Neural activation during inhibition predicts initiation of substance use in adolescence

Abstract

Background

Problems inhibiting non-adaptive behaviors have been linked to an increased risk for substance use and other risk taking behaviors in adolescence. This study examines the hypothesis that abnormalities in neural activation during inhibition in early adolescence may predict subsequent substance involvement.

Methods

Thirty eight adolescents from local area middle schools, ages 12–14, with very limited histories of substance use, underwent functional magnetic resonance imaging (fMRI) as they performed a go/no-go task of response inhibition and response selection. Adolescents and their parents were then followed annually with interviews covering substance use and other behaviors. Based on follow-up data, youth were classified as transitioning to heavy use of alcohol (TU; n = 21), or as healthy controls (CON; n = 17).

Results

At baseline, prior to the onset of use, youth who later transitioned into heavy use of alcohol showed significantly less activation than those who went on to remain non to minimal users throughout adolescence. Activation reductions in TU at baseline were seen on no-go trials in 12 brain regions, including right inferior frontal gyrus, left dorsal and medial frontal areas, bilateral motor cortex, cingulate gyrus, left putamen, bilateral middle temporal gyri, and bilateral inferior parietal lobules (corrected p < .01, each cluster ≥32 contiguous voxels).

Conclusions

These results support the hypothesis that less neural activity during response inhibition demands predicts future involvement with problem behaviors such as alcohol and other substance use.

Introduction

For many, substance (e.g., alcohol and cannabis) use initiation begins in middle school, with rates of use increasing dramatically through adolescence. For instance, according to recent community surveys, alcohol is the most widely used intoxicant among 8th graders, with 15% reporting drinking in the past 30 days, expanding to 44% by 12th grade. More alarmingly, 8% of 8th graders and 25% of 12th graders report past 2-week heavy drinking (≥5 drinks per occasion) (Johnston et al., 2009). Similarly, 12% of 8th graders report cannabis use in the past year, and rates increase to 30% by 12th grade; other drug use increases from 7% to 17% from 8th to 12th grade for annual or more frequent use (Johnston et al., 2009). These data indicate that a sizable proportion of adolescents use substances.

A range of risk factors influences the probability of initiation and intensity of substance use in adolescence. Youth with a family history of alcohol use disorder (AUD) are more likely to initiate drinking and develop AUD than youth without such familial background (Cloninger et al., 1986, Schuckit, 1985). Initiating alcohol use at an earlier age (i.e., age <14) is also linked to an increased risk for future AUD (DeWit et al., 2000, Grant and Dawson, 1997). Additional factors influencing adolescent substance use risk are childhood behavioral problems (Stice et al., 1998, Zucker et al., 2008) and disinhibitory features (Hill et al., 1999, Kirisci et al., 2006) such as impulsivity (Ernst et al., 2006, McGue et al., 2001), aggression (Brook et al., 1992, Clapper et al., 1995), and conduct disorder (Boyle and Offord, 1991, King et al., 2004, Kuperman et al., 2005). Early pubertal onset (Downing and Bellis, 2009), positive alcohol or other drug use expectancies (Killen et al., 1996, Simons-Morton, 2004), and depressive symptoms (Henry et al., 1993) also potentiate the likelihood of substance involvement. Understanding the mechanisms by which these factors contribute to adolescent substance use may refine the foci of secondary prevention efforts.

Inhibition, the ability to suppress inappropriate or impulsive behaviors, is a risk factor of particular interest as it may relate to future adolescent substance use either directly or indirectly, through externalizing behaviors (Iacono et al., 2008). Inhibition matures across childhood with the capacity to inhibit a response apparent as early as infancy (Diamond and Goldman-Rakic, 1989) yet the ability to consistently inhibit a behavior or action developing well into adolescence (Luna et al., 2004, Luna et al., 2001). Multiple inhibitory tasks show increasing cognitive control with age, including Stroop (Tipper et al., 1989), go/no-go (Luciana and Nelson, 1998), stop signal (Greenberg and Waldman, 1993, Ridderinkhof et al., 1999, Williams et al., 1999), and antisaccade tasks (Fischer and Weber, 1998, Fukushima et al., 2000, Luna et al., 2004, Munoz et al., 1998). For example, go/no-go studies have shown decreasing false alarm rates across development with young adults recording significantly less false alarms than children (Casey et al., 1997, Jonkman, 2006). Similarly, antisaccade studies have shown that children commit more errors than adults, with rates of correct inhibitions improving through adolescence into adulthood (Fischer and Weber, 1998, Fukushima et al., 2000, Luna et al., 2004, Munoz et al., 1998).

Underlying the development of more efficient inhibitory control are changes in brain structure and function subserving these higher order processes. Synaptic pruning, cortical thinning, and myelination of white matter tracts occur throughout adolescence (Gogtay et al., 2004, Sowell et al., 2002). Prefrontal cortex activity appears to change prominently as inhibitory networks develop. Increased BOLD response in prefrontal regions has been reported among children (age <12) compared to adolescents and adults during go/no-go (Booth et al., 2003, Casey et al., 1997) and antisaccade tasks (Velanova et al., 2008). However, less prefrontal response in children and adolescents compared to adults has been reported during flanker (Bunge et al., 2002), stop (Rubia et al., 2000), and Stroop tasks (Adleman et al., 2002), and one investigation found adolescents and adults to activate prefrontal cortex similarly during an antisaccade task (Velanova et al., 2008). These inconsistent results could be due to small sample sizes, chance, or differences in developmental stages across studies.

Some investigations offer explanations for the mixed results in the development of inhibition and the prefrontal cortex. Tamm et al. (2002) suggest a differential effect for regions within the frontal lobe, with increasing activation in inferior frontal, orbital frontal, and insular gyri, and decreasing activation in middle and superior frontal areas across adolescence. A meta-analysis of functional magnetic resonance imaging (fMRI) studies using various go/no-go tasks in adults suggest that differences among studies are task dependent; tasks with greater cognitive demands evoke additional brain regions that underlie cognitive functions beyond response selection (e.g., working memory) (Simmonds et al., 2008). The same theory could be applied to all developmental studies of inhibition where differences in neural recruitment during inhibition could be accounted for by the type (e.g., go/no-go stop, Stroop, antisaccade) and complexity of the inhibitory tasks.

In addition to the prefrontal cortex, other neural regions are recruited during tasks of inhibition. Activation within the anterior cingulate, parietal and occipital lobes have been observed during tasks requiring response inhibition, and have been more pronounced among adults relative to children (Adleman et al., 2002, Velanova et al., 2008). Subcortically, activation has also been shown in the caudate, though more so in adolescents compared to adults (Rubia et al., 2000). These investigations highlight the complexity of the neural mechanisms responsible for the maturation of inhibition.

Understanding the development of inhibition at the behavioral and neural level is of great importance when considering inhibition within the context of substance use. The few studies examining brain function and response inhibition in youth at risk for developing substance use disorders suggest that less BOLD response in frontal regions is related to risk for substance use; however, the relationship between this frontal activation and future substance use among adolescents is yet to be determined. Among adults with AUD, deficits in executive functioning, specifically in inhibition, persist into abstinence periods, suggesting that functions underlying planning, decision making and response inhibition may be uniquely affected by alcohol consumption (Davies et al., 2005, Fein et al., 2004, Noel et al., 2007). The question remains whether aberrant inhibitory networks in childhood and adolescence are themselves risk factors for substance use or rather substance use interrupts inhibition and the maturation of neuronal substrates of inhibition. The aim of the present study was to investigate the hypothesis that altered neural activation during inhibition in adolescents with limited substance use experience predicts subsequent substance involvement.

To that end, adolescents who had not yet had any notable substance use involvement (i.e., at project baseline) completed a go/no-go task during fMRI, and were then followed quarterly to assess for emergence and persistence of substance use (i.e., follow up). Based on follow-up (M = 4.2 years) alcohol use information, adolescents were classified into one of two alcohol use groups: Transitioned to Heavy Use (TU) and Healthy Controls (CON). Some TU youth also endorsed other substance use at follow-up, predominantly cannabis. We examined the relationship between baseline BOLD response during inhibitory trials of a go/no-go task, and follow-up substance use outcome. Previous studies using this go/no-go task in high-risk adolescents have observed less activation in frontal and posterior parietal regions than low-risk youth (Anderson et al., 2005, Schweinsburg et al., 2004b). Based on previous fMRI studies concerning development, familial risk for substance use, and prediction of relapse (Paulus et al., 2005), we hypothesized that at baseline, when no youths had more than 3 lifetime substance use episodes, those who would transition into heavy alcohol use by mid-adolescence (TU youth) would show less BOLD response in frontal regions than youth who would remain relatively substance-free throughout this developmental period (CON youth), and this would be specific to inhibition trials rather than response selection trials.