Emotionally Valenced Stimuli Impact Response Inhibition in Those with Substance Use Disorder and Co-Occurring Anxiety and Depression Symptoms

Abstract

Background:

Substance use disorder (SUD) is associated with impaired response inhibition. Given the deficits in emotion regulation associated with SUD, it is unclear if this impairment is exacerbated by emotionally valenced stimuli. Co-occurring conditions may further exacerbate these impairments as many co-occurring conditions further impact emotion regulation. It was hypothesized that negative stimuli may further impact response inhibition for this population.

Methods:

The current study used the stop-signal task to examine response inhibition to negative, neutral and positive stimuli in a sample of those with a history of SUD and co-occurring depression and anxiety symptoms.

Results:

Response inhibition was poorer for negative stimuli relative to neutral stimuli. There was no difference between negative and positive stimuli. Depression severity moderated the difference between response inhibition for negative and neutral stimuli. At elevated depression, response inhibition was worse and there was no difference across emotional stimuli. At low depression, there was a significant difference between negative and neutral stimuli. This effect was not found for anxiety symptoms.

Limitations:

Study participants presented with polysubstance use of varying duration and amount. It is unclear whether findings are attributed to specific substances, or substance use broadly. Additionally, happy, angry, and calm facial emotions were used to represent positive, negative, and neutral valences respectively. It is unclear whether these findings are generalizable to other emotional expressions.

Conclusion:

Results suggested that emotionally valenced stimuli affected response inhibition among those with low symptom severity. At elevated symptom severity, response inhibition to all stimuli were impaired.

Substance use disorder (SUD) is associated with significant impairment across the lifespan and increased mortality and is thus a major public health concern (Onyeka, 2015; Borges, Walters, & Kessler, 2000; Whiteford et al., 2013). Co-occurring conditions are common among those with SUD with depression (35%) and anxiety disorders (45%) being the most common (Merikangas et al., 1998; Grant et al., 2004). Those with co-occurring disorders experience more severe mood and anxiety symptoms as well as substance-related impairments (Merikangas et al., 1998; Regier et al., 1990; Swendsen & Merikangas, 2000). Understanding the processes that underlie this impairment is necessary to improve treatments for these conditions. One such factor that may be impacted is inhibitory functioning, defined as the ability to suppress a behavioral response (Verbruggen & Logan, 2009).

Poor inhibitory functioning in those with SUD is thought to map onto altered responding to salient drug cues (Nigg, 2000; Jentsch & Taylor, 1999; de Wit, 2009). Several studies examining participants who reported substance use found that these individuals had poorer response inhibition relative to individuals who did not use substances (Fillmore & Rush, 2002; Monterosso, Aron, Cordova, Xu, & London, 2005). A study examining response inhibition in adolescents with substance use using fMRI and the Go-No-Go Task (GNG) discovered atypical activation in areas outside of the ventromedial prefrontal (vmPFC) and anterior cingulate cortices (ACC) (Mahmood, et al., 2013). Activation of the vmPFC and ACC during the GNG task was shown to be predictive of increased substance use at an 18-month follow up. Taken together, these results suggest that poorer response inhibition is associated with sustained SUD.

Much of the work examining response inhibition in those with a history of SUD have used neutral stimuli, such as arrows, shapes, or letters, to control for extraneous effects, such as emotional reactivity (Lawrence, Luty, Bogdan, Sahakian, & Clark, 2009; Mahmood et al., 2013; Heitzeg et al., 2014). There is considerable evidence to suggest that those with SUD often have a strong reaction to emotionally relevant stimuli, however. Several studies examining those who reported substance use found that pertinent trauma-related stimuli increased drug cravings, arousal, and increased negative affect (Coffey et al., 2002; McHugh, Gratz, & Tull, 2017). In a recent study of those with current opioid dependence, exposure to a stressful stimulus was associated with greater cue reactivity as defined by elevated stress, anger, and craving (Back et al., 2015). This elevated reactivity to emotional stimuli among those with a history of SUD may impact response inhibition toward such stimuli. Specifically, stronger reactions to negative stimuli may further impair response inhibition relative to neutral stimuli.

Prior work with healthy adults has suggested that the emotional valence of a stimulus affects response inhibition. Kalanthroff, Cohen, and Henik (2013) used peripheral negative images during the stop signal task (SST) and found that reaction times on stop trials were worse if preceded by a negative image as opposed to a neutral image. A study completed by Pessoa, Padmala, Kenzer, and Bauer (2012) found that when healthy participants were exposed to low-threat faces, their SST worsened relative to neutral faces. Rebetez, Rochat, Billieux, Gay, & Van der Linden (2015) used emotional faces as stimuli in the SST with healthy adults and found that SSRT for angry faces was longer than that for neutral faces. In an fMRI study with healthy adults, participants were asked to complete the SST using neutral or fearful faces (Sagaspe, Schwartz, & Vuilleumier, 2011). GO_RT for fearful faces was slower than for neutral faces, however there were no differences in response inhibition. Despite the lack of a behavioral difference in response inhibition, there were differences in brain activity such that trials involving negative stimuli recruited the orbitofrontal cortex (OFC), which is involved in outcome expectations for emotional stimuli (Davidson, Putnam, & Larson, 2000). This difference in brain activation suggests that the lack of a difference in response inhibition may be attributed to the recruitment of emotion regulation areas of the brain. Persistent use of substances has been shown to affect these regions, especially the orbitofrontal cortex (OFC) (Schoenbaum, Roesch, & Stalnaker, 2006). As such, it is hypothesized that response inhibition to emotionally valenced stimuli may be further impacted among those with a history of SUD.

In understanding how emotionally valenced stimuli affects response inhibition, it is necessary to consider how response inhibition is impacted by the presence of co-occurring psychopathology. There has been relatively little work examining the effect of mood and anxiety disorders on response inhibition. Several studies have suggested that those with depression demonstrate slowed SSRT to neutral stimuli when compared to non-depressed controls (Aker, Bø, Harmer, Stiles, & Landrø, 2016; Bredemeier, Warren, Berenbaum, Miller, & Heller, 2016). Taken together, these findings suggest that depression symptoms may further impair response inhibition in those with SUD, with the worst response inhibition observed for negative stimuli.

The effect of anxiety pathology on response inhibition is unclear. Initial pre-clinical work suggested that elevated anxiety may facilitate response inhibition, presumably because the individual is better primed to withhold a response to a possible threat (Grey & McNaughton, 2000). Consistent with this hypothesis, two studies found that elevated state anxiety during a GNG task improved response inhibition (Robinson, Krimsky, & Grillon, 2013; Grillon, Robinson, Mathur, & Ernst, 2016). State anxiety in these studies was induced by threat of electrical shock for incorrect responses. In contrast, several studies demonstrated that elevated trait anxiety was unrelated to response inhibition (Karch et al., 2008; Righi et al., 2009; Li, Zinbarg, Boehm, & Paller, 2008). Recently, it was suggested that elevated state and trait anxiety may have an interactive effect on response inhibition such that when both are elevated, response inhibition is further impaired (Grillon et al., 2017). Preliminary data suggest that this hypothesis is supported in that those with anxiety disorders presented with excessive response inhibition (Grillon et al., 2017). Excessive response inhibition is defined as improved response inhibition at the expense of overall reaction time when perceiving threat. In contrast, when healthy individuals were under threat, response inhibition improved but overall reaction time was unaffected. These results suggest that the type of stimulus (threatening vs non-threatening) may interact with anxiety to affect response inhibition. It is unclear how the presence of comorbid SUD may further affect these factors. Given that those with SUD are prone to having elevated anxiety symptoms, it is hypothesized that elevated trait anxiety would further impair response inhibition for negative stimuli.

The present study had two primary aims. The first was to examine how response inhibition was affected by stimuli of a specific emotional valence in those with a history of SUD. It was hypothesized that response inhibition would be worse for negative stimuli relative to neutral stimuli. To determine if this difference is attributed to negative emotional stimuli as opposed to emotional stimuli, positive emotional stimuli were also included. The second aim was to determine if the presence of co-occurring psychopathology further impacted response inhibition to emotional stimuli. It was hypothesized that participants with higher levels of self-reported anxiety and depression would demonstrate poorer response inhibition towards negative stimuli.

Method

Data for the current study came from a larger study conducted from the Northeastern United States designed to examine the relation between substance use, mental health disorders, and response inhibition (Price et al., 2018). For the study visit, participants completed diagnostic interviews, self-report measures, and the SST.

Participants

Participants were 72 individuals with a history of SUD. All participants recruited because they had a prior history of opioid use, defined as misusing prescription opioids or heroin for at least 1 year (Heroin M = 4.12 years, SD = 4.52; Prescription opioids: M = 6.95 years, SD = 6.21), and opioids being the drug of choice. A majority of participants (n = 46, 69%) reported misusing opioids in the last month. Predominantly, participants were also involved in methadone maintenance treatment (n = 52, 72%). Furthermore, history in years of lifetime use of other substances were as follows: Alcohol: M = 10.56, SD = 8.81; Cocaine: M = 6.12, SD = 6.59; Cannabis: M = 13.34, SD = 9.45. Taken together, although the sample all had a history of opioid misuse, there was evidence of consistent chronic substance use. No participants were under the influence of substances at the time of the clinical interview.

Clinical interviews were used to corroborate participants” reports of depression and anxiety. These ratings mirrored rates of self-reported depression and anxiety used in the study. The sample mean for the BAI was 27.79 (SD = 15.51) falling within the “Severe” range of anxiety (Beck & Steer, 1990). The sample mean for the PHQ was 13.31 (SD = 6.73) falling within the “Moderate” range for depression (Kroenke, Spitzer, & Williams, 2001). This data suggests that self-reported rates of depression and anxiety were similar to clinician-rated diagnostic information for depression and anxiety. Of the N = 72 participants, 84.7% (n = 61) met diagnostic criteria for at least one anxiety disorder or PTSD as determined by clinical interviews. Prevalence of these diagnoses were as follows: generalized anxiety disorder (n = 29), social anxiety disorder (n = 32), panic disorder (n = 24), agoraphobia (n = 10), obsessive compulsive disorder (n = 15), posttraumatic stress disorder (n = 50), and specific phobias (n = 13). Approximately one third met criteria for major depression (n=28). The majority of the sample met diagnostic criteria for at least 1 or more diagnosis: 1 (n = 12, 16%), 2 (n = 15, 21%), 3 (n = 12, 16%), 4 (n = 13, 18%), 5 (n = 4, 7%), 6 (n = 6, 8%). A small portion of the sample did not meet criteria for any diagnosis (n = 10, 14%).

The average number of years of education was 12.67 (SD = 2.14) years and the majority of participants (n = 41, 56.2%) earned less than $10,000 per year, which is below the federal poverty level. The sample was predominantly male (n = 41), 56.2% with an average age of 34 (SD = 7.25). Participants self-identified as White 86% (n = 61), African American 1.4% (n = 1), Pacific Islander/Native Alaskan 1.4% (n = 1), Asian American 2.8% (n = 2), Native American 5.6% (n = 4), and Bi-Racial 2.8% (n = 2). Exclusion criteria included a lack of English fluency and active psychosis.

Procedure

Participants attended an in-person study visit during which they completed a battery of questionnaires as part of a larger study examining how PTSD effects behavior inhibition in those who habitually use opioids. Trained research assistants administered the Structured Clinical Interview for DSM-5 (SCID-5) and Addiction Severity Index Lite (ASI-Lite) and reviewed for consistency by a licensed clinical psychologist (First, Williams, Karg, & Spitzer, 2015; McLellan et al., 1992). Later, participants were asked to complete the SST. Before beginning the experimental trials, participants had to achieve at least 80% accuracy on a practice trial. The experimental trial SST paired the trained visual stimuli with images of faces with one of 3 emotional valences; angry, happy, or calm.

Measures

Stop-Signal Task (SST; Logan, 1985).

The SST is a cognitive task designed to measure a participant’s ability to inhibit behaviors in response to an auditory signal in STOP trials (87) and their reaction time to visual stimuli in GO trials (177). A variant of the SST developed by Logan (1994) was used. Each task trial began with a fixation cross presented for 250 milliseconds (ms). A target stimulus was then presented that consisted of a face from the NimStim stimulus set for 1400 ms (Tottenham et al., 2009). Color images were of men and women who were making neutral, happy, or angry facial expressions with open and closed mouths and were equally divided and randomized across trial types. Angry faces were chosen to represent negative stimuli in this study. Although some previous research on depression and negative stimuli have utilized sad faces, multiple studies have used angry faces (Goldstein et al., 2007; Hare et al., 2005; Pawliczek et al., 2013; Shafritz et al., 2006) or have found no difference between anger and other negative facial emotions on reaction time tasks (Leyman, De Raedt, Schacht, & Koster, 2006; Domes, Normann, Heinrichs, 2016; Mogg, Millar, Bradley, 2000). Participants were asked to identify the gender of the individual as quickly as possible. For STOP trials, the onset of the stop signal was dynamic in that successful STOP trials resulted in an increase in the onset of the stop signal by 50ms and failed STOP trials reduced the onset of the stop signal by 50ms. This adaptive timing was used to obtain a successful stop rate of approximately 50%. The three outcomes of interest were: (1) GO_RT on correct GO trials, (2) GO_SD variability in the reaction time for GO trials, and (3) SSRT, which is the difference between the mean GO_RT and the mean STOP signal presentation for trials in which the participant successfully stopped. Longer SSRTs indicate poorer response inhibition.

Addiction Severity Index Lite (ASI-Lite; McLellan et al., 1992).

The ASI-Lite is an abbreviated adaptation of the ASI-5. It is a 169-item structured interview designed to assess rates of drug or alcohol addiction as well as levels of associated impairment in life domains such as alcohol use, drug use, legal history, family/social history, psychological history, medical history, and employment status. The ASI-Lite was used in this study to determine substance use frequency.

Patient Health Questionnaire (PHQ-9; Kroenke, Spitzer, & Williams, 2001).

The PHQ-9 is a widely used 9-item self-report tool used to assess depression as defined by DSM-IV. It is used to determine severity and frequency of depressive symptoms as well as the patient’s level of function. The PHQ-9 uses a Likert scale (0–3) to indicate how often patients are experiencing symptoms. This measure demonstrated excellent internal reliability (a = .93).

Beck Anxiety Inventory (BAI; Beck & Steer, 1990).

The BAI is comprised of 21 self-report items used to assess common symptoms of anxiety, including physical symptoms. Answers are based on a Likert scale (0–3), with higher scores indicating higher levels of symptom severity. The BAI demonstrated excellent internal consistency (a = .96).

Data Analysis Plan

Repeated-measures analyses of variance (RM-ANOVAs) were used to determine if there was a difference between SSRT, go reaction times (GO_RT), and go standard deviation (GO_SD), for each emotional face (angry, happy, and calm). To determine the moderating effect of depression and anxiety, mixed models were used to examine the interaction between anxiety, depression, and SSRT to emotional faces. Models contained main effects for valence in which emotional stimuli were compared to calm stimuli and for psychopathology. Interaction terms were included between emotional valence (neutral/emotional) and the psychopathology variable (depression and anxiety). Separate models were used for anxiety and depression. For significant interactions the models were probed at high (+1 SD) and low (−1SD) levels of psychopathology.

Results

Mean scores and standard deviations of SST performance are reported in Table 1. These data indicated that the SST was administered correctly. The proportion of correct stop trials was M = 44%, SD = 9%, which suggested that the task maintained a consistent stop rate. There were relatively few omissions (3%), which suggested that participants consistently responded to GO trials. These data suggest that, in aggregate, this modified SST performed correctly.

Table 1.

Descriptive statistics for stop signal tasks

Task Variable	Mean (SD)
GORT	758.49 (214.77)
Angry	762.59 (215.12)
Happy	752.17 (208.59)
Calm	748.84 (211.93)
GOcRT	738.66 (198.93)
GOSD	197.50 (62.30)
SSRT	253.83 (68.22)
Angry	262.10 (71.28)
Happy	255.32 (68.95)
Calm	244.37 (76.12)
SSRT PRS	0.44 (0.09)
SSRT Delay	418.88 (176.08)
Omission Error Rate	0.03 (0.04)

Note. GO_RT = Go Reaction Time. GOc_RT = Go Reaction Time on correct trials. GO_SD = Go Reaction Time Standard Deviation. SSRT PRS= Probability of responding on a stop trial. SSRT Delay = Delay between presentation of the go stimulus and the stop-signal.

There was a significant difference in SSRT across emotional faces, F (2, 71) = 6.17, p < .01. Post-hoc pairwise comparisons indicated that the difference between SSRT to negative faces (M = 262.00, SD = 71.28) was significantly longer than that to neutral faces (M = 244.37, SD = 76.12, p < .01, d = 0.41). The difference between SSRT to positive faces (M = 255.32, SD = 68.95) and neutral faces (p = .06, d = 0.26) was not significant. There were also no differences in SSRT between negative and positive faces (p = .34, d = 0.13). For GO_RT, there was a significant difference across emotional stimuli, F (2, 71) = 5.67, p < .01. There was a significant difference between GO_RT for negative faces (M = 762.59, SD = 215.12) and neutral faces (M = 748.84, SD = 211.93, p < .01, d = 0.38) such that negative GO_RT was slower than GO_RT for neutral faces. GO_RT for negative faces was slower than for positive faces (p = .04, d = 0.25). There were no differences in GO_RT between positive and neutral faces (p < .72, d = 0.08). Finally, there were significant differences in GO_SD across emotional faces, F (2, 71) = 3.96, p = .02. There was a significant difference between GO_SD for neutral (M = 189.22, SD = 66.93) and positive faces (M = 201.95, SD = 71.8), such that the GO_SD for positive faces was larger than for neutral faces (p = .01, d = 0.30). There were other no significant differences for GO_SD.

Mean scores, standard deviations, and correlations for psychopathology variables can be found in Table 2. There was a significant depression by negative face interaction (B = −1.50, SE = 0.69, p = .03) as well as a significant depression by positive face interaction (B = −1.51, SE = 0.69, p = .03). The moderating effect of depression scores on the relation between faces and SSRT was then examined. The significant interactions were probed at high and low (± 1 SD) levels for psychopathology (Table 3). These methods were chosen in accordance with guidelines suggested by Aiken, West, and Reno in 1991 and have been found acceptable to be used to probe interactions. Graphs in Figure 1 are in reference to the categorical variables created by the standard deviation split and are meant to aid in interpretation of reported continuous interactions. In the low depression group, there was a significant difference between both negative (B = 30.78, SE = 5.64, p < 0.01) and positive faces (B = 19.48, SE = 6.13, p < .01) from neutral faces. These findings suggest that at low depression, SSRT for negative and positive faces were significantly longer than for neutral faces. For the high depression group, there were no differences between negative (B = 1.96, SE = 6.41, p = 0.76) or positive (B = 2.07, SE = 7.06, p = 0.77) faces relative to neutral faces. This finding suggests there were no difference between emotional faces at high levels of depression (Figure 1). Also, SSRTs increased for all emotional faces at high depression.

Table 2.

Correlation Matrix and Descriptive Statistics

Measure	1.	2.	3.	4.	5.	6.	.7
1. PHQ	-
2. BAI	0.76	-
3. SSRT Angry	0.02	0.06	-
4. SSRT Calm	0.23	0.15	0.85	-
5. SSRT Happy	0.11	−0.01	0.83	0.85	-
6. GORT	0.02	0.06	0.13	0.06	0.08	-
7. GOSD	0.28	0.30	0.26	0.22	0.19	0.73	-
M (SD)	1.	2.	3.	4.	.5	.6	.7
	13.31 (6.73)	27.79 (15.51)	262 (71.28)	244.37 (76.12)	255.32 (68.95)	754.65 (210.87)	197.5 (62.30)

Note. PHQ = Patient Health Questionnaire. BAI = Beck Anxiety Inventory. SSRT = Stop Signal Reaction Time. GO_RT = Go Reaction Time. GO_SD = Go Reaction Time Standard Deviation.

Table 3.

Interactions and Simple Effects

Emotion BAI interaction	B	SE	P
BAI	0.75	0.55	.18
Angry	29.11	9.48	<.01
Happy	32.50	9.48	<.01
Angry*BAI	−0.47	2.98	.12
Happy*BAI	−.079	2.98	<.01
Emotion PHQ interaction
PHQ	2.55	1.27	.05
Angry	36.03	10.21	<.01
Happy	30.54	10.21	<.01
Angry*PHQ	−1.50	0.69	.03
Happy*PHQ	−1.51	0.69	.03
Emotion BAI simple effects
High BAI*Angry	11.62	7.23	.12
Low BAI*Angry	21.11	5.43	<.01
High BAI*Happy	−1.38	6.99	.85
Low BAI*Happy	23.74	5.70	<.01
Emotion PHQ simple effects
High PHQ*Angry	1.96	6.41	.76
Low PHQ*Angry	30.78	5.64	<.01
High PHQ*Happy	2.07	7.06	.77
Low PHQ*Happy	19.48	6.13	<.01

Note. PHQ = Patient Health Questionnaire. BAI = Beck Anxiety Inventory.

Figure 1.

Note. Relation between anxiety and depression symptom classification and SSRT across emotional valence. SSRT = Stop Signal Reaction Time. PHQ = Patient Health Questionnaire. BAI = Beck Anxiety Inventory.

A different set of findings was obtained for anxiety. There was a significant anxiety by positive face interaction for SSRT (B = −0.79, SE = 2.98, p = .009). The negative face by anxiety score interaction for SSRT was not significant (B = −0.47, SE = 2.98, p = .12). At low levels of BAI, there was a significant difference between negative faces and neutral faces (B = 21.11, SE = 5.43, p < 0.01) and between positive faces and neutral faces (B = 23.74, SE = 5.70, p < 0.01). This suggested that at low anxiety, SSRT for both negative and positive faces were significantly longer than for neutral faces. However, at high levels of anxiety there was no effect of emotional faces for SSRT in the high anxiety group for either negative (B = 11.62, SE = 7.23, p = 0.12) or positive faces (B = −1.38, SE = 6.99, p = 0.85).

Discussion

The current study contributes to a small body of literature on the impact of emotionally valenced stimuli on response inhibition. It is among the first to demonstrate that, among those with a history of SUD, the presence of a negative stimulus (angry face) slows overall reaction times and reduces response inhibition. Others (Rebetez et al., 2015) have described this difference as interference such that the processing of the emotional component of the stimulus supersedes the inhibition component of the task. The effect size of the difference in emotional responding for negative stimuli was comparable to that obtained in a sample with healthy controls (d = 0.51, Rebetez, et al., 2015), suggesting that this interference effect may be comparable across populations. These findings are relevant to those with SUD as it suggests that the ability to appropriately withhold a response may be further compromised to a negative stimulus, which is a possible mechanism by which the disorder is maintained.

The present study did not find a significant difference between response inhibition for positive stimuli relative to neutral stimuli. However, the effect size of the difference between positive and neutral stimuli was in the small-medium range (d = 0.26), suggesting this difference is potentially meaningful. Although less examined than negative stimuli, prior work on positive stimuli and response inhibition has suggested these stimuli may also have an interference effect, albeit smaller (Rebetez et al., 2015). Continued exploration of the role of positive stimuli on response inhibition is warranted, especially in the SUD population. Recent reviews have highlighted the importance of understanding how reactions to positive stimuli are altered in those with SUD (Weiss, Tull, Sullivan, Dixon-Gordon, & Gratz, 2015; Weiss, Forkus, Contractor, & Schick, 2018).

Depression moderated the difference in response inhibition in a manner that was different than hypothesized. At low levels of depression, there was evidence of an interference effect for emotional stimuli relative to neutral stimuli. Specifically, response inhibition to emotional stimuli was slower than that to neutral stimuli. However, at elevated levels of depression, two differences were observed. First, there was no significant difference in response inhibition across stimuli. Second, response inhibition worsened, suggesting that elevated depression further impaired response inhibition. Related work has suggested that those with depression are less impacted by the emotional valence of stimuli, particularly faces (Stuhrmann, Suslow, & Dannlowski, 2011). Specifically, neutral faces are interpreted comparably to emotional faces (for review see Bourke, Douglas, & Porter, 2010). This phenomenon may explain the reduction in the difference between SSRTs across emotional stimuli associated with increased depression.

Brain-based studies provide additional information on how emotional stimuli may impacts response inhibition. As discussed previously, work with healthy controls on response inhibition has shown that when the emotional valence of the stimuli varies, the orbitofrontal cortex (OFC) is recruited for inhibition (Sagaspe et al., 2011). The OFC is involved in numerous executive and emotional functions with its primary role being learning and consolidating outcome expectations for emotionally relevant stimuli (for a review see Schoenbaum et al., 2006). Those with a history of SUD have been shown to have impaired OFC functioning, which is thought to maintain the disorder. OFC impairment in this population has been associated with limited recognition of reward reduction and increased negative outcomes associated with continued substance misuse. Reviews have suggested that unipolar depression is also associated with diminished OFC functioning (Rogers et al., 2004). Specifically, those with depression have reduced OFC volume (Drevets, 2007) and fail to show a shift in responding when the reward or punishment contingences for a given task change. The presence of both a history of SUD and elevated depression symptoms may be attributed to significantly altered functioning of the OFC. This impairment may be associated with the overall worsening of response inhibition that was observed in the current study. Additional neuroimaging work is necessary to test this hypothesis, however.

Anxiety symptoms were found to moderate the difference in SSRT between positive and neutral stimuli, but not between negative and neutral stimuli. This result is consistent with related work suggesting that state anxiety alone has a minimal impact on response inhibition (Grillon et al., 2017). However, elevated trait and state anxiety further impair response inhibition. A similar effect was found in the current study in that the highest SSRTs were found for those with elevated trait anxiety when inhibiting to a negative stimulus. However, elevated state anxiety alone did not further exacerbate this difference. Participants with elevated anxiety did, however, demonstrate less of a difference between neutral and positive stimuli than those with low anxiety. This shift is attributed to the speed with which those with anxiety process threat. Anxiety disorders are associated with a bias towards threatening stimuli that is characterized by faster processing of threatening information (Shechner & Bar-Haim, 2016). For those with elevated anxiety, positive faces in the current study were likely quickly interpreted as non-threatening, which allowed for improved SSRT. However, negative and neutral faces were likely processed as threatening, which resulted in an interference effect. Additional work on the processing of positive information, however, is needed to verify this hypothesis.

The current study had several limitations of note. First, all participants presented with a history of polysubstance use. It is unclear if these findings are specific to the use of a specific substance or set of substances or if they are characteristic of substance use more broadly. Second, the amount and duration of substance use varied considerably across participants. It is unclear if these findings would be obtained in a sample of those with a history of chronic use or in a sample of those using actively. Participants’ psychopathology was measured by self-report surveys. Use of clinician-administered measures of psychopathological severity should be considered for future studies. Finally, the current study used angry facial expressions as the negative stimuli. Related work has used more salient emotional stimuli such as images of traumatic events or drug related cues to evoke a particular emotion. It is unclear how more salient emotional stimuli would affect the results. In addition, happy, calm, and angry facial expressions were chosen to represent positive, neutral, and negatively valenced stimuli respectively. It is unclear to what extent other types of emotional expressions (e.g., excitement, fear, sadness, etc.) may impact SST performance in this population. Finally, the ethnic and racial diversity of the sample was quite low due to the location in which the study was conducted. Efforts should be taken to replicate these findings in a more diverse set of individuals to determine the presence or absence of cultural variables on the observed effects.

In conclusion, the findings of the present study demonstrate how psychopathology, particularly depression, and emotional stimuli interact to affect response inhibition in those with a history of SUD. The results highlight that both emotion regulation and an understanding of the context in which stimuli are encountered are necessary to determine an individual’s ability to appropriately inhibit a response. Poorer response inhibition among those with depression and anxiety towards negative stimuli may explain why individuals with co-occurring SUD and psychopathology have greater impairment and poorer treatment outcomes (Davis et al., 2010). The results highlight the need for treatment that addresses co-occurring conditions.

Highlights.

• Psychopathology is associated with poor response inhibition in SUD history.

• Depression moderated the relation between emotional stimuli and response inhibition.

• Low anxiety was associated with poor response inhibition to emotional stimuli.

• Response inhibition did not differ between positive and negative emotional stimuli.