Exclusionary criteria at each follow-up time-point consisted of endorsement of an emergent Axis I disorder as measured by a structured diagnostic interview . Participants were asked to refrain from alcohol and substance use for at least 24 hours prior to all baseline or follow-up assessments, verified via breath alcohol concentration and urine drug screen. The University of California San Diego Human Research Protections Program approved the study protocol and procedures . Data for the current project was selected among the first 15 years of annual follow-up assessments. Participants were included in the present analysis if they: transitioned to frequent binge drinking, averaging ≥ one binge episode per week for at least a one-year period, at any point during the 15-year follow-up period, and provided usable neuroimaging data within 3 years of their 18th birthday. fMRI Go/No-go data was selected from available scan time points to be closest to their 18th birthday and prior to their transition to frequent binge drinking . This time point was selected because all participants had begun moderate drinking by this time and it represented a proximal time point to the average age of transition , thus reducing the potential for influence from extraneous developmental factors. Time to transition to frequent binge drinking was calculated as the difference in time between age of transition onset and age at scan. The neuroimaging data of 8 participants in this sample are also included in the report by Wetherill et al., 2013. Youths were administered comprehensive interviews at baseline, including the assessment of demographics, living situations, and alcohol and drug use. The Hollingshead Index of Social Position score , an index of socioeconomic status , was calculated for each subject using parental socioeconomic background information to characterize the youth’s rearing environment. Higher values indicate lower SES .
Corroborative information from an informant was used to support youth report on demographic background and family history topics. The annual follow-up assessments were similarly structured. The Customary Drinking and Drug Use Record structured interview was used to assess history of alcohol consumption and alcohol-related problems, as well as additional substance use information. For the purposes of this study, a binge drinking episode was defined as the consumption of ≥5 alcoholic drinks for males, or ≥4 drinks for females, in a single occasion. Consistent with previous analysis on the parent data set ,vertical farming units imaging data were collected using a 3.0 Tesla General Electric short bore Excite-2 system with an eight channel phase-array head coil. A high-resolution T1-weighted sequence including a sagittally acquired spoiled gradient recalled sequence was acquired. BOLD signal was measured with T2*-weighted axially acquired echo-planar images . Field maps with two different echo times were used to measure signal dropout and field inhomogeneities. Stimuli for the task were back-projected from a laptop to a screen at the foot of the scanner bed and were visible via an angled mirror attached to the head coil. Task performance and behavior was recorded using a fiber-optic response box compatible with MRI . Response inhibition was assessed during scanning via an event-related Go/No-Go paradigm . The task consisted of a serial presentation of blue shapes, which included 64 large circles, 16 small circles, 43 large squares, and 57 small squares. The duration of each stimulus was 200ms and the intertrial interval was 1,500ms. Participants were asked to press a button each time a large circle, small circle, or large square shape was presented but to withhold their response when a small square was presented . Baseline constituted ~114 seconds scattered throughout the task. Primary analyses contrasted BOLD response during no-go correct rejection trials relative to no-go false alarm trials . The no-go correct rejection versus go contrast was also evaluated. Correct rejections were determined by the absence of a motor response during no-go trials.
False alarms were defined as a button press following a no-go stimulus. Processing of imaging data was conducted using the bug-corrected version of the Analysis of Functional NeuroImages software . Abnormal signals and artifacts were removed from the data, and the time series data were aligned temporally and coregistered to a maximally stable base volume using an iterated least squares algorithm . AFNI’s 3dSkullStrip was used to skull strip each participant’s high-resolution T1-weighted image. Participants’ anatomical and functional data sets were co-registered and warped to Talairach space . Functional data were resampled to 3mm2 voxels, and activations maps were spatially smoothed using a 5mm full-width half maximum Gaussian filter. Motion was estimated for each participant and used as a control in task analyses . 3dDespike was used to detect outliers in the motion parameters. Significant outliers in the time-series data were censored or despiked. Analysis of time series data utilized multiple regression controlling for linear drift, baseline signal, and motion from 6 motion parameters calculated above. Regressors of interest and no interest convolved with a modified gamma variate function that modeled anticipated hemodynamic response. Beta weights were converted to percent signal change which were used for further analysis.Performance measures evaluated in the Go/No-Go task included percent correct on inhibitory trials, β , and d′ . Similar to methods previously used by the authors , activation was masked by an average skull-stripped anatomical image from all participants. Whole brain, voxel-level analysis on the masked data was conducted using a paired t-test to contrast no-go correct rejection vs. no-go false alarm trials. To control for variability in age at scan acquisition, time to transition to frequent binge drinking was chosen for the correlation analysis and entered as the covariate of interest in the model. The Clustsim nonparametric randomization/permutation option of 3dttest++ was used with a conservative voxel-wise alpha of 0.001 and cluster-wise alpha of 0.05, resulting in an estimated a cluster size threshold of 18 contiguous voxels. This method of Type I error control has been shown to produce false positive rates compatible with the nominal 95% confidence interval .
A second model not containing the covariate of interest was run to validate the task by showing task-relevant activation in this sample using the same Type 1 error correction as above. As expected, successful inhibition in this sample of frequent binge drinkers was associated with activation in a number of regions previously implicated in the literature, including the fronto-striatal system , validating the use of this paradigm in our high risk sample. In the correlation analysis, a single cluster of activation, including portions of the left insula, inferior frontal gyrus , and precentral gyrus, elicited during successful inhibitory control, was found to predict time to transition to high-risk frequent binge drinking in adolescents who were already engaged in moderate alcohol use. Specifically, greater BOLD response in this cluster predicted longer time to transition, implying that lower magnitude of activation during successful inhibition could serve as a temporal warning of future high-risk impulsive behavior. The IFG, precentral gyrus,weed drying room and insula have been consistently implicated as critical regions involved in response inhibition . Although the right IFG/insula are most commonly implicated , a number of studies have also implicated key roles for the left IFG/insula in this process . This study provides further support for the involvement of the left IFG/insula in inhibitory control by demonstrating the predictive utility of activation in these top-down executive control regions for the onset of impulsive binge drinking behavior. The correlation with time to transition to frequent binge drinking also supports the notion that this pattern of alcohol consumption is likely driven, at least in part, by deficiencies in inhibitory control and suggests opportunities for intervention prior to the onset of this very high-risk behavior. Inhibitory control interventions, particularly those utilizing Go/No-go paradigms, have demonstrated effectiveness for short-term health behavior change , which may be all that is needed to delay onset of this high-risk drinking pattern beyond the critical neurodevelopmental stage of adolescence. The results of this study should be interpreted within the context of its strengths and limitations. The prospective correlational design is a strength of the study, as it avoids issues related to the selection of comparable controls and addresses the question of whether the magnitude of the BOLD signal during inhibitory control contains clinically relevant predictive information. Another strength is the well-characterized sample of frequent binge drinkers who were already engaged in moderate alcohol use at scan acquisition. Few attempts have been made to identify unique risk factors for adolescents that are already engaged in moderate alcohol use, despite the exceptionally high prevalence of adolescent alcohol users. Limitations of the study include the relatively small sample size for fMRI studies and small number of no-go false alarm trials included in the analysis. The use of conservative statistical thresholding and the consistence of implicated regions with the extent literature on inhibitory control provides support for the validity of the results; however, additional studies using more difficult tasks within larger samples are needed to confirm these within-subjects effects. Furthermore, the sample is comprised predominately of White adolescents with high educational attainment, potentially limiting the generalizability of the results. Thus, replication within a more diverse sample of adolescents is warranted. In conclusion, this study suggests that BOLD response in portions of the IFG, insula, and precentral gyrus during successful inhibitory control could prove valuable as a temporally specific risk marker for future frequent binge drinking behavior.
Early identification of adolescents at-risk for this pattern of alcohol use is of great importance given the potential for neural consequences associated with alcohol use during neurodevelopment . The increased study of risk factors for youth already engaged in moderate alcohol use could provide additional insights into meaningful pathways for intervention that were previously overlooked.Alcohol use disorder is a highly prevalent, chronic relapsing disorder with a high disease burden in the United States. Despite current and lifetime prevalence rates of 13.9% and 29.1%, respectively, it remains largely untreated as only 7.7% of those with 12-month and 19.8% of those with lifetime diagnoses sought treatment in 2012– 2013. In spite of low treatment rates, pharmacotherapy offers a promising treatment method for AUD. The Federal Drug Administration has approved of four medications for AUD: disulfiram , oral naltrexone , extended-release injectable naltrexone , and acamprosate . However, these currently approved pharmacotherapies are only modestly effective, so there is still a great need to develop more effective interventions. Medications development is a very costly, cumbersome, and inefficient process that can take nearly 20 years from discovery to market. In particular, the development of treatments for alcoholism has been difficult with over 20 medications having been tested in humans yet only three were able to receive FDA approval, the last of which was granted over a decade ago. Therefore, there is a pressing need to develop valid and efficient methods to decrease the cost and length of medications development to better shepherd novel compounds from the lab to dissemination. The development of novel medications for AUD is a high priority research area, but the drug development process is long and challenging, with many compounds stuck in the transition from preclinical to clinical testing, also known as the “valley of death”. Beyond the “valley of death,” there is an overall need to develop effective methodologies for efficiently running clinical trials, particularly in screening novel compounds in early phase 2 trials. Early phase 2 trials, also known as “proof-of concept” studies, help determine if a novel medication is safe, tolerable, and efficacious using clinically relevant phenotypes such as cue-induced craving or subjective response to alcohol. These trials largely incorporate human laboratory paradigms to assess medication efficacy, providing valuable information on whether or not the medication warrants a larger clinical trial. However, human laboratory paradigms have not always demonstrated translational validity and often lack the ecological validity of clinical trials where medication efficacy is established through clinically meaningful endpoints. Therefore, there are major opportunities to refine this process of screening novel medications by combining the internal validity of human laboratory models and the external validity of clinical trials. To that end, the current study aims to develop and validate a novel early efficacy paradigm to screen medications for AUD. This early efficacy paradigm is the practice quit attempt model adapted from the smoking cessation medication development literature. In the original practice quit attempt model, individuals who report intrinsic motivation to quit smoking undergo a 7-day practice quit attempt while taking study medication. Individuals with high intrinsic motivation to quit smoking fared better on active medication, compared to placebo, on increased abstinence, while individuals with low intrinsic motivation showed no effect of active medication.