Relevant covariates thought to affect risk-taking performance and self-report of risk-taking were entered as covariates on Block 1 of the regression analyses. These included: age, family history of alcoholism, average # of alcoholic drinks per month, # of binge episodes in the prior 3 months, # of days/month of cannabis use in prior 3 months. BOLD response signal change variables for each ROI were entered on Block 2. The R 2 ∆ for the second step represented the degree to which BOLD activation was associated with baseline risk-taking performance or self-report of risk taking behavior, above and beyond the covariates listed in Step 1. There were a total of 28 tests run for Hypothesis 3 . Alpha was set at 0.05 for each test, as this was the first fMRI study of risk-taking in adolescent alcohol users and the analyses for Hypothesis 3 were considered exploratory. However, it should be noted that the probability of Type I error is increased due to the large number of tests that were run. The goals of this study were to use fMRI to: identify brain regions where neural activation associated with three separate stages of risky decision-making differed between heavy drinking adolescents and controls; examine whether neural activity associated with risky decision-making changed across a five-week period of abstinence and whether the trajectory of change over time differed for heavy drinkers vs. controls; and determine whether neural activation in regions showing baseline group differences during risky decision-making could predict differences in neuropsychological functioning , risk-taking performance, or self-report of risk taking behavior and impulsivity. Both ROI and whole-brain analyses were conducted to examine neural functioning in brain regions expected to activate during risky decision making as well as brain regions not implicated as fundamental, but which may still be relevant. With regard to group differences in neural activation patterns in ROIs at baseline ,grow systems for weed heavy drinkers showed less BOLD response relative to controls in the right insula during the anticipation phase of decision-making when previous balloon trials had popped.

Although it was originally hypothesized that heavy drinking adolescents would show increased activation in the insula during this stage of decision-making, the finding of reduced activation fits with previous investigations that suggest the insula is primarily involved in the experience of loss avoidance . With this interpretation, it may be that heavy drinkers were less concerned with avoiding further losses than controls, even when faced with a reminder of loss from the previous trial. During the experience of negative outcome evaluation , heavy drinkers showed increased activation in bilateral regions of the VMPFC compared to nondrinkers. This result is consistent with expectation given the widely replicated findings outlining the VMPFC as central to reward processing . Specifically, studies have suggested that the VMPFC is involved in encoding the reward value of a choice , with greater activation associated with appraisals of a higher valued item or result . While it was expected that heavy drinkers would show increased activation in the VMPFC relative to controls during evaluation of both “win” and “loss” outcomes, this effect was only observed during the evaluation of “loss” outcomes. Thus, it is possible that heavy drinkers and controls experience “winning” similarly; however, heavy drinkers may be less affected by negative outcomes, finding them more rewarding. This explanation is consistent with findings from Bogg and colleagues’ study, in which participants with greater recent alcohol consumption showed greater medial prefrontal cortex activity while experiencing outcomes achieved through riskier behavior . With regard to changes in BOLD activation during the stages of risky decision making across the five-week period of abstinence , findings were mixed. In brain regions where differences in BOLD activation were observed at baseline , no between-group differences were evident after two to three weeks of abstinence. This pattern of change was anticipated, though group differences were not expected to be non-significant until the third time point . However, the observed pattern is consistent with results of the preliminary study by Pulido and colleagues , in which baseline BOLD activation differences between heavy drinkers and controls to an alcohol cue reactivity task were non-existent in most brain regions after two to three weeks of abstinence. On the other hand, some surprising group differences in the trajectory of BOLD response to risky decision-making across the five-week abstinence period were seen.

Specifically, controls showed increasing activation over time in the right anterior cingulate during the pre-response assessment phase of decision-making, while heavy drinkers did not show changes over time in this region during this phase of decision-making. Instead, heavy drinkers showed increasing activation over time in the right ventromedial prefrontal cortex/anterior cingulate during the anticipation phase of decision-making, while controls did not show this change. The anterior cingulate has been linked with a variety of separate functions at different stages in the decision-making process. Specifically, during the pre-response assessment phase of decision-making, it has been shown to have a primary role in cognitive control processes such as error and performance monitoring . During anticipation of reward and reward evaluation, the anterior cingulate has also shown high reactivity, particularly when there is a higher degree of effort that must be undertaken to “earn” a reward . Other studies have suggested that, like the insula, the anterior cingulate is associated with loss aversion during the anticipation of risks, with greater activation observed as the probability of a negative outcome increases . Thus, increasing BOLD response in the anterior cingulate over time during the pre-response assessment phase may be indicative of learning, or an increased attention to cognitive strategies for optimal performance on the task as participants become increasingly comfortable with the task format. As controls showed this pattern but heavy drinkers did not, it is possible that heavy drinkers did not increase attention to cognitive strategies/performance monitoring over repeated assessments to the same degree as controls. Similarly, increased BOLD response in the VMPFC/anterior cingulate during the anticipation phase of decision-making may be indicative of increasing loss aversion or greater attention to the probability of loss with increasing abstinence. As the heavy drinkers showed this pattern but controls did not, this suggests that alcohol may alter brain functioning in regions responsible for aversion to loss which in turn, could contribute to greater risk taking; in addition, short-term abstinence may contribute to neural recovery in these regions. Other regions of the brain demonstrated main effects of group on BOLD response averaged across time points, indicating that some neural functioning differences between heavy drinkers and controls persisted across the abstinence period. Specifically, heavy drinkers had less BOLD response relative to controls in bilateral regions of the DLPFC during the pre-response assessment phase of the task , in the left insula during anticipation , and in the right VMPFC during anticipation .

The finding that heavy drinkers exhibited hyporeactivity relative to controls in the DLPFC during pre-response assessment is consistent with expectation as well as with previous studies indicating that adolescents engage the DLPFC to a lesser extent than do adults during risky decision-making . In addition,cannabis indoor grow system hypore activity in widespread brain regions during a risky decision-making task has been observed in a sample of adolescent males with substance use problems who had been abstinent from substances for at least 30 days. It is important to note that all adolescents in this sample had comorbid conduct disorder, so it is impossible to determine whether the observed neural abnormalities resulted from substance use, psychiatric factors, or both. The finding that heavy drinkers displayed decreased left insular activity averaged across time points during anticipation is consistent with between-group differences in baseline BOLD response observed in this study, where heavy drinkers showed decreased right insular activity during anticipation, when previous balloons had popped. However, it is interesting that differences in right insular activity did not persist across the abstinence period as is observed in the left insula. This may be because the left insular activity was observed after previous “winning” trials while the right insular activity was observed after previous “loss” trials . If insular activity represents a marker for loss aversion, it may be that loss aversion is triggered more quickly when the memory of loss is more salient. In other words, heavy drinkers may continue to be less averse to loss during anticipation of an uncertain outcome when the memory of a recent reward is more salient. The finding that heavy drinkers showed increased left VMPFC activity averaged across time points during anticipation is consistent with the idea that adolescent heavy alcohol users may pay greater attention to the potential rewarding properties of an uncertain outcome compared to nondrinkers. This interpretation also fits with the possibility that alcohol use may alter neural functioning patterns related to reward sensitivity, and that these alterations may persist for longer periods than do alterations related to loss aversion. However, the finding that controls had greater BOLD response than heavy drinkers averaged across time in the right VMPFC during anticipation is somewhat surprising as we might expect the opposite pattern. With regard to baseline BOLD response as a predictor of executive functioning, risk-taking task performance, and self-report of risk-taking and impulsive behavior , baseline BOLD response during risky decision-making was not associated with executive functioning performance or performance on the risk-taking task.

However, greater activation in the VMPFC during negative outcome evaluation was predictive of self-report of greater risk-taking in the naturalistic environment. This is consistent with other studies that have reported similar relationships between BOLD response and ratings of risk-taking in adolescents. For example, Galvan and colleagues found that BOLD response in the nucleus accumbens during a reward processing task was positively correlated with adolescents’ ratings of both the likelihood of engaging in risky behaviors in the near future and ratings of anticipated positive consequences as a result of such risky behavior. Nucleus accumbens activity was negatively correlated with ratings of anticipated negative consequences of risky behavior. Findings from the exploratory whole-brain analysis suggest that heavy drinkers and controls may show differential neural response to risky decision-making in regions not identified as ROIs. During the baseline anticipation phase, heavy drinkers showed less BOLD response than controls in the right middle frontal gyrus and left anterior cingulate . During the baseline outcome evaluation phase, heavy drinkers showed greater BOLD response than controls in bilateral middle temporal and cingulate gyrus, right supramarginal gyrus, inferior/medial/superior frontal, and precentral gyrus areas. These baseline group differences were not evident after two to three weeks of abstinence, except for the right middle frontal gyrus during anticipation, where BOLD response was greater for controls, averaged across time points. During the pre-response assessment phase of decision-making, controls showed greater BOLD response than heavy drinkers averaged across time points, in widespread regions including bilateral medial temporal, inferior parietal, inferior/middle/medial frontal, anterior cingulate, right thalamus, and right superior frontal areas. In contrast, during the anticipation phase, heavy drinkers showed more BOLD response relative to controls averaged across time points in bilateral cingulate cortices, middle/inferior frontal regions, and occipital areas. During outcome evaluation, heavy drinkers showed greater BOLD response relative to controls averaged across time points in bilateral inferior/middle/medial frontal, right inferior parietal, right superior temporal, left occipital, and left thalamic regions. Overall, results of the whole brain analysis provide evidence that there are other regions implicated in the stages of risky decision-making, which show neural functioning differences between heavy drinkers and controls. Some neural functioning differences in these areas appear to resolve after two to three weeks of abstinence, while others persist for at least 4+ weeks. Some limitations should be mentioned. First, among the heavy drinking adolescents, there was considerable variability in the reported length of abstinence from alcohol prior to entering the study. Although the ideal length of abstinence at study entrance was between 4-10 days, some subjects reported much longer periods of abstinence , while others reported shorter periods . Although the mean number of days of reported abstinence in the heavy drinking group was consistent with the ideal length of abstinence it is possible that the variability may have skewed the results.