The cross-sectional nature of the majority of studies examining adolescent neurocognitive functioning makes it difficult to determine the influence of alcohol and drug use on adolescent neurocognition. Therefore, ongoing longitudinal neuroimaging studies are essential to ascertain the degree to which substance intake is linked temporally to adverse changes on indices of brain integrity, or whether neural abnormalities reflect pre-existing patterns. In cross-sectional or longitudinal work, several methodological features are critical to evaluate the potential influence of adolescent substance use on neurocognition. These issues pertain to ensuring participant compliance, accurately assessing potential confounds, and maximizing participant follow-up.
Adolescent compliance as a research participant can be maximized by attending to rapport, building trust, and ensuring privacy of self-report data to the extent that is ethical and feasible to the setting. For behavioral tasks within or outside of imaging, it is critical to ensure participants comprehend task instructions, are fully trained on fMRI tasks, and then are given reminders just prior to task administration. Motion during scan acquisition is detrimental to the quality of imaging data, and is often worse in younger adolescents than older teens or adults. Adolescent head motion can be minimized by the following steps: discuss the importance and rationale for keeping the head still multiple times before and at the scan appointment; model and practice how to say “yes” and “no” when communicating with the research subject from the scanner; model and practice techniques for relaxing and ensuring subjects are in a position suitable for long-term comfort (e.g., legs are not crossed) before scanning begins; maximize participant comfort by using soft cushions around the head and under the knees; and many studies, especially those with younger participants, find practicing scanning in a less expensive mock scanner results in improved participant comfort and more reliable data during data acquisition.
Accurately measuring and accounting for confounds frequently present in adolescent substance-using populations is essential for elucidating the true effect of substance use on adolescent neurocognitive functioning. Common confounds in this population include head injury, depression, ADHD, conduct disorder, prenatal exposure to neurotoxins, family history-related effects, and polysubstance involvement. Conversely, excluding subjects for the aforementioned confounds may impede the generalizability of results. The tradeoff between minimizing confounds and having meaningful, ecologically valid results is an important study design decision, especially given the high cost of fMRI sessions.
Accurately measuring abstinence is another important consideration in substance-related research protocols. If abstinence is required for participation (and compensation) in a study, the dynamics of self-report could change. While biological data may help confirm self-report, these measures are imperfect and do not pinpoint the quantity of specific timing of substance intake 65, 66
. Regarding abstinence from cannabis, obtaining serial quantitative THC metabolite levels, normalized to creatinine, is the best approach for guarding against new use episodes 67
Tracking participants over time is a critical part of many clinical issues when interested in the degree to which a variable (e.g., alcohol or marijuana use) might result in neural changes. Although some statistical approaches can help manage attrition, effective tracking procedures are more desirable to ensure study integrity. To maximize participant follow-up, frequent contact with participants must be maintained 68
. Having a well-trained, friendly staff experienced with the population also helps retain participants and parents, and ensures that all participants fully understand the tasks and expectations during the study. Collecting comprehensive contact information can help track adolescents over time in case they should relocate. Additionally, follow-up measures and procedures should be as similar as possible to baseline, except to mitigate learning and practice effects 69
. For imaging studies, field map unwarping of EPIs (e.g., fMRI and DTI) should also be considered, as this technique appears to produce more consistent localization of activations 70
. Finally, as technical problems are common, back up plans for each piece of equipment used in the neuroimaging session should be in place.