The research team, led by Runye Shi and Shitong Xiang of Fudan University together with senior authors Xiaolei Lin, Tianye Jia and Jianfeng Feng, combined data from the Adolescent Brain Cognitive Development study, the IMAGEN study, the Human Connectome Project and the UK Biobank. Together these cohorts provided 53,199 neuroimaging scans from 51,467 participants aged between 9 and 70 years. To guard against the possibility that their findings were an artefact of any single dataset, the team validated their trajectories in an independent collection of 2,127 people aged 12 to 89 drawn from the Nathan Kline Institute-Rockland Sample, the Cambridge Centre for Ageing and Neuroscience and the Brains and Minds in Transition study. In total, 6,992 participants in the discovery sample met criteria for problematic substance use, alongside 18,659 healthy controls.

A central methodological challenge in lifespan neuroscience is that brain volumes differ systematically with age, sex, scanner and study, making it difficult to compare a 14-year-old in one cohort with a 55-year-old in another. The researchers tackled this with normative modelling built on the Generalized Additive Model for Location, Scale and Shape, a statistical framework that establishes age-specific and sex-specific reference ranges for grey matter volume in 24 brain regions implicated in addiction. These regions spanned the ventral and dorsal striatum, amygdala, hippocampus, insula, anterior cingulate cortex and prefrontal cortex. Each participant’s volumes were then expressed as centiles relative to healthy controls of the same age, allowing the team to fit continuous trajectories of group differences across the entire lifespan rather than relying on isolated snapshots.

The resulting map revealed three clearly demarcated phases. Before the age of 25, problematic substance use was associated with a striking developmental imbalance: participants showed higher grey matter volume in the bilateral putamen, part of the dorsal striatum richly innervated by dopamine, but lower volume in the nucleus accumbens, insula, frontal pole, orbitofrontal and superior frontal cortices and the anterior cingulate cortex. Notably, while healthy controls showed declining volumes in the medial orbitofrontal cortex and nucleus accumbens through adolescence, individuals with problematic use exhibited the opposite, an increasing trend. The researchers interpret this as evidence of maturational imbalance between reward-driven striatal circuits and the prefrontal systems that normally exert top-down control.

Behavioural data collected within each cohort reinforced this interpretation. Adolescents and young adults with problematic substance use scored higher on rule-breaking, conduct problems and impulsivity, particularly sensation-seeking. Yet the picture was nuanced: they did not differ from controls on laboratory measures of stop-signal reaction time, cognitive flexibility or spatial working memory, and they showed no consistent preference for immediate rewards in monetary-choice tests. What did emerge were lower fluid and crystallized cognitive abilities, with effect sizes of 0.26 and 0.40 respectively. In mediation analyses, rule-breaking statistically mediated the link between whole-brain volumetric patterns and later substance use in the ABCD cohort, while impulsivity mediated the link between striatal-prefrontal volume contrasts and subsequent use, suggesting that these behavioural traits are not merely correlates but potential pathways connecting early brain structure to the initiation of risky use.

Between the ages of roughly 25 and 45, the structural signature shifted. Rather than deficits, the dominant pattern was enlargement: individuals with problematic substance use showed relatively higher grey matter volume in the bilateral insula, rostral anterior cingulate and lateral orbitofrontal cortex, regions that together constitute the brain’s salience network, which mediates conscious emotion, interoception and the conscious craving for addictive substances. These positive differences emerged from about age 30, peaked in the mid-40s and faded after 60. When the team projected their regional findings onto cortical gene-expression atlases, a dopaminergic marker combining FDOPA, the dopamine transporter, D1 receptors and NMDA receptors significantly explained the volumetric differences during this window, with the strongest effect around age 34. Genomically, this stage was anchored by a locus on chromosome 13 mapping to SUCLA2, a gene involved in mitochondrial energy metabolism, a pathway previously implicated in the neural effects of psychoactive substances and dopaminergic function.

After age 45, the pattern inverted again, this time toward widespread reduction. Significant differences emerged after approximately 56 years in subcortical structures and 68 years in cortical regions, with the amygdala, hippocampus, nucleus accumbens, medial orbitofrontal and superior frontal cortices showing progressively lower volumes and faster rates of decline in individuals with problematic use. The cognitive factor most strongly associated with these brain differences also shifted with age, moving from reward sensitivity in early adulthood toward general cognitive impairment and inhibitory control in mid-to-late adulthood. A genetic locus on chromosome 14, associated with alcoholism and tobacco use disorder, mapped to COA8, a gene involved in mitochondrial function and protection against oxidative stress. The authors suggest that these late-life reductions may reflect cumulative neurotoxicity and oxidative damage from prolonged substance exposure, potentially consistent with the premature brain aging reported in alcohol use disorder and drug addiction.

Sex shaped these trajectories in ways that were most pronounced early in life. In adolescence and early adulthood, females showed greater substance-related differences in the right nucleus accumbens, anterior cingulate, insula and orbitofrontal cortex, regions tied to reward motivation, interoception and emotional regulation, whereas males showed larger differences in the dorsal striatum and frontal pole, implicated in action selection and externalizing behaviour. The differences converged during mid-adulthood, then re-emerged in later life, with females showing greater hippocampal and caudate reductions and males greater orbitofrontal and superior frontal losses. The authors note this mirrors established sex-specific internalizing and externalizing pathways in psychiatric research, and argue that prevention efforts may need to be tailored differently for boys and girls.

Genetic correlations across the cohorts added a temporal dimension to the biology. Studies whose participants were closer in age showed higher genetic correlations with one another, and the correlation with a published genome-wide association study of addiction risk factors increased steadily with age, reaching a significant 0.72 in the UK Biobank sample. Combined with the identification of distinct shared loci at each life stage, CCDC91 on chromosome 12 in childhood, linked to white matter maintenance and neurodevelopmental disorders; SUCLA2 in early-to-mid adulthood; and the chromosome 14 locus in later life, the evidence points to gradually evolving genomic contributions to problematic substance use rather than a single static genetic vulnerability.

The authors are careful to spell out the limitations. Problematic substance use was defined using developmentally appropriate but not clinically equivalent criteria across cohorts, since consistent diagnostic assessments were unavailable, and most age ranges were dominated by a single cohort, raising the possibility of cohort effects despite statistical adjustment and independent replication. Grey matter trajectories were estimated primarily from cross-sectional data, and self-reported impulsivity did not always align with laboratory measures of inhibitory control. Even so, the framework that emerges is one in which addiction is not a static state but a moving target: a neurodevelopmental vulnerability before 25, a salience-network and dopaminergic reinforcement phase between 25 and 45, and a cumulative neurotoxic burden thereafter. Each stage, the authors argue, demands a different public health response, from school- and family-based prevention targeting impulsivity in adolescence, to early detection and intervention during the reinforcement years, to integrated monitoring of cognitive decline and broader health in older adults who use substances.

Subject of Research: Lifespan neuroimaging of brain structural and behavioural changes associated with problematic substance use

Article Title: Behavioural and brain volumetric changes associated with problematic substance use through the lifespan

Article References: Behavioural and brain volumetric changes associated with problematic substance use through the lifespan. (n.d.). https://doi.org/10.1038/s44360-026-00204-z

Image Credits: AI Generated

DOI: 10.1038/s44360-026-00204-z

Keywords: addiction, problematic substance use, grey matter volume, normative modelling, adolescence, dopamine, salience network, UK Biobank, genetics, impulsivity, brain aging, public health