Brain network signatures of spatial memory in adolescents at risk for substance use

Examining youth before engagement in risky behaviors may help identify neurobiological signatures that prospectively predict susceptibility to initiating and escalating alcohol and other substance use. Given that frontal and medial temporal (e.g., hippocampal) regions continue developing during adolescence, identifying vulnerabilities in these systems is critical.

This study evaluated baseline brain activation during spatial memory performance using a virtual functional magnetic resonance imaging (fMRI) Morris water task (MWT). Participants were healthy, substance-naïve adolescents aged 13-14 years (n = 51, 27 females) who underwent brain imaging annually for 3 years and were evaluated quarterly for substance use. For the present report, only baseline imaging data were analyzed. During quarterly follow-ups, 22 participants initiated alcohol or other substance use, while 29 remained substance-naïve. Network activation strength was extracted for the retrieval > motor contrast in the dorsal attention network (DAN), salience network (SN), and medial temporal lobe subnetwork of the default mode network (MTL-DMN).

DAN and SN activation strengths were significantly lower in initiators than in noninitiators. No group differences emerged for MTL-DMN activation. While minimal group or sex differences were observed in task performance, greater MTL-DMN activation was associated with better performance across the full sample, whereas better performance was associated with reduced DAN activation only in noninitiators.

Adolescents who later initiated substance use showed distinct baseline network patterns in systems supporting spatial memory and attention (DAN) and task control and salience detection (SN). In contrast, hippocampal network activation (MTL-DMN) was not evident before initiation but may emerge later as a consequence of use. Because substance use status was categorized for group comparisons, early variability may not be fully captured. These findings underscore the need for future longitudinal analyses to disentangle neurobiological markers of risk that precede adolescent substance use from alterations that arise as a result of use.