The rapid event-related design of the BART fMRI task created some special challenges for analysis. Because the task moved very quickly through the Think, Pump, Wait, Inflate, Pop or Win, and Rest conditions , it is possible that there was overlap of the hemodynamic response across task conditions. However, the deconvolution process is generally effective at disentangling the individual hemodynamic response functions so that the BOLD response pattern specific to each task condition can be appropriately measured. Rapid event-related designs are also limited by a lower signal-to-noise ratio, compared to blocked designs or slower event-related designs. This ultimately leads to a loss of statistical power. Also, some of the task conditions may not have been perceived as separate, distinct events to participants, and thus BOLD response may not be qualitatively different between these conditions. For example, the “Wait” control condition occurred immediately after subjects inputted their number of pumps. Although the active “Inflate” condition was designed to measure the anticipation stage of decision making, the “Wait” condition may have captured some anticipatory response as well. The visual display of a balloon inflating likely added an element of anticipation above and beyond the “Wait” phase ; however, both “Rest” and “Wait” conditions were used as control conditions for the “Inflate” conditions, in order to address this issue. It is somewhat surprising that heavy drinkers and controls showed very few differences on BART task performance variables. This contrasts with previous studies using BART paradigms which have found group differences between adolescent smokers and non-smokers and adolescents with “serious substance use and conduct problems” and controls . However other studies have not shown differences between adolescent substance users and non-users . One reason for the lack of performance differences may be that the version of the BART used in this study differs from the original design of the task where participants are asked to sequentially input discrete pumps, one by one, to pump up the balloons . It may be that inputting each pump separately allows for anticipation to build to higher levels,indoor cannabis grow system leading to an increased response to reward or loss which may further risk-taking performance on the task.
Some studies have also analyzed the first and second half of the BART separately, to examine group differences in risky decision-making as the task progresses. A preliminary study of the BART in the current sample used this method and found that heavy drinking adolescents had a greater number of pumps on the second half of the task compared to controls, though after 5 weeks of abstinence, the groups were equivalent . In addition, at baseline, number of pumps on the second half of the task was positively correlated with number of recent alcohol binges and number of drinks per day. Although the lack of group differences in BART task performance within the current study allows for easier interpretation of the fMRI data, it is also important to consider the real-world generalizability of any laboratory measure of an abstract concept like risky decision-making. Another limitation to this study and all fMRI studies is that other variables may have affected the magnitude of the BOLD response other than the construct of interest . For example, heavy drinkers and controls differed in their use of cannabis and nicotine. However, findings remained unchanged after controlling for cannabis use. While tobacco use was not controlled for in this study, the level of use in the heavy drinking group was relatively low, and no participants met criteria for tobacco dependence. Therefore it is unlikely that tobacco use could have accounted for variability in BOLD response. Heavy drinkers and controls also differed in terms of verbal intellectual functioning and level of externalizing problems. However, level of externalizing problems was not controlled for, primarily because the difference observed on this variable was thought to represent a naturally occurring difference between adolescents who use substances and those who do not. Cerebral blood flow was not measured in this study and is known to have some effect on BOLD response. As resting state perfusion can affect the magnitude of the BOLD response , it is possible that differences in cerebral blood flow could explain BOLD response differences between heavy drinkers and controls.
A recent study by Jacobus and colleagues found that for heavy adolescent marijuana users, cerebral blood flow was reduced in four cortical regions and increased in one region at baseline ; however, after four weeks of abstinence, no between-group differences in cerebral blood flow were found. In extrapolating these findings to adolescent heavy alcohol users, it seems possible that cerebral blood flow could have an impact on the baseline between-group differences in BOLD response in this study, but may be less likely to impact BOLD response differences at the +2weeks and +4weeks time points. The issue of test-retest reliability of the BOLD response should be considered. There has been some controversy about this topic in the literature, as some studies have found high test-retest BOLD signal reliability , and others have reported considerable within-subject variation in BOLD signal change across scan sessions . Changes in BOLD signal response across different scan sessions for the same subject can occur for several reasons, such as fluctuating mood/anxiety/alertness states , levels of effort , motion, scanner drift , physiological changes , and developmental maturation. Many of these variables were measured and controlled for in this study. Specifically, acute anxiety and alertness were found to be equivalent between groups and are therefore unlikely to affect results. In addition, subpar effort was controlled for by removing subjects with five or more “Too Slow” outcomes within a given scan session, as it was assumed that subjects with a high degree of “Too Slow” responses were not adequately engaged in the task. Excess motion was controlled for by removing participants with more than 20% repetitions containing excessive head motion. Field maps applied to the fMRI acquisitions also helped to control the BOLD signal stability across scans, as this process minimizes warping and signal dropout, as well as reduces mislocalization errors, especially in frontal regions. Physiological changes and scanner drift were more difficult to control for and thus there is a small probability that these variables could have affected the reliability of the BOLD signal across repeated assessments in this study. Finally, the probability of Type I error in this study is likely higher than desired due to the large number of tests that were run within each hypothesis.
Although Hypotheses 1 and 2 included corrections for multiple comparisons within each ROI , there was no comparable correction for the number of comparisons examined across ROIs . Hypothesis 3 also did not employ any multiple comparison corrections, and these analyses should be considered exploratory. Taken together, the results of this study suggest that heavy drinkers display abnormalities in neural functioning during risky decision-making compared to their non-drinking peers, particularly in the right insula during anticipation and the ventromedial prefrontal cortex during evaluation of negative outcomes. Abnormalities in these regions appear to resolve after two to three weeks of abstinence. In addition, heavy drinkers showed some changes in BOLD response across repeated assessments , which provide further support for a neural recovery hypothesis. However, after five weeks of abstinence, heavy drinkers and controls show some differences in neural functioning that persist across time. This suggests that other regions of the brain may take longer to fully recover, or, that there are pre-existing differences in these regions,indoor weed growing accessories which could represent vulnerabilities for future substance use. In addition, this study suggests that differences in neural functioning in reward-related regions can effectively predict real-world report of risk-taking behavior. These findings are important as they suggest that neural functioning may be used as a potential biomarker for risk-taking vulnerability in the future. Future directions for this study should first include replication within a larger sample, to increase confidence in the findings. Second, an examination of the effect of length of abstinence from alcohol prior to study entry on BOLD response in the heavy drinking group will be necessary to determine whether variability in length of abstinence may have contributed to the between-group differences observed in this study. It would also be informative to analyze the first and second half of the BART task separately, as Hansen and colleagues did with the non-fMRI BART task. In addition, an examination of gender differences could be important, given results from previous studies suggesting that females may be especially susceptible to the effects of heavy alcohol use. Measures of hangover severity and withdrawal effects could also be investigated as possible moderating factors of group differences in BOLD response to risky decision-making, as these have been implicated as significantly related to performance on cognitive tasks in heavy alcohol users. The ultimate goal of this study and others like it should be to disseminate findings to youth and families, and to provide psycho education through the creation of prevention materials and public service campaigns. By understanding how the brain responds to risky decision-making after recent alcohol use, and how the brain’s circuitry may “repair” itself with sustained abstinence, adolescents could become motivated to remain abstinent from alcohol, which ultimately, would reduce the rates of accidents and deaths in this age group as a result of risky behavior. This work is being prepared for submission for publication as “fMRI Correlates of Risky Decision-Making in Adolescent Alcohol Users: The Role of Abstinence.” The dissertation author will be the primary author of this material along with co-authors Alan Simmons, Ph.D., Carmen Pulido, Ph.D., Susan Tapert, Ph.D., and Sandra Brown, Ph.D.
Two endogenous agonists of cannabinoid receptors have been well characterized and are now widely used in research: anandamide , and 2-arachidonoylglycerol . Both molecules derive chemically from the polyunsaturated fatty acid, arachidonic acid, which is used in nature as the starting material for other important signaling compounds, such as the eicosanoids. Additional endocannabinoid-related compounds present in the body include virodhamine, which may act as an endogenous antagonist of CB1 receptors, and arachidonoylserine, which may engage an as-yet-uncharacterized cannabinoid-like receptor expressed in the vasculature. As is well-known, the Cannabis plant contains more than 60 cannabinoids, which include -D9 -tetrahydrocannabinol , cannabigerol, cannabidiol, cannabinol, cannabichromene and cannabicyclol. Attention has been mostly focused on D9 -THC, because of its multiple biological properties. Nevertheless, less studied compounds such as cannabidiol may also be important, although we do not yet know at which receptors they may act to achieve their effects. D9 -THC is the only natural cannabinoid presently used in the clinic. In addition to these plant-derived cannabinoids, an extensive set of synthetic cannabinergic agonists has been developed over the last 30 years. Products of these efforts include CP-55940 , created by opening one of the rings of the tricyclic D9 -THC structure and introducing other small changes in its structure; HU-210 , a very potent cannabinoid agonist resembling some D9 -THC metabolites; and WIN55212-2 , which belongs to an altogether different class of chemicals, the aminoalkylindoles. Additionally, the metabolically stable synthetic analog of anandamideR-methanandamide is routinely used as a pharmacological probe to circumvent the short half life of the natural substance. Two important new additions to this armamentarium under discussion at the workshop include a peripherally acting cannabinoid agonist in preclinical development by Novartis for the treatment of neuropathic and inflammatory pain , andBAY-387271 , a centrally acting cannabinoid agonist in Phase II clinical studies for the treatment of stroke. The interest of the pharmaceutical industry in the application of cannabinoid agonists to the treatment of pain conditions is not recent. Indeed, most of the compounds now in experimental use derive from such an interest. Historically however cannabinoid agonist development has not proved clinically fruitful, largely because of the profound psychotropic side effects of centrally active cannabinoid agonists, hence the attention given to peripherally acting cannabinoids, which exhibit significant analgesic efficacy and low central activity in animal models. Neuroprotection is a relatively new area for cannabinoid agonists, but one that appears to be already well advanced. Preclinical studies have made a convincing case for the efficacy of cannabinoid agents not only in experimental brain ischemia, but also in models of Parkinson’s disease and other forms of degenerative brain disorders.