDab MDDr (F(1,39) = 6.64, p = 0.014), and MDDr vs. HC (p > 0.05).
F.T. Maalouf et al. / Psychiatry Research 198 (2012) 28–33
3.3. Omission errors There was no significant effect of group, shift condition, valence or significant interaction between valence and shift condition, or valence and group, or valence and group and shift condition upon omissions. 3.4. Exploratory analysis There was no statistically significant correlation between EBI and depression severity as measured by CDRS scores, age of illness onset, total duration of depressive episodes, or number of depressive episodes in MDDa (p > 0.05 for all). EBI and total commission errors were not statistically different in MDDa between individuals with and without a history of suicide attempt, those with and without a history of suicidal ideation, those on and those off SSRIs, and those with and without a history of anxiety disorders (p > 0.05 for all). 4. Discussion The aim of this study was to examine the extent to which bias towards negative emotions was a state or trait marker in major depressive disorder (MDD) in adolescents. To do so, we assessed the performance of adolescent participants in a current MDD episode (MDDa), participants with MDD in remission (MDDr) and healthy controls (HC) on the affective go/no go task. We showed that bias towards emotionally salient stimuli with negative valence is a state marker of adolescent depression, present in MDDa and not in MDDr. Specifically, under the shift condition, three-group comparison showed that MDDa participants, when compared to the pattern observed in controls and MDDr, were faster on the blocks where target valence shifted to negative as compared to positive shifts. The fact that this difference was only observed under the shift condition may be due to the higher demand shift tasks place on dorsolateral prefrontal cortical regions to inhibit incongruent responses (Bermpohl et al., 2005). While non-shift blocks also demand that a previous incongruent response be inhibited, this process is more pronounced for shift blocks where the carryover effect of a previous congruent stimulus is not present. These results point to impaired inhibitory control processes in acutely depressed adolescents. Inhibitory control is a particularly effective regulatory process that helps depressed individuals to stop focusing on their mood-related thoughts, which maintain their negative mood (Joormann and Gotlib, 2007). Indeed, reduced inhibition when processing negative material has been proposed to underlie self-reported rumination in depression (Joormann and Gotlib, 2010), which has been associated with a poor course for major depressive episodes (Kuehner and Weber, 1999). Recent evidence also shows that impaired inhibitory control predicts the maintenance of depressive symptoms and rumination over a period of 6 months in healthy individuals (Zetsche and Joormann, 2011). It has also been linked to suicidal thinking in individuals in this population, since it is believed that individuals think of suicide as a solution to their problems, because they cannot think of any alternative course of action (Miranda and Nolen-Hoeksema, 2007). Other studies of MDD in adolescents have shown more errors on happy targets than sad targets in a depressed cohort with recent diagnosis of depression, and fewer errors on sad targets as compared to controls, but no group differences were detected on response latency (Kyte et al., 2005). The differences between the findings observed on the dependent measures of response latency in our study and response accuracy in Kyte et al.'s (2005) study may be due to a number of methodological factors. For example, in our study, depressed and remitted depressed adolescents were tested as separate groups and compared to controls. In Kyte et al.'s (2005) study, the case group included both depressed and remitted patients. Another difference was the inclusion of only first episode depressed adolescents by
31
Kyte et al. (2005), while our study did not select the sample based on number of depressive episodes. Finally, while response accuracy is a measure of automatic biases, response latency, especially at longer inter stimulus interval presentations, may be a better measure of voluntary processes of emotion regulation (Lawson and MacLeod, 1999; Sears et al., 2011), which was the main focus of the current study. Findings regarding commission rates indicated that, relative to the HC and MDDr groups, MDDa were more likely to respond to false targets regardless of the target's emotional valence, although this difference fell short of statistical significance. This suggests that the high commission rate in MDDa relative to HC and MDDr may be driven by impulsive responding rather than a bias towards certain emotions. This finding also replicates our previous findings of impulsivity emerging as a state-dependent marker for depression in adolescents (Maalouf et al., 2011), which is also consistent with other studies exploring impulsivity in children with MDD (Palladino et al., 1997; Kyte et al., 2005). These findings also suggest that when examining biases towards emotional stimuli using behavioral tasks in pediatric MDD, latency, rather than commissions, may be a more reliable outcome measure to differentiate between response styles to positively and negatively valenced stimuli as commission errors tap into more automatic processes of emotion regulation. Our data also provide some evidence for a lack of impairment in the processing of emotional stimuli in remitted adolescents with MDD. The lack of negative bias in our remitted group is in contrast to most adult studies that showed residual impairment in the period of remission in processing of emotional stimuli (Leppanen et al., 2004; Atchley et al., 2007; LeMoult et al., 2009) and bias towards sad stimuli (Joormann and Gotlib, 2007). This is consistent, however, with a few adult studies showing that remitted depressed participants do not differ from healthy controls in their inhibition of negative material (Joormann and Gotlib, 2010), and with other pediatric studies showing that on a recall task, remitted children showed no attentional bias to negative information (Timbremont and Braet, 2004). The discrepancy between our findings and the adult findings may be due to the fact that our paradigm did not include a priming strategy such as negative mood induction preceding the affective go/no go task. While some adult studies have been able to detect negative bias in remitted individuals without a priming manipulation (Joormann and Gotlib, 2007), most studies suggest that negative mood induction, simulating real life stressful events, is needed in order to detect the negative bias in remitted individuals (Scher et al., 2005; Ramel et al., 2007). Future studies that utilize mood priming paradigms are therefore necessary to either replicate or refute our findings under these settings. Alternatively, the trait-related negative bias in depression in adults may be due to the cumulative effect of recurrent depressive episodes, a potential “scar effect”, which our adolescent sample did not have, as the average number of episodes was 1.6 in MDDr (Robinson and Sahakian, 2008). This also highlights the importance of early intervention in order to prevent recurrence of episodes and potential development of a persistent negative bias. One of the strengths of our study is that our sample was well categorized into acute and remitted MDD subgroups, and we excluded adolescents with history of ADHD which, if present, could have impacted response style on our chosen behavioral task. This could, however, limit the generalizability of our findings, since ADHD is a prevalent comorbid disorder in adolescents with MDD in clinical practice (Biederman et al., 1995). A limitation is the cross-sectional design of the present study, which precludes us from drawing strong conclusions regarding the subsidence lack of negative bias with remission of depressive symptoms. Future studies should therefore aim at examining emotional bias during acute episode and in remission in the same sample of adolescents with MDD. It is also to be noted that the effect sizes that we observed for the differences between MDDa and HC and MDDr and MDDa on EBI were large (Cohen's d = 1) and small (Cohen d = 0.3) respectively. When
32
F.T. Maalouf et al. / Psychiatry Research 198 (2012) 28–33
considering our null findings for MDDr vs. HC on the EBI, we did also have adequate statistical power to detect any large effect size between MDDr vs. HC on EBI given our sample size. The statistical power for this comparison given a threshold of p b 0.05 and an effect size comparable to the effect size observed in our main finding – MDDa vs. HC – (Cohen's d = 1) would be 80%. A larger sample size would, however, be needed to detect smaller effect sizes between MDDr and HC with adequate statistical power. In summary, this is the first study to examine the extent to which bias towards negative emotional stimuli in an inhibitory control paradigm represents a state vs. a trait marker of depression in adolescents. We showed that bias towards negative emotions, as measured by the affective go/no go task, appeared to be present in the acute stage of MDD and absent in remission, suggesting that it is a state-specific marker in adolescent depression. Studies have shown that antidepressants modulate emotion processing much earlier than the effect on mood appears (Harmer, 2008). Given that the behavioral task that we used in this study can be easily employed in different clinical settings, it shows promise as a task to identify state-specific markers in MDD that may serve as early predictors of treatment response in MDD. Disclosure of interest Dr. Maalouf is on the speaker bureau of Eli Lilly. Dr. Brent receives research support from the National Institute of Mental Health, receives royalties from Guilford Press and is UpToDate Psychiatry Editor. Dr. Luke Clark consults for Cambridge Cognition plc. Dr. Sahakian consults for Cambridge Cognition. She has consulted for Novartis, Shire, GlaxoSmithKline, Lilly Boehringer-Ingelheim and F Hoffman-La Roche Ltd. She has also received honoraria for Grand Rounds in Psychiatry at Massachusetts General Hospital (CME credits) (Boston, 27 April 2007) and for speaking at the International Conference on Cognitive Dysfunction in Schizophrenia and Mood Disorders: Clinical Aspects, Mechanisms and Therapy (Brescia, 17–19 January 2007). She was on the Medical Research Council Neurosciences and Mental Health Board (2010) and on the Science Coordination Team for the Foresight Project on Mental Capital and Wellbeing, 2008 (Office of Science, The Department of Innovation, Universities and Skills,). She is currently on Panel LS5 for the European Research Council. As an Associate Editor, she also receives an honorarium from Psychological Medicine. Dr. Phillips and Ms. Tavitian report no biomedical financial interests or potential conflicts of interest. Acknowledgment This study was funded by the American Academy of Child and Adolescent Psychiatry — Eli Lilly Pilot Research Award and the Junior Faculty Scholar Program at the University of Pittsburgh/Western Psychiatric Institute and Clinic (R25 MH060473-10 Pilkonis (PI)). References Atchley, R.A., Stringer, R., Mathias, E., Ilardi, S.S., Diane Minatrea, A., 2007. The right hemisphere's contribution to emotional word processing in currently depressed, remitted depressed, and never-depressed individuals. Journal of Neurolinguistics 20 (2), 145–160. Bermpohl, F., Fregni, F., Boggio, P.S., Thut, G., Northoff, G., Otachi, P.T.M., Rigonatti, S.P., Marcolin, M.A., Pascual-Leone, A., 2005. Left prefrontal repetitive transcranial magnetic stimulation impairs performance in affective go/no-go task. NeuroReport: For Rapid Communication of Neuroscience Research 16 (6), 615–619. Biederman, J., Faraone, S., Mick, E., Lelon, E., 1995. Psychiatric comorbidity among referred juveniles with major depression: fact or artifact? Journal of the American Academy of Child and Adolescent Psychiatry 34 (5), 579–590. Emslie, G.J., Heiligenstein, J.H., Wagner, K.D., Hoog, S.L., Ernest, D.E., Brown, E., Nilsson, M., Jacobsen, J.G., 2002. Fluoxetine for acute treatment of depression in children and adolescents: a placebo-controlled randomized clinical trial. Journal of the American Academy of Child and Adolescent Psychiatry 41 (10), 1205–1214. Erickson, K., Drevets, W.C., Clark, L., Cannon, D.M., Bain, E.E., Zarate Jr., C.A., Charney, D.S., Sahakian, B.J., 2005. Mood-congruent bias in affective go/no-go performance
of unmedicated patients with major depressive disorder. The American Journal of Psychiatry 162 (11), 2171–2173. Fleck, D.E., Kotwal, R., Eliassen, J.C., Lamy, M., Delbello, M.P., Adler, C.M., Durling, M., Cerullo, M.A., Strakowski, S.M., 2011. Preliminary evidence for increased frontosubcortical activation on a motor impulsivity task in mixed episode bipolar disorder. Journal of Affective Disorders 133 (1–2), 333–339. Gorlyn, M., Keilp, J., Grunebaum, M., Taylor, B., Oquendo, M., Bruder, G., Stewart, J., Zalsman, G., Mann, J., 2008. Neuropsychological characteristics as predictors of SSRI treatment response in depressed subjects. Journal of Neural Transmission 115 (8), 1213–1219. Harmer, C.J., 2008. Serotonin and emotional processing: does it help explain antidepressant drug action? Neuropharmacology 55 (6), 1023–1028. Joormann, J., Gotlib, I.H., 2007. Selective attention to emotional faces following recovery from depression. Journal of Abnormal Psychology 116 (1), 80–85. Joormann, J., Gotlib, I.H., 2010. Emotion regulation in depression: relation to cognitive inhibition. Cognitive Emotion 24 (2), 281–298. Kaplan, J.S., Erickson, K., Luckenbaugh, D.A., Weiland-Fiedler, P., Geraci, M., Sahakian, B.J., Charney, D., Drevets, W.C., Neumeister, A., 2006. Differential performance on tasks of affective processing and decision-making in patients with panic disorder and panic disorder with comorbid major depressive disorder. Journal of Affective Disorders 95 (1–3), 165–171. Kaufman, J., Birmaher, B., Brent, D., Rao, U., Flynn, C., Moreci, P., Williamson, D., Ryan, N., 1997. Schedule for Affective Disorders and Schizophrenia for School-Age Children—Present and Lifetime Version (K-SADS-PL): initial reliability and validity data. Journal of the American Academy of Child and Adolescent Psychiatry 36 (7), 980–988. Kuehner, C., Weber, I., 1999. Responses to depression in unipolar depressed patients: an investigation of Nolen–Hoeksema's response styles theory. Psychological Medicine: A Journal of Research in Psychiatry and the Allied Sciences 29 (6), 1323–1333. Kyte, Z.A., Goodyer, I.M., Sahakian, B.J., 2005. Selected executive skills in adolescents with recent first episode major depression. Journal of Child Psychology and Psychiatry 46 (9), 995–1005. Ladouceur, C.D., Dahl, R.E., Williamson, D.E., Birmaher, B., Ryan, N.D., Casey, B.J., 2005. Altered emotional processing in pediatric anxiety, depression, and comorbid anxiety–depression. Journal of Abnormal Child Psychology 33 (2), 165–177. Lawson, C., MacLeod, C., 1999. Depression and the interpretation of ambiguity. Behaviour Research and Therapy 37 (5), 463–474. LeMoult, J., Joormann, J., Sherdell, L., Wright, Y., Gotlib, I.H., 2009. Identification of emotional facial expressions following recovery from depression. Journal of Abnormal Psychology 118 (4), 828–833. Leppanen, J.M., Milders, M., Bell, J.S., Terriere, E., Hietanen, J.K., 2004. Depression biases the recognition of emotionally neutral faces. Psychiatry Research 128 (2), 123–133. Lewinsohn, P.M., Rohde, P., Seeley, J.R., 1998. Major depressive disorder in older adolescents: prevalence, risk factors, and clinical implications. Clinical Psychology Review 18 (7), 765–794. Maalouf, F.T., Brent, D., Clark, L., Tavitian, L., McHugh, R.M., Sahakian, B.J., Phillips, M.L., 2011. Neurocognitive impairment in adolescent major depressive disorder: state vs. trait illness markers. Journal of Affective Disorders 133 (3), 625–632. Mayberg, H.S., 2007. Defining the neural circuitry of depression: toward a new nosology with therapeutic implications. Biological Psychiatry 61 (6), 729–730. Mayberg, H.S., Brannan, S.K., Mahurin, R.K., Jerabek, P.A., Brickman, J.S., Tekell, J.L., Silva, J.A., McGinnis, S., Glass, T.G., Martin, C.C., Fox, P.T., 1997. Cingulate function in depression: a potential predictor of treatment response. Neuroreport 8 (4), 1057–1061. Miranda, R., Nolen-Hoeksema, S., 2007. Brooding and reflection: rumination predicts suicidal ideation at 1-year follow-up in a community sample. Behaviour Research and Therapy 45 (12), 3088–3095. Murphy, F.C., Sahakian, B.J., Rubinsztein, J.S., Michael, A., Rogers, R.D., Robbins, T.W., Paykel, E.S., 1999. Emotional bias and inhibitory control processes in mania and depression. Psychological Medicine: A Journal of Research in Psychiatry and the Allied Sciences 29 (6), 1307–1321. Palladino, P., Poli, P., Masi, G., Marcheschi, M., 1997. Impulsive–reflective cognitive style, metacognition, and emotion in adolescence. Perceptual and Motor Skills 84 (1), 47–57. Petersen, A.C., Crockett, L., Richards, M., Boxer, A., 1988. A self-report measure of pubertal status: reliability, validity, and initial norms. Journal of Youth and Adolescence 17 (2), 117–133. Phillips, M.L., Ladouceur, C.D., Drevets, W.C., 2008. A neural model of voluntary and automatic emotion regulation: implications for understanding the pathophysiology and neurodevelopment of bipolar disorder. Molecular Psychiatry 13 (9), 833–857 (829). Poznanski, E.O., Grossman, J.A., Buchsbaum, Y., Banegas, M., Freeman, L., Gibbons, R., 1984. Preliminary studies of the reliability and validity of the children's depression rating scale. Journal of the American Academy of Child Psychiatry 23 (2), 191–197. Ramel, W., Goldin, P.R., Eyler, L.T., Brown, G.G., Gotlib, I.H., McQuaid, J.R., 2007. Amygdala reactivity and mood-congruent memory in individuals at risk for depressive relapse. Biological Psychiatry 61 (2), 231–239. Robinson, O.J., Sahakian, B.J., 2008. Recurrence in major depressive disorder: a neurocognrtive perspective. Psychological Medicine: A Journal of Research in Psychiatry and the Allied Sciences 38 (3), 315–318. Rubinsztein, J.S., Michael, A., Underwood, B.R., Tempest, M., Sahakian, B.J., 2006. Impaired cognition and decision-making in bipolar depression but no ‘affective bias’ evident. Psychological Medicine: A Journal of Research in Psychiatry and the Allied Sciences 36 (5), 629–639.
F.T. Maalouf et al. / Psychiatry Research 198 (2012) 28–33 Scher, C.D., Ingram, R.E., Segal, Z.V., 2005. Cognitive reactivity and vulnerability: empirical evaluation of construct activation and cognitive diatheses in unipolar depression. Clinical Psychology Review 25 (4), 487–510. Sears, C.R., Suzie Bisson, M.A., Nielsen, K.E., 2011. Dysphoria and the immediate interpretation of ambiguity: evidence for a negative interpretive bias in error rates but not response latencies. Cognitive Therapy and Research 35 (5), 469–476.
33
Timbremont, B., Braet, C., 2004. Cognitive vulnerability in remitted depressed children and adolescents. Behaviour Research and Therapy 42 (4), 423–437. Zetsche, U., Joormann, J., 2011. Components of interference control predict depressive symptoms and rumination cross-sectionally and at six months follow-up. Journal of Behavior Therapy and Experimental Psychiatry 42 (1), 65–73.