Children Prenatally Exposed to Cannabis Show Thinner Cortex in Frontal and Parietal Regions

Children whose mothers used cannabis during pregnancy show measurable differences in brain structure at school age, according to a small but carefully controlled study that tracked exposure using biological samples rather than memory alone. In a prospective cohort of 56 children aged seven to eight years in the United States, those with prenatal cannabis exposure displayed reductions in a multivariate measure of cortical thickness concentrated in frontal and parietal regions of the brain, even after accounting for several confounding variables. The findings, published in Discover Neuroscience, add to a limited and inconsistent body of evidence on how cannabis exposure before birth might shape the developing brain.

The research drew on the Lifestyle and Early Achievement in Families (LEAF) study, a follow-up cohort recruited from women enrolled during pregnancy in the Ohio Perinatal Research Network Perinatal Research Repository. Rather than relying solely on maternal self-report, which previous work suggests can underestimate or misclassify exposure, the researchers confirmed cannabis exposure through stored urine samples assayed with chromatography-tandem mass spectrometry for metabolites of THC, the primary psychoactive component of cannabis. Exposure was considered confirmed if at least one specimen exceeded a defined threshold, if the mother reported use during pregnancy, or if cannabis use appeared in obstetric medical records. All personnel involved in recruitment and outcome measurement were blind to exposure status.

Of 171 children who participated in the LEAF study at age seven, 61 took part in a magnetic resonance imaging sub-study conducted between 2018 and 2020 on a 3-Tesla scanner at Nationwide Children’s Hospital in Columbus, Ohio. Five unexposed comparison participants were excluded because of excess motion or imaging artifacts, leaving 56 children in the final analyses: 22 with prenatal cannabis exposure and 34 unexposed comparisons. Structural scans were processed with FreeSurfer to derive cortical thickness and surface area for 68 cortical regions defined by the Desikan-Killiany atlas.

Because the number of participants in each group was smaller than the number of brain regions, the team used principal component analysis to condense the 68 regional measurements into summary metrics. The first principal component captured 28.37 percent of the variance in cortical thickness and 44.72 percent of the variance in surface area across the whole sample. Children with prenatal cannabis exposure showed a significantly lower mean score on the first cortical thickness component than their unexposed peers, a difference of moderate effect size. The regional loadings indicated that the difference was driven predominantly by frontal and parietal cortices, areas that include the bilateral supramarginal gyrus and inferior parietal regions overlapping with the brain’s ventral attention network. The corresponding surface area component also trended lower in the exposed group, but that difference did not reach statistical significance.

The neurocognitive results were more equivocal. On the NIH Toolbox Early Childhood Cognition Battery, children with prenatal exposure scored significantly lower than comparisons on the List Sorting Working Memory test, with a moderate effect size. However, this gap was attenuated and became non-significant when the researchers controlled for prenatal tobacco exposure, which was substantially more common among the cannabis-exposed children. Sensitivity analyses showed that the working memory difference persisted when adjusting for maternal education and maternal cognitive function, but disappeared when household income was included as a covariate. No significant group differences emerged on the other cognitive tasks, which measured inhibitory control and attention, cognitive flexibility, episodic memory, and planning.

The study also explored whether brain structure related to cognitive performance, though these analyses were explicitly hypothesis-generating. In uncorrected analyses across the combined sample, greater cortical thickness in the left precentral gyrus, right inferior parietal, and right supramarginal regions was associated with better performance on a cognitive flexibility task, and thicker left supramarginal gyrus tracked better inhibitory control, attention, and planning. Counterintuitively, greater surface area in several temporal and frontal regions was associated with worse performance on the attention measure. None of these correlations survived correction for multiple comparisons, and follow-up tests found no significant interaction between cortical morphology and exposure group, meaning the brain-behavior patterns could not be attributed specifically to cannabis exposure.

The authors are careful about interpretation. They note that the clinical and functional relevance of the cortical differences remains unknown, and that the small sample size left the study underpowered to robustly account for confounding variables such as experiences of poverty, which are themselves associated with lasting changes in brain structure. The cohort also lacked information on the dose, duration, frequency, and timing of cannabis exposure during pregnancy, factors that could substantially influence neurodevelopmental effects. Because the imaging data came from a subset of participants recruited for an MRI sub-study, the results may not be representative of the full cohort.

The findings sit within a conflicting literature. An earlier study of similarly aged children with prenatal cannabis exposure reported significantly greater, not lesser, frontal cortical thickness, and an analysis of data from the Adolescent Brain Cognitive Development Study found no significant morphometric differences in children with exposure identified by maternal self-report alone. The authors suggest that differences in software versions used for cortical reconstruction, sample characteristics, and exposure ascertainment methods may all contribute to these discrepancies. They also emphasize that maternal self-report is less reliable than biometric measures, an advantage of the current design.

The study’s environmental findings carry a cautionary message for the broader field. Household income attenuated the group difference on the working memory task, and the cortical thickness difference became non-significant when maternal education, maternal cognitive function, and household income were controlled together. These patterns echo previous research showing that associations between prenatal substance exposure and child outcomes often weaken after adjusting for socioeconomic and parental factors, underscoring how difficult it is to disentangle the effects of cannabis itself from the circumstances that frequently accompany its use during pregnancy.

Cannabis use during pregnancy appears to be increasingly common, with an estimated seven percent of pregnant individuals in the United States reporting past-month use in 2021, a trend driven partly by widespread legalization and by the perception that cannabis is a safe remedy for nausea, mood, and sleep problems. The researchers argue that their results may inform clinical guidance about potential harms of prenatal cannabis use, while stressing that larger-scale longitudinal investigations are needed to clarify whether the observed structural differences persist, worsen, or matter functionally as exposed children grow into adolescence and adulthood.

Subject of Research: Cortical morphology and neurocognitive function in children with prenatal cannabis exposure

Article Title: Anomalies in regional cortical morphology and associations with neurocognitive function in children with prenatal cannabis exposure and unexposed comparisons

Article References: Gimbel, B. A., Ontiveros, F., Mattson, W. I., Keim, S. A., Klebanoff, M. A., Boone, K. M., Yeates, K. O., Fried, P., & Nelson, E. E. (2026). Anomalies in regional cortical morphology and associations with neurocognitive function in children with prenatal cannabis exposure and unexposed comparisons. Discover Neuroscience, 21(1), Article 99. https://doi.org/10.1186/s13064-026-00327-9

Image Credits: AI Generated

DOI: 10.1186/s13064-026-00327-9

Keywords: prenatal cannabis exposure, cortical thickness, MRI, neurocognitive function, working memory, child neurodevelopment, prospective cohort, THC, executive function, principal component analysis, confounding factors, Anomalies