Understanding exactly why cannabis elevates psychosis risk has eluded researchers for decades. A neurochemical mechanism — not merely behavioral or social factors — may finally be coming into focus, with implications for how clinicians assess vulnerability in people who use cannabis and those already on the psychosis spectrum.
Using ultra-high-field 7 Tesla magnetic resonance spectroscopy, investigators measured glutamate concentrations in the anterior cingulate cortex (ACC) of 79 participants spanning three groups: typically developing controls, individuals at clinical high risk for psychosis, and patients with an established psychotic disorder. All participants completed detailed cannabis use surveys corroborated by urine toxicology and dimensional clinical assessments covering positive, negative, depressive, and manic symptom domains. The central finding was an interactive relationship: lower ACC glutamate was associated with more severe positive psychosis symptoms specifically among cannabis users, an effect absent in non-users. Independently, both reduced glutamate and heavier cannabis use each predicted worse positive and negative symptom burden. Exploratory analyses further suggested that glutamatergic deficits may amplify cannabis-related depressive and manic symptoms as well.
This work slots into a growing literature implicating the glutamate system — particularly NMDA receptor hypofunction — as a convergence point between cannabis pharmacology and psychosis neurobiology. Delta-9-THC is known to modulate glutamatergic signaling in prefrontal circuits, and the ACC is a hub for cognitive control and salience processing, functions disrupted in psychosis. The sample size of 79 is modest, limiting statistical power for interaction effects and warranting replication in larger cohorts. The cross-sectional design precludes causal inference: it remains unknown whether low glutamate predisposes individuals to cannabis use, whether cannabis depletes glutamate over time, or both. Still, the use of 7T spectroscopy — offering superior signal-to-noise resolution compared to standard 3T scanners — lends methodological credibility. This is best characterized as a mechanistically important but hypothesis-generating finding rather than a practice-changing one.