Sensitivity to the Disinhibiting Effect of Alcohol: The Role of Trait Impulsivity and Sex Differences

Abstract

Objective:

Higher trait impulsivity is associated with more impulsive responding on certain behavioral measures of disinhibition. Additionally, behavioral disinhibition is acutely elevated following alcohol consumption. The current study examined the possibility that trait impulsivity may predict individual differences in sensitivity to the disinhibiting effect of alcohol. Specifically, the current study tested the hypothesis that those with elevated trait impulsivity also experience heightened sensitivity to the disinhibiting effect of alcohol, which might further compound their tendency toward impulsive action.

Method:

To test this hypothesis, data from six studies were aggregated to comprise a sample of 190 young adults. Participants completed the Barratt Impulsiveness Scale-11, and behavioral disinhibition was assessed using a cued go/no-go task following consumption of 0.65 g/kg alcohol and a placebo.

Results:

Alcohol increased disinhibition overall, but higher impulsivity did not predict increased sensitivity to alcohol-induced disinhibition. In men, higher levels of trait impulsivity predicted heightened disinhibition in the unintoxicated state following placebo, but this relationship was not present in women.

Conclusions:

These findings suggest significant sex differences in the relationship between trait impulsivity and disinhibition. This sex difference may explain inconsistent research findings in studies assessing links between trait and behavioral measures of impulsivity. The data also point to trait impulsivity and sensitivity to alcohol-induced disinhibition as independent constructs.

Introduction

Impulsivity has been defined as acting rashly without forethought or consideration for future consequences, and it is a symptom of many mental health disorders (e.g., conduct disorder, attention deficit hyperactivity disorder [ADHD], borderline personality disorder, and alcohol use disorder [AUD]; American Psychological Association, 2013). There are well-established sex differences in typical levels of impulsivity such that males tend to be more impulsive than females (Cross et al., 2013; Shulman et al., 2014). This likely contributes to certain high-risk behaviors which are more commonly found in men. For example, men make up 97% of dangerous driving offenses (Beattie, 2008; Norris et al., 2000) and are more likely to engage in extreme and dangerous sports (Harris et al., 2006). This disparity may also relate to the higher rates of AUD observed in men versus women (Grant, 2015; Keyes et al., 2008; Wilsnack et al., 2009). However, the magnitude of sex differences in impulsivity varies greatly depending on the facet of impulsivity being assessed and the method with which it is measured (Adan et al., 2017; Cross et al., 2011). Furthermore, heightened sensitivity to alcohol’s impairing effects on behavioral aspects of impulsivity may be associated with at-risk drinking (Fillmore, 2003). This study tests this relationship by examining the degree to which individual differences in trait impulsivity predict individuals’ sensitivity to the disinhibiting effect of alcohol among young adults in the laboratory. The study also tests sex differences in this relationship as the literature examining this relationship shows mixed findings.

The multifaceted nature of impulsivity has resulted in the development of a breadth of measures to assess it. For example, personality-based measures of impulsivity conceptualize impulsivity as an enduring and relatively stable trait that varies normatively within the population

(Patton et al., 1995). As a trait, impulsivity is typically measured using self-report questionnaires, which are thought to capture both cognitive and affective processes (DeYoung, 2010). Commonly used self-report questionnaires include the UPPS-P (Whiteside & Lynam, 2001) and the Barratt Impulsiveness Scale (BIS-11; Patton et al., 1995; Vasconcelos et al., 2012). Additionally, subscales of common personality inventories such as the Eysenck Personality Inventory and the Minnesota Multiphasic Personality Inventory (MMPI-2; Butcher et al., 1989) are also sometimes used to assess impulsivity. Generally, trait impulsivity measures contain subscales which capture more specific impulsivity facets. Sex differences are inconsistent across different subscales and are more pronounced for certain facets (Adan et al., 2016).

Trait measures of impulsivity have reliably shown relationships with alcohol-related problems including alcohol dependence and AUD. Heightened scores on measures of impulsivity as well as on more targeted assessments of impulsivity facets are associated with greater risk for heavy drinking (Colder & O’Connor, 2002; Field et al., 2008). Higher scores of impulsivity, as measured by the BIS-11, are associated with heavier drinking in samples of light and moderate drinkers without a history of AUD (Berey et al., 2017). Additionally, studies using clinical samples comprised of those with AUD suggest that they endorse elevated trait impulsivity compared to healthy controls (for review, see, Verdejo-García et al., 2008).

Despite strong links between impulsivity and risk for heavy drinking and AUD, personality-based measures are limited in that they rely on self-report, which in turn is dependent on individuals’ level of insight and ability to accurately report their behavioral patterns. To overcome these limitations, laboratory tasks have been developed as a means of capturing acute, more transient impulsive actions. These tasks tap into a number of cognitive processes and specific behavioral aspects of impulsive behavior including behavioral disinhibition, which is especially relevant for considering liability for alcohol-related problems. Behavioral disinhibition is the inability to inhibit a prepotent or instigated response, and poor performance on these tasks reflects an inability to suppress behavioral responses (Fillmore, 2003). Increased behavioral disinhibition may account for continued alcohol consumption despite cues that one should stop drinking (e.g., feelings of drunkenness, sleepiness, sickness, incoherence, and loss of motor coordination; Fillmore, 2003; Gan et al., 2014; Giancola et al., 2010). Laboratory tasks allow for the identification of deficits in mechanisms that may be implicated in heightened impulsivity, including behavioral patterns characteristic of those with ADHD (Adams et al., 2011; Tannock, 1998) and DUI offenders (Kasar et al., 2010; Miller & Fillmore, 2014). They also allow for the evaluation of state-dependent changes in behaviors underlying impulsivity, including changes under the influence of drugs or alcohol (Fillmore, 2003; Marczinski & Fillmore, 2003; Vera et al., 2018) or during different affective states (Herman et al., 2018).

Like measures of trait impulsivity, sex differences are observed in certain task-based assessments of impulsive action. Men tend to show greater behavioral disinhibition, higher scores on measures of risk taking (Cross et al., 2011) and poorer performance on motor impulsivity tasks (Lage et al., 2013). However, for some domains of impulsive behavior, findings related to sex differences are equivocal, suggesting that sex differences may depend on the specific task used and the behavioral aspect of impulsivity being measured. On tasks of behavioral disinhibition, sex differences appear to be moderated by drinking habits. Heavy drinking women display greater disinhibition compared to heavy drinking men; however, this sex difference is not present among non-risk drinkers (Weafer et al., 2015). Furthermore, heavy drinkers show poorer behavioral disinhibition relative to non-risk drinkers (Adams et al., 2011; Tannock, 1998). Rubio et al. (2008) provided evidence that this relationship is causal by demonstrating that impulsive responding on a behavioral disinhibition task (i.e., the Stop-signal Task) prospectively predicted alcohol dependence at a 4-year follow-up. Additionally, another prospective study found further support for the causal role of heightened levels of behavioral disinhibition in predicting at-risk drinking by demonstrating that increased disinhibition in adolescence predicted heavier alcohol consumption 6 months later (Fernie et al., 2013).

Alcohol produces acute increases in disinhibition as measured by behavioral impulsivity tasks (Fillmore & Vogel-Sprott, 1999; Marczinski et al., 2005; Reynolds et al., 2006), and this may be particularly important in explaining the well-established relationship between alcohol use and impulsivity. Alcohol-induced behavioral disinhibition explains several socially inappropriate or dangerous behaviors characteristic of alcohol intoxication including aggression, risky sexual behavior, and driving under the influence (Babor et al., 1983; Fillmore & Van Dyke, 2020; Leeman et al., 2007). Additionally, performance on disinhibition tasks is sensitive to the effects of alcohol in a dose-dependent manner such that larger doses of alcohol are associated with higher levels of disinhibition (Marczinski & Fillmore, 2003). Studies from our laboratory have shown that increased sensitivity to the impairing effects of alcohol on behavioral disinhibition is associated with heavier drinking (Allen et al., 2021; Marczinski et al., 2007; Weafer & Fillmore, 2008). Moreover, the effects of alcohol on disinhibition are long-lasting (Miller & Fillmore, 2014), and individuals who are prone to sustained elevations in disinhibition after a dose of alcohol are at increased risk for over-consumption of alcohol and binge drinking (Allen et al., 2021). Furthermore, studies in our laboratory have also demonstrated a sex difference in the effects of alcohol on inhibitory control whereby men are more sensitive to alcohol-induced increases in disinhibition than women (Fillmore & Weafer, 2004).

In sum, sex differences in trait and task-based measures of impulsivity are commonly observed such that men are generally more impulsive than women. Heightened impulsivity is also associated with heavier alcohol use, and alcohol increases aspects of impulsivity such as behavioral disinhibition, an effect that is more pronounced in heavier drinkers (Fillmore & Weafer, 2004). Trait impulsivity may also be associated with individual differences in the acute effects of alcohol. In particular, those with elevated trait impulsivity may also experience heightened sensitivity to the disinhibiting effect of alcohol, further compounding their tendency toward impulsive action (Fillmore & Weafer, 2004). Some small-scale studies have examined the relationship between trait impulsivity and acute sensitivity to alcohol’s effects on behavioral aspects of impulsivity. These suggest that those with higher levels of sensation seeking and ADHD symptomology are more sensitive to the disinhibiting effect of alcohol (Fillmore et al., 2009; Hendershot et al., 2015). Although these studies hint that impulsive traits might account for individual differences in sensitivity to alcohol’s disinhibiting effect, they are limited in terms of sample size. The relationship between trait impulsivity and sensitivity to the disinhibiting effect of alcohol has important implications, specifically sustained levels of disinhibition may prolong a drinking episode and lead to alcohol overconsumption or binge drinking. Thus, the purpose of this study was to test this relationship in a large sample of healthy adult drinkers (N = 190).

The current study tested the hypothesis that drinkers with greater trait impulsivity might display greater sensitivity to the acute disinhibiting effect of alcohol compared to those with lower trait impulsivity. Reliable tests of such associations require large sample sizes, and studies utilizing laboratory tasks to assess the acute effects of alcohol are often limited in sample size, precluding the assessment of individual differences. The current study overcomes this limitation by aggregating data from prior studies completed in our laboratory. The study tested the degree to which participants’ self-reported trait impulsivity predicted individual differences in their sensitivity to the acute disinhibiting effect of 0.65 g/kg alcohol. Data were aggregated from six separate studies to test the relationship, yielding a sample size of 190 participants. The studies were placebo-controlled experiments that measured trait impulsivity and the acute disinhibiting effect of alcohol in groups of healthy young adults. Based on considerable sex differences in impulsivity and its behavioral aspects, and given we were statistically powered to do so, we also evaluated sex differences in this relationship.

Method

Participants

An aggregate sample of young adult social drinkers was comprised of participants from six separate studies conducted in the investigators’ laboratory. Participants (N = 190) were between 21- and 33-years-old (M = 22.84, SD = 2.54) and were recruited from the community between 2005 and 2014 for participation in research projects assessing the effects of alcohol on behavioral and cognitive function (Fillmore et al., 2005, 2008; Fillmore & Weafer, 2012; Marczinski et al., 2007; Miller & Fillmore, 2014; Roberts et al., 2013). Only studies with common population, methodology, and testing procedures were included. These criteria included young adult participants with no history of AUD, counterbalanced, repeated dose design with two alcohol dose conditions (0.0 g/kg and 0.65 g/kg), dose effect measurement following the same time course, and common assessment of behavioral disinhibition and trait impulsivity. Volunteers were excluded from participation if they self-reported a history of head trauma or other central nervous system injury. Potential participants were also excluded if they reported a psychiatric disorder or a SUD. To further screen for AUD, the Short-Michigan Alcohol Screening Test (S-MAST; Selzer et al., 1975) was administered. The S-MAST is a 13 item self-report questionnaire with scores ranging from 0 to 13, with higher scores being indicative of more severe alcohol-related problems. Volunteers with a score of five or higher were excluded from participation.

Urine samples were tested for the presence of metabolites of amphetamine, methamphetamine, barbiturates, benzodiazepines, cocaine, opiates, methadone, phencyclidine, tricyclic antidepressants, and tetrahydrocannabinol (THC; ICUP Drugscreen; Instant Technologies, Norfolk, VA). Volunteers who tested positive for cannabis use and reported use within the past 4 days and those who tested positive for any other drug metabolites were excluded from participation. No volunteers who were pregnant or breast feeding participated in the research, as was determined by self-report measures and urine human chorionic gonadotropin levels. The aggregate sample was 84 women and 106 men. In terms of racial makeup, participants self-identified as Caucasian (n = 163), African American/Black (n = 21), or as other (n = 6).

Measures

Trait Impulsivity

Trait impulsivity was measured by the Barrett Impulsiveness scale-11 (BIS-11; Patton et al., 1995). This 30-item self-report questionnaire includes questions such as “I plan tasks carefully” and “I do things without thinking.” Participants respond to each item, indicating how frequently each applies to them on a Likert scale from 1 (rarely/never) to 4 (almost always/always). Higher scores on the measure suggest higher levels of impulsivity, and the highest possible score is 120. The BIS-11 includes three second-order subscales including attentional, motor, and non-planning impulsiveness. Attentional impulsiveness is defined as the inability to concentrate, motor impulsiveness is the tendency to act without thinking, and non-planning impulsiveness is the lack of future planning or forethought. The BIS-11 demonstrates appropriate internal consistency, test-retest reliability, and criterion validity (Patton et al., 1995).

In our sample, Cronbach’s alpha was used to calculate internal consistency for the BIS-11 in its entirety and for each second-order subscale (BIS-11 Total, α = 0.82; Attention, α = 0.68; Motor, α = 0.61; Non-planning, α = 0.63).

Disinhibition

A cued go/no-go reaction time task was used to measure participants’ response inhibition to no-go targets and their reaction time to go targets (e.g., Fillmore and Weafer 2004). The task required finger presses on a keyboard and measured the ability to inhibit prepotent behavioral responses of executing a key press. Cues (vertical and horizontal rectangles) provided preliminary information regarding the type of target stimulus (i.e., go or no-go) that was likely to follow, and the cues had a high probability of signaling the correct target. Participants were instructed to press the forward slash (/) key on the keyboard as soon as a go target (green rectangle) appeared and to suppress the response when a no-go target (blue rectangle) was presented. The go cue conditions were of particular interest. Go cues generate response prepotency which speeds response time to go targets. However, subjects must overcome this response prepotency to inhibit the response if a no-go target is subsequently displayed. Response inhibition was measured by the proportion of no-go targets in which subjects failed to inhibit a response (p-inhibition failures) during the test. Disinhibition was indicated by a higher proportion of inhibition failures (i.e., greater p-inhibition failures). A test required approximately 15 minutes to complete. The validity of these types of tasks is well-established. Cued go/no-go tasks are sensitive to inhibitory deficits characteristic of brain injury (Malloy et al., 1993), ADHD (Tannock, 1998), and the impairing effects of alcohol (Fillmore & Weafer, 2012; Weafer & Fillmore, 2016).

Drinking habits

Drinking habits were measured using the Personal Drinking Habits Questionnaire

(PDHQ; Vogel-Sprott, 1992). This self-report questionnaire was used to sample participants’ typical drinking patterns, including consumption of excessive quantities of alcohol or frequent drinking. Participants responded to items according to their current and typical drinking habits at the time of administration. The PDHQ measured three aspects of participants’ drinking behavior including (1) frequency, or the typical number of drinking occasions per week, (2) dose, or milliliters of absolute alcohol consumed during a typical drinking episode per kilogram of body weight, and (3) duration, or typical time span (in hours) of a drinking occasion.

Procedure

Participants were recruited from the community using flyers, posters, online advertisements, and newspaper advertisements. All volunteers provided informed consent prior to participation and all parent studies were approved by the University of Kentucky Medical Institutional Review Board. All participants were compensated for their participation. Volunteers responded to advertisements by calling the laboratory and participated in a telephone screening procedure conducted by a research assistant. All eligible volunteers were told that the purpose of the study was to examine the effects of alcohol on behavior. All sessions were conducted at the Human Behavioral Pharmacology Laboratory at the University of Kentucky between the hours of 10 AM and 6 PM. Participants were required to abstain from alcohol for 24 h prior to each session and fast for 4 h prior to each session. Before testing sessions began, body weight was measured, a BAC of 0.0 g/kg was verified using a breathalyzer (Intoxilyzer, Model 400 [CMI Inc., Owensboro, KY]), and urine samples were collected to ensure that participants were negative for recent drug use and pregnancy.

Familiarization session

During the intake and familiarization session, participants were introduced to laboratory procedures and completed a practice session of the cued go/no-go task. Participants completed the BIS-11 to collect data on trait impulsivity. Information regarding drug and alcohol use history was obtained using the PDHQ and S-MAST. Information regarding participants’ health status, and general demographic information was also obtained via self-report.

Behavioral testing sessions

Participants were tested on two separate occasions under different doses of alcohol, 0.0 g/kg (placebo) and 0.65 g/kg. Each session, including familiarization, was conducted on a different day separated by a minimum of 1 day and a maximum of 7 days. The dose order was counterbalanced across participants and dose order was independent of sex (χ² = 0.13, p = 0.90). For women, 50% (n = 42) of participants received each of the two orders (placebo followed by alcohol, and vice versa). For men, 49% (n = 52) received placebo followed by alcohol, and 51% (n = 54) received alcohol followed by placebo.

During test sessions, participants received either a dose of 0.65 g/kg alcohol or a placebo dose of 0.0 g/kg. The size of the 0.65 g/kg dose was calculated based on participants’ body weight and was equally divided into two glasses with each containing one part 96.4% alcohol and three parts carbonated mix. Participants were given 2 min to consume each drink, and the second beverage was served 4 min after consumption of the first. This dose has been shown to produce an average peak BAC of 80 mg/100 ml about 60 min after consumption and has been shown to impair inhibitory control and slow reaction time as measured by the cued go/no-go task (Marczinski & Fillmore, 2003). The placebo beverages contained four parts carbonated mix at an equal volume to the active dose. A small amount of alcohol (3 ml) was floated on the surface of the beverage after dividing it into two equal glasses. The glasses were sprayed with an alcohol mist that resembled condensation and provided a strong scent of alcohol. Prior research indicates that participants believe these beverages contain alcohol (Fillmore & Vogel-Sprott, 1998). The timing of consumption of placebo beverages was identical to that of the active dose.

The cued go/no-go task was administered between 25 – 35 min after participants’ beverage consumption. BACs were verified using a breathalyzer. Breath samples were also obtained at 60 min, when peak BAC was expected. Because participants were recruited for the purpose of different studies, they also completed an additional task during test sessions, including a driving simulation task, a motor coordination task, a choice reaction time task, or a multi-sensory cued go/no-go task. However, these tasks were administered after the cued go/no-go task, so it is unlikely that cued go/no-go task performance would have been reactive. Participants remained in the lab until their BAC fell below 20 mg/100 ml, and transportation home by taxi was provided after the sessions. Upon completion of the final session, participants were paid and debriefed.

Data Analyses

The acute disinhibiting effect of alcohol in men and women was tested by a 2 Dose (placebo, 0.65 g/kg alcohol) × 2 Sex analysis of variance (ANOVA). Regression analyses tested the degree to which individual differences in trait impulsivity (i.e., BIS-11 scores) predicted participants’ unintoxicated level of disinhibition (i.e., following placebo) and their sensitivity to the disinhibiting effect following alcohol. Participants’ sensitivity to the disinhibiting effect of alcohol was calculated by subtracting their p-inhibition failures in the placebo condition from the p-inhibition failures following alcohol, such that larger sensitivity scores represented greater alcohol-induced increases in disinhibition. All variables were standardized (i.e., z-transformed) prior to analysis. Regression analyses also included sex and the interactions between sex and trait impulsivity to examine sex differences in relationships between trait impulsivity and disinhibition. Exploratory regression analyses were conducted to test for sex differences in the relationship between BIS-11 subscales of trait impulsivity (i.e., Attention, Motor, and Non-planning impulsiveness) and sensitivity to alcohol-induced disinhibition. All analyses were tested using a significance threshold of p < 0.05.

Results from analyses of difference scores for p-inhibition failures were supported by regression models testing the relationship between trait impulsivity and in the observed p-inhibition failures following alcohol while controlling for the observed p-inhibition failures following placebo. There were no differences in results between these covariance analyses and analyses using change scores. Because change scores provide a more direct indication of the magnitude of participants’ response to the dose administered, all reported analyses and figures use these scores to better illustrate dose effects.

Results

Drinking Habits

The sample’s drinking habits are reported in Table 1 for men and women. The sample reported a mean typical quantity of 4.91 drinks (SD = 2.46) and an average weekly drinking frequency of 2.16 days (SD = 1.14). The mean duration of their typical drinking occasion was 3.73 h (SD = 1.35). Results of two-sample t tests comparing drinking habits between men and women are reported in Table 1. Men drank significantly more frequently than women and reported a higher typical quantity and dose of alcohol consumed.

Table 1.

Descriptive statistics of participants’ drinking habits, BIS-11 scores, p-inhibition failures, and t test statistics comparing means in among men and women.

	Mean	SD	Minimum	Maximum	t	p-value
Frequency
Full Sample	2.16	1.14	0.08	7.00
Women	1.94	1.16	0.08	5.00	−2.41	0.02*
Men	2.34	1.10	0.21	7.00
Quantity
Full Sample	4.91	2.46	1.00	12.00
Women	3.80	1.78	1.00	8.00	−6.30	< 0.001*
Men	5.79	2.57	1.00	12.00
BIS-11
Full Sample	57.77	14.90	0.00	99.0
Women	52.85	15.64	0.00	79.00	−5.11	< 0.001*
Men	61.54	13.18	1.00	99.00
BIS-Attention
Full Sample	15.32	4.15	8.00	29.00
Women	14.37	3.51	8.00	24.00	−6.34	< 0.001*
Men	16.19	4.46	8.00	29.00
BIS-Motor
Full Sample	22.63	3.56	15.00	35.00
Women	21.42	2.83	15.00	29.00	−3.57	< 0.001*
Men	23.60	3.79	15.00	35.00
BIS-NP
Full Sample	21.81	5.28	10.00	39.00
Women	20.86	4.55	10.00	34.00	−5.20	< 0.001*
Men	22.55	5.70	10.00	39.00
Non-intoxicated pf
Full Sample	0.21	0.30	0.00	0.95
Women	0.12	0.21	0.00	0.95	−3.90	< 0.001*
Men	0.28	0.34	0.00	0.95
Sensitivity Scores
Full Sample	0.06	0.12	−0.35	0.56
Women	0.05	0.12	−0.16	0.52	−0.35	0.73
Men	0.06	0.123	−0.35	0.56

Frequency, typical number of days alcohol consumed per week; Quantity, typical number of drinks consumed per drinking occasion; Full Sample (N = 190), Women (N = 84), Men (N = 106). Non-intoxicated pf is p-inhibition failure following placebo. Sensitivity to alcohol-induced disinhibition was calculated by subtracting participants’ p-inhibition failures in the placebo condition from p-inhibition failures following alcohol.

^*Sig indicates a value of p < 0.05

Trait Impulsivity

Trait impulsivity scores on the BIS-11 and t tests comparisons of men and women are reported in Table 1. The mean impulsivity score for the entire sample was 57.8 (SD = 14.9) and men reported higher scores than women. Table 1 also reports statistics for BIS-11 second-order subscales, Attention, Motor, and Non-planning Impulsiveness. Men self-reported higher scores than women on all three subscales.

Blood Alcohol Concentrations

The 0.65-g/kg alcohol dose produced a mean peak BAC of 80.8 mg/100 ml (SD = 14.5) approximately 60 to 70 min post alcohol administration. The mean BAC of participants during their performance of the cued go/no-go task was 70.4 mg/100ml (SD = 20.2). There were no significant BAC differences between men and women at peak BAC or during testing (p’s >

0.262). No detectable BAC was observed during the placebo session.

Acute Effect of Alcohol on Disinhibition

A 2 Dose × 2 Sex ANOVA obtained a significant main effect of dose, F(3,188) = 37.95, p > 0.001, and main effect of sex, F(3, 188) = 15.30, p < 0.001. No interaction was observed. Means are plotted in Figure 1, which shows p-inhibition failures increased following alcohol (M = 0.27, SD = 0.30) compared with placebo (M = 0.21, SD = .30). Figure 1 also shows that, overall, men committed more inhibitory failures than women following the placebo and active dose of alcohol.

Figure 1.

Mean p-inhibitory failures and RT among men and women following placebo and alcohol dose conditions. The capped vertical lines show the standard errors of the mean.

A 2 Dose × 2 Sex ANOVA of RTs in the go cue condition obtained a significant main effect of dose, F(3, 188) = 4.94, p = 0.03. Figure 1 shows that alcohol slowed RT compared with placebo. No significant main effect of sex or interaction was observed.

Trait Impulsivity as a Predictor of Behavioral Disinhibition in the Unintoxicated State

Trait impulsivity was examined as a predictor of participants’ disinhibition on the cued go/no-go task when alcohol-free (i.e., following placebo) using a hierarchical regression model. The regression analysis also examined the role of sex, which was dummy coded, and the interaction between sex and trait impulsivity as predictors of disinhibition. Results from these analyses are reported in Table 2. In step one of the regression model, sex and trait impulsivity scores were entered as factors predicting disinhibition. A significant main effect of sex was obtained whereby men reported higher levels of trait impulsivity than women. A main effect of trait impulsivity score was also obtained; Higher p-inhibition failures were associated with higher trait impulsivity. In step two, the interaction term was entered and accounted for a significant proportion of variance in trait impulsivity scores. The interaction is plotted in Figure 2. The figure shows a positive relationship between trait impulsivity and disinhibition in men but not in women. This was confirmed by tests of the slopes for the zero-order regression of trait impulsivity on disinhibition which was significant for men (b = 0.55, p < 0.001), but not for women (b = 0.12, p = 0.12).

Table 2.

Statistics obtained from regression models of sex and BIS-11 scores on disinhibition in a non-intoxicated state and sensitivity scores

	Non-Intoxicated Disinhibition (df = 2, 183)			Sensitivity to Alcohol-Induced Disinhibition (df = 3, 182)
Model	R2	F-Ratio	p-value	R2	F-Ratio	p-value
Step 1	0.16			0.005
Sex		4.82	0.03*		0.34	0.56
BIS-Total		19.54	< 0.001*		0.74	0.39
Step 2	0.20			0.008
Sex		5.40	0.021*		0.31	0.58
BIS-Total		21.32	< 0.001*		0.79	0.38
Sex × BIS-Total		8.68	0.004*		0.62	0.43
Step 1	0.21			0.03
Sex		7.17	0.008*		0.66	0.42
BIS-Attention		32.49	<0.001*		5.12	0.03*
Step 2	0.23			0.03
Sex		8.40	0.004*		0.57	0.45
BIS-Attn		22.33	< 0.001*		3.84	0.05*
Sex × BIS-Attention		4.24	0.04*		0.25	0.62
Step 1	0.10			0.001
Sex		8.01	0.005*		0.13	0.72
BIS-Motor		6.79	0.01*		0.06	0.81
Step 2	0.11			0.002
Sex		8.19	0.005*		0.09	0.67
BIS-Motor		4.94	0.03*		0.004	0.95
Sex × BIS-Motor		0.21	0.65		0.16	0.69
Step 1	0.17			0.006
Sex		9.85	0.002*		0.45	0.51
BIS-Non-Planning		22.62	< 0.001*		0.91	0.34
Step 2	0.18			0.007
Sex		10.40	0.001*		0.42	0.52
BIS-NP		16.72	< 0.001*		0.67	0.41
Sex × BIS-Non-Planning		1.43	0.23		0.05	0.82

Non-intoxicated disinhibition is p-inhibition failure following placebo. Sensitivity to alcohol-induced disinhibition was calculated by subtracting participants’ p-inhibition failures in the placebo condition from p-inhibition failures following alcohol. Larger sensitivity scores represented greater alcohol-induced disinhibition. All variables were standardized prior to regression models, and Sex was dummy coded (female = 1, male = 2).

^*Sig indicates a value of p < 0.05

Figure 2.

Regression lines relating trait impulsivity (BIS-11 scores in SD units; Low = −1 SD, Average = 0 SD, and High = +1 SD) and behavioral disinhibition following placebo for males and females.

Trait Impulsivity as a Predictor of Sensitivity to Alcohol-Induced Disinhibition

Trait impulsivity was also examined as a predictor of participants’ sensitivity to the acute disinhibiting effect of alcohol. Acute sensitivity for each participant was calculated by subtracting their p-inhibition failures in the placebo condition from their p-inhibition failures following alcohol. Larger difference scores indicated greater sensitivity to the disinhibiting effect of alcohol. In step one of the hierarchical regression model, sex and trait impulsivity were entered as factors predicting sensitivity to alcohol-induced disinhibition. Step two tested the degree to which their interaction accounted for additional variance in sensitivity to alcohol-induced disinhibition. Results from this analysis are reported in Table 2, and no predictors (i.e., sex, impulsivity, interaction term) accounted for significant variance in participants’ sensitivity to alcohol-induced disinhibition (ps > 0.35).

Given the possibility that drinking habits might have influenced participants’ sensitivity to the acute effect of alcohol on disinhibition in the active dose condition (0.65 g/kg), we tested the same regression models including participants’ drinking habits (typical number of drinks and weekly frequency of drinking) as covariates. Results obtained from regression analyses controlling for these drinking habits remained the same. Moreover, zero-order correlations showed no relationships between

Exploratory Analyses of Second Order BIS-11 Factors as Predictors of Disinhibition

Given the relationship observed between BIS-11 full-scale trait impulsivity scores and unintoxicated levels of behavioral disinhibition in men but not women, we further explored the relationships of the second-order factors of trait impulsivity to men and women’s levels of behavioral disinhibition. Second order factors of the BIS-11 (i.e., Attention, Motor, and Non-planning Impulsiveness) are frequently utilized to consider how independent impulsivity facets may be differentially relevant within an individual or in their relationship(s) with other outcome variables. In separate regression analyses for each higher-order factor, we regressed p-inhibition failures following placebo on the higher-order factor. Regression analyses examined the possible role of sex by including sex and its interaction with the higher-order factor as additional predictors in the regression model. Table 2 reports the results and shows that all second-order factors of trait impulsivity accounted for significant variance in disinhibition. Higher scores on each impulsivity factor were associated with greater disinhibition. In addition, Attention Impulsiveness scores interacted with sex to account for significant variance in disinhibition whereby the relationship between attentional impulsivity and disinhibition was stronger for men than women.

A similar set of regression models examined the role or BIS-11 subscales, sex, and their interaction as predictors of sensitivity to alcohol-induced disinhibition. Table 2 reports the results and shows that only scores on the Attention Impulsiveness subscale were associated with sensitivity to the disinhibiting effect of alcohol, where higher Attention Impulsiveness is associated with lower sensitivity to disinhibition following alcohol, controlling for the effect of sex. Motor and Non-planning Impulsiveness scores did not relate to sensitivity scores and there were no significant sex differences in relationships between BIS-11 subscales and sensitivity scores.

Discussion

This study examined the degree to which individual differences in impulsivity predicted sensitivity to the disinhibiting effects of alcohol among young adults. In order to be sufficiently statistically powered to test individual differences and sex differences, data were aggregated from prior studies conducted in our laboratory. Participants completed a measure of trait impulsivity and a cued go/no-go task that measured behavioral disinhibition following an active dose of 0.65-g/kg alcohol and a placebo. Alcohol significantly increased levels of disinhibition and slowed reaction time compared with placebo. Men self-reported more frequent and heavier drinking than women. Men also reported higher levels of trait impulsivity and, on average, were more disinhibited on the cued go/no-go task. However, there was no sex difference observed in sensitivity to the disinhibiting effect of alcohol. Considerable individual differences were observed in trait impulsivity and in the sensitivity to the disinhibiting effect of alcohol. Trait impulsivity in men predicted greater disinhibition following placebo but no such relationship was found among women. Trait impulsivity did not predict individual differences in sensitivity to the disinhibiting effect of 0.65 g/kg alcohol in either men or women.

Our findings suggest a relationship between trait impulsivity and a task-based assessment of behavioral disinhibition in an unintoxicated state, which is consistent with studies showing similar associations between trait impulsivity and task-based measures of the specific components of impulsivity, such as response inhibition (Enticott et al., 2006; Logan et al., 1997; Castellanos-Ryan et al., 2011; Swann et al., 2002; Gay et al., 2008; Spinella, 2004). However, other studies have observed little to no association between trait impulsivity and task-based measures of its components (Jacobs et al., 2010; Horn et al., 2003; Reynolds et al., 2006; Franken & Muris, 2005). There may be many reasons for such equivocal findings. It is often argued that task-based assessment of behavioral control may have little overlap with self-report assessments of impulsivity as a stable trait. Trait measures are based on subjective reports about one’s ability to exercise some degree of sustained control over emotional and motivational states in everyday situations that often involves anticipating future rewards and punishments for immediate actions (Baskin-Sommers et al., 2012; Dougherty et al., 2005). By contrast task-based measures assess impulsivity as the simple and extremely brief control over one’s immediate reaction to a stimulus presented in a laboratory (Baskin-Sommers et al., 2012). Compounding the problem, studies of the relation between trait impulsivity and task-based measures of its components are often based on small sample sizes (i.e., between 20 and 70; Franken & Muris, 2005; Horn et al., 2003; Jacob et al., 2010; Reynolds et al., 2006; Wingrove & Bond, 1997). The size of our sample (N = 190) may have provided the statistical power necessary to detect the relationship in this case.

We found that the relationship between trait impulsivity and unintoxicated behavioral disinhibition was present only in men, which is consistent with findings from a meta-analysis which found a stronger relationship between self-report and lab impulsivity task performance in studies with greater numbers of men (Cyders & Coskunpinar, 2011). Based on our sizeable sample of women (n = 86), it is unlikely that the failure to observe the relationship in women was a result of being underpowered. Additionally, restricted range is an unlikely explanation as ranges for trait impulsivity and behavioral disinhibition scores were nearly identical in the men and women.

One factor which may relate to the lack of the relationship between trait impulsivity and unintoxicated disinhibition among women is the possibility that the BIS-11 may have poorer construct validity for women than men. The nature and magnitude of sex differences in trait impulsivity varies depending on the specific facet of impulsivity being examined (Adan et al., 2017). In the current study, men reported higher impulsivity on all three subscales. Therefore, the BIS-11 may tap into facets which are especially relevant for patterns of impulsivity seen in men compared to women. Prior research suggests that affective context, especially negative emotional experiences, may be particularly relevant for the onset of impulsive behavior among women (d’Acremont & Van der Linden, 2005; Costa et al., 2001), and these factors are not captured by the BIS-11. For example, drinking to cope with emotions is thought to be especially relevant for women (Becker et al., 2012). Furthermore, the tendency to act rashly when experiencing negative emotions (i.e., negative urgency) has a strong association with certain psychopathologies and symptoms including AUD, BPD, aggression, and suicidality (Berg et al., 2015). Negative urgency is also thought to contribute to disordered eating (Davis-Becker et al., 2014; Eichen et al., 2016) and risky sexual behavior in women (Deckman & DeWall, 2011).

Examination of the relationships between each trait impulsivity subscale and disinhibition following placebo yielded sex differences, which were not consistent across subscales. Attentional impulsivity was especially predictive of behavioral disinhibition among men, while motor and non-planning impulsivity facets predicted unintoxicated behavioral disinhibition equally among men and women. This is consistent with previous research suggesting attention may be especially important in modulating inhibitory control in men relative to women (Omura & Kusumoto, 2015). It also points to inconsistent sex differences in impulsivity facets. These inconsistent sex differences in addition to the possibility of differential validity of impulsivity measures among men and women may be implicated in inconsistent research findings across studies. Furthermore, many studies over-sample women due to reliance on convenience samples comprised primarily of undergraduate psychology students (Castellanos-Ryan et al., 2011; Franken & Muris, 2005; Jacob et al., 2010; Swann et al., 2002), which could inadvertently limit their ability to detect relationships between trait and task-based measures of impulsivity.

No relationship between trait impulsivity and sensitivity to alcohol-induced behavioral disinhibition was observed following a 0.65-g/kg dose of alcohol. This finding is unlikely to be a result of restriction of range as there were pronounced individual differences in sensitivity to the disinhibiting effect of alcohol. We tested for the possibility that participants’ drinking habits might have mediated the relationship between trait impulsivity and sensitivity to alcohol’s disinhibiting effect. However, drinking habits bore no relationship to participants’ sensitivity to the disinhibiting effect of alcohol. This is not surprising as participants were young adult drinkers with no history of an AUD, especially the prolonged patterns of heavy drinking that would be necessary to produce tolerance to its acute behavioral effects. This study examined only one dose of alcohol 0.65 g/kg yielding an approximate BAC of 80 mg/100 ml. It is possible that trait impulsivity might be more predictive of alcohol-induced disinhibition at higher BACs when inhibitory control is under greater disruption by the drug. Our knowledge of how factors such as impulsivity might predict the impairing effects of alcohol following different doses or in response to different BACs remains quite limited.

The cued go/no-go task used in the current study used geometric shapes and colors as stimuli to signal the execution and inhibition of responses. Visual images that signal alcohol (e.g., pictures of drinks and bottles) can also be used to signal such actions and could have greater ecological validity for understanding the role of inhibitory control on drinking behavior. Indeed, studies from our laboratory have shown that exposure to alcohol-related cues may exacerbate alcohol’s disruptive effect on inhibitory control and with greater cue-induced disinhibition predicting heavier drinking (e.g., Weafer & Fillmore, 2015). Future research examining the relationship between trait impulsivity and behavioral disinhibition following exposure to alcohol-related cues is needed.

The current study is also limited by its use of only a single behavioral task, disinhibition. It could be that the impairing effects of alcohol on other behavioral aspects of impulsivity are more heavily influenced by trait impulsivity. For example, higher trait impulsivity may not relate one’s acute inability to suppress a behavioral response while intoxicated but may increase sensitivity toward impulsive decision-making about the future while drinking. Delay discounting is an aspect of impulsivity, and it measures the inability to wait weeks or months for a larger reward. This is markedly different from behavioral disinhibition tasks focused on the inability to inhibit a prepotent response in a matter of milliseconds. Delay discounting may be of interest for future research because sex differences in performance on behavioral disinhibition and delay discounting tasks are inconsistent. Some research suggests that on delay discounting tasks, unlike behavioral disinhibition tasks, women behave more impulsively than men (Weafer & de Wit, 2014). Therefore, future research may wish to assess delay discounting as individual differences might show a stronger relationship with trait impulsivity among women. This study is also limited in that it utilizes only one measure of trait impulsivity. As previously mentioned, the BIS-11 may better reflect individual differences in impulsivity for men than women. Assessment tools that consider affective states which may contribute to impulsive behavior may help to elucidate the relationship between trait impulsivity and behavioral aspects of impulsivity among women.

The current study examined the relationship between trait impulsivity and behavioral disinhibition in a large sample of young adult drinkers. It also examined sex differences in this relationship given prior literature demonstrating differences in levels of impulsivity facets among men and women. We found no relationship between trait impulsivity and sensitivity to alcohol-induced disinhibition, suggesting that levels of trait impulsivity may not be directly related to the degree to which alcohol increases behavioral disinhibition at the BACs tested. The study also found that higher trait impulsivity scores on the BIS-11 were associated with higher behavioral disinhibition following placebo for men, while no such relationship was found for women. This finding may help to further elucidate mixed findings in studies examining the relationship between trait impulsivity and tasks designed to measure specific behavioral aspects of impulsivity. Additionally, our findings raise the possibility that the BIS-11 may have poorer construct validity for women compared to men.

Public Significance Statement.

The current study found that, among young adults, trait impulsivity relates to behavioral disinhibition as measured by a laboratory task that assesses the ability to inhibit a pre-potent response. However, trait impulsivity did not predict individual differences in sensitivity to alcohol-induced increases in disinhibition. The study also found that trait impulsivity better predicts disinhibition among men compared to women. These findings further elucidate our theoretical understanding of impulsivity facets, indicating the importance of considering sex differences in future impulsivity research as well as pointing to trait impulsivity and disinhibition as independent indicators of heavy drinking.