We obtained written informed consent and assent from interested teens and their guardians, approved by the University of California San Diego Human Research Protections Program. Adolescents were administered a 90-min telephone screening interview to ascertain family history of substance use and psychiatric diagnoses using the Family History Assessment Module screener , lifetime substance use and abuse/dependence criteria using the Customary Drinking and Drug Use Record , and history of psychiatric disorders using the Diagnostic Interview Schedule for Children . Collateral interviews were administered to a guardian, usually a parent. Exclusion criteria included history of head injury with loss of consciousness >2 min, neurological or medical problems, learning disabilities, DSM-IV psychiatric disorder other than conduct disorder, current psychotropic medication use, significant maternal drinking or drug use during pregnancy, family history of bipolar I or psychotic disorder, and left handedness. Teens meeting criteria for conduct disorder were not excluded due to high comorbidity with substance use disorders . Eligible participants were ages 15−17, and groups were demographically similar . Controls had little experience with alcohol or other drugs. AUD adolescents met DSM-IV criteria for current alcohol abuse or dependence, but had limited experience with marijuana . Only two AUD teens disclosed marijuana use in the month before scanning . MAUD adolescents met DSM-IV criteria for both current marijuana and alcohol abuse or dependence, had ≥100 lifetime experiences with marijuana and had used ≥10 days/month in the three months before scanning. One MAUD participant reported stopping marijuana use 4 months prior to scanning; however,rolling grow table the urine toxicology screen indicated recent use. Twelve other MAUD teens reported marijuana use in the week before the scan, with last use 3.3 ± 1.7 days prior to scanning.
Participants in each group had little experience with drugs other than alcohol and marijuana , had not used other drugs for 30 days prior to imaging, and had not used marijuana or alcohol for at least 48 h before scanning. Importantly, AUD and MAUD youths demonstrated similar alcohol use disorder characteristics . Both AUD and MAUD teens were primarily weekend heavy drinkers, as evidenced by an overall average 15.13 days since last drink and typical blood alcohol concentration reaching 0.107. Two AUD teens and one MAUD teen reported abstinence from alcohol in the month before scanning. AUD and MAUD teens displayed similar cigarette smoking patterns, but more MAUD teens had experiences with other drugs than AUD and control teens, although such use was limited . Although MAUD and AUD teens had higher rates of conduct disorder than control teens, severity was mild to moderate reflected by the normal range Child Behavior Checklist externalizing scores . Substance involvement and abuse/ dependence diagnoses were assessed using the CDDR . The CDDR collects lifetime and past 3-month information on alcohol, nicotine, and other drug use, and assesses DSM-IV abuse and dependence criteria, withdrawal symptomatology, and other negative consequences associated with substance use. The CDDR also obtains information necessary to estimate typical blood alcohol concentrations reached using the Widmark method, i.e. amount consumed, duration of drinking, height, weight, and gender . Strong internal consistency, test–retest, and inter-rater reliability have been demonstrated with adolescent CDDR assessments . The Timeline Follow back obtained detailed substance use patterns for the 30 days prior to scanning. On the day of the scanning session, all participants submitted samples for Breathalyzer and urine drug toxicology analyses. Participants were asked to abstain from substance use for at least 48 h before imaging to avoid intoxication and acute withdrawal during scanning.
Imaging sessions were held Thursday evenings between 8 and 10 p.m. to maximize recovery from weekend binge drinking and maintain consistent circadian influence across subjects. According to self-report on the Timeline Follow back , the most recent alcohol use was 72 h and marijuana use was 48 h before scanning. Upon arrival for the imaging session, all participants submitted samples for Breathalyzer and urine drug toxicology for THC, ethanol, amphetamines, methamphetamines, barbiturates, benzodiazepines, cocaine, codeine, morphine, and PCP. No participant had a positive breath alcohol concentration. Due to experimenter error, toxicology screens were unavailable for one control teen, one AUD teen, and five MAUD teens. Based on available data, only MAUD participants produced toxicology screens positive for cannabinoids, and no toxicology screens were positive for any drug other than cannabinoids. Although it is possible that MAUD teens were over-reporting marijuana use, self-reported marijuana use has been an accurate predictor of verified use .Imaging data were processed and analyzed using the Analysis of Functional NeuroImages package . We first applied a motion-correction algorithm to the time series data . Second, we correlated the time series data with a set of reference vectors that represented the block design of the task and accounted for delays in hemodynamic response , while covarying for estimated motion and linear trends. Next, we transformed imaging data to standard coordinates then resampled the functional data into 3.5 mm3 voxels. Finally, we applied a spatial smoothing Gaussian filter to account for anatomic variability. After processing functional data, we examined average BOLD response to the SWM task in each group using one sample t-tests, and determined regions that showed greater response to SWM relative to simple attention , reduced response during SWM relative to rest , and greater simple attention response than SWM response.
We next compared response during SWM relative to simple attention between groups with ANOVAs, and performed pairwise comparisons between groups. We performed group comparisons on the whole brain, rather than discrete regions thought to be activated by the task, because previous studies by our group and others have suggested neural reorganization and use of alternate brain systems during working memory among individuals with AUD. To control for Type I error in group analyses, we required significant voxels to form clusters ≥1072 μl , yielding a cluster-wise α < .0167 . We utilized the Talairach Daemon and AFNI to confirm gyral labels for clusters. Previous research has suggested that neuropsychological deficits among adult marijuana users are associated with lingering effects of recent use, and that these impairments dissipate with extended abstinence . To understand whether group differences in the current study relate to recent marijuana use, we performed post-hoc regressions within the MAUD group. First, we extracted the average fit coefficient for each MAUD participant from each cluster where we observed a difference between MAUD and control or AUD teens. Next, we used regression analyses to examine whether days since last marijuana use predicted brain response within each group difference cluster. Groups did not significantly differ on any neuropsychological performance measure . SWM accuracy was 86 ± 9% in the control group, 91 ± 5% in the AUD group, and 92 ± 5% in the MAUD group, revealing a trend for MAUD to be more accurate than controls . However, one control performed at 60% accuracy,cannabis grow equipment which was >2.5 standard deviations below the mean for that group, and exclusion of this participant removed the group difference in SWM accuracy. This raised the concern that this individual impacted the fMRI group analyses. Upon further examination, we determined that this participant’s brain response was within the normal range for each significant cluster described below. Groups did not differ on simple attention accuracy or reaction time to either condition. The overall pattern of BOLD response to the SWM condition relative to simple attention was similar in all three groups. Participants showed SWM activation in several regions, including bilateral prefrontal, premotor,cingulate, and posterior parietal areas . Groups showed SWM deactivation in medial prefrontal cortex, a large posterior midline region including posterior cingulate and cuneus, and several temporal regions . Although groups demonstrated similar patterns of response localization, several significant group differences emerged. The response differences between AUD and control teens are detailed elsewhere . Briefly, AUD teens showed less SWM response than controls in the left precentral gyrus and midline precuneus/posterior cingulate, but more SWM activation than controls in bilateral posterior parietal cortex . MAUD participants evidenced altered BOLD response compared to controls in several regions: bilateral inferior frontal gyri, right superior temporal/supramarginal gyri, right middle and superior frontal gyri , and anterior cingulate . In both right inferior frontal and superior temporal regions, MAUD teens demonstrated less SWM response than controls. Moreover, while controls showed SWM activation in the right superior temporal gyrus, MAUD teens showed greater simple attention response than SWM response. In right dorsolateral prefrontal cortex, MAUD youths showed more SWM activation than controls. Both controls and MAUD evidenced SWM deactivation in the anterior cingulate; however, MAUD showed a greater intensity of deactivation than controls. MAUD also demonstrated deactivation in the left inferior frontal gyrus, where controls showed no significant activation or deactivation.
MAUD teens showed different response intensity relative to AUD teens in the right inferior frontal gyrus/insula, left precuneus, right middle temporal/supramarginal gyri, left superior temporal gyrus, and a large cluster spanning anterior cingulate and bilateral inferior frontal gyri . In the precuneus, groups showed SWM activation, yet AUD teens showed greater response than MAUD teens. Similar to controls, AUD teens showed SWM activation in right inferior frontal and middle temporal areas, while MAUD teens evidenced greater simple attention response than SWM response. In the left superior temporal gyrus, AUD showed SWM deactivation, while MAUD demonstrated no significant activation or deactivation. Finally, a group difference was observed in a large cluster spanning anterior cingulate and bilateral inferior frontal gyri. In this cluster, both AUD and MAUD showed deactivation, but MAUD showed greater intensity and spatial extent of deactivation. Days since last marijuana use did not significantly predict brain response among MAUD teens in any cluster where MAUD teens had significantly different SWM response than controls or AUD teens. A trend was found for more recent use to be associated with reduced brain response in the right middle temporal gyrus , where MAUD teens showed less SWM response than AUD teens. This study investigated the neural correlates of SWM in adolescents with comorbid marijuana and alcohol use disorders, teens with alcohol use disorders alone, and demographically similar non-abusing adolescents. The groups showed similar neuropsychological abilities, SWM task performance, and general BOLD response localization patterns. However, MAUD teens demonstrated significantly more dorsolateral prefrontal SWM activation and anterior cingulate deactivation, and significantly less right inferior frontal and superior temporal response compared to control teens. Similarly, MAUD youths also showed significantly more medial frontal deactivation as well as less right inferior frontal and bilateral temporal activation compared to AUD teens. As noted above, MAUD teens showed more SWM activation than control teens in the right dorsolateral prefrontal cortex, a brain region consistently active during working memory . A recent fMRI study of heavy cannabis using adults also demonstrated greater dorsolateral prefrontal recruitment relative to controls during SWM 6 to 36 h after last marijuana use, despite similar task performance . More intense and widespread fMRI response despite intact behavioral performance has also been observed among adult alcoholics, suggesting that while some task-related areas demonstrate deficient processing, other ancillary regions may become active to compensate, resulting in an altered functional network among alcoholics . Similarly, the MAUD teens in this study may compensate for subtle neuronal disruption with increased task-related neural recruitment in frontal regions, observed in fMRI as heightened activation. However, MAUD teens did not show the aberrant parietal response we expected given the role of parietal cortex in SWM tasks . While greater SWM task difficulty is associated with increased activity in both frontal and parietal cortices , increased dorsolateral prefrontal activation may be associated with general task difficulty, whereas greater parietal response relates to visuospatial demands . Therefore, the increase in response among MAUD teens in frontal regions, but not parietal cortex, may suggest a greater difficulty with general task demands, despite similar task performance. However, given a more difficult task, frontal regions may no longer be able to compensate, and activation may decrease in parallel with decreasing task performance. Although all three groups demonstrated SWM deactivation in the anterior cingulate, MAUD teens showed significantly more deactivation than controls and AUD-only adolescents. The anterior cingulate is highly active at “rest” , during which it is thought to monitor various environmental and internal processes .