The authors applied latent profile analysis (LPA) to derive person-level profiles of mental health using the mean and standard deviation of three measures: distress (K10), wellbeing (COMPAS-W) and suicidal ideation (SIDAS). LPA produced three discrete clusters among 135 adolescents (688 observations): a low distress profile (N=110; 50% female), a moderate distress with greater suicidal ideation profile (N=14; 64% female) and a moderate distress with lower suicidal ideation profile (N=11; 91% female). The profiles reflect differences in both symptom level and stability over the observation period as captured by person-level summary statistics.
Predicted brain-age (BrainAGE) was estimated using the CentileBrain Global-BrainAGE pipeline. Input features were morphometric measures derived from FreeSurfer processing of structural MRI. BrainAGE was operationalized as the difference between predicted brain-age and chronological age (predicted minus chronological age), consistent with established brain-age gap approaches.
Data came from the Longitudinal Adolescent Brain Study and included 135 adolescents aged 12.0–16.9 years, with a total of 688 observations across the study window. The report presents cross-sectional and longitudinal analyses using these repeated observations. The manuscript is a preprint and has not undergone peer review.
The primary analyses tested whether profile membership related to BrainAGE. After adjusting for chronological age and sex, the moderate distress with lower suicidal ideation cluster showed a significantly higher BrainAGE relative to the low distress cluster (B = 1.87, SE = 0.76, p = .015). By contrast, the moderate distress with greater suicidal ideation cluster did not differ significantly from the low distress cluster on BrainAGE.
Chronological age itself was positively associated with BrainAGE in the sample (B = 0.67, SE = 0.07, p < .001), indicating that the predicted brain-age gap increased with older chronological age within the studied developmental window.
Associations were evaluated using linear regression and linear mixed-effects models, with adjustments for chronological age and sex. The notable effect size reported was an increase of 1.87 years in BrainAGE for the moderate-distress/low-suicidality group versus the low-distress group, with standard error 0.76 and p = .015. The moderate-distress/high-suicidality group did not show a significant BrainAGE difference compared to the low-distress group in these adjusted models. Details on other covariates or sensitivity analyses were not reported in the abstract and thus are not available from the source text.
The authors interpret findings as evidence that distinct adolescent mental-health profiles—characterized by different combinations of distress, wellbeing, and suicidal ideation—may be associated with differences in MRI-derived brain-age. The specific association found for the moderate-distress/low-suicidality cluster suggests that elevated BrainAGE can occur in profiles marked by moderate psychological distress even when suicidal ideation is lower or less stable. These observations support continued investigation of BrainAGE as a potential marker of brain maturation or biological ageing in adolescent mental health research.
Clinically, the findings highlight heterogeneity in adolescent mental-health presentations and suggest that neurodevelopmental timing or maturation captured by BrainAGE may be differentially related to distress and suicidality. However, because this work is preliminary and reported as a preprint, it should not yet guide clinical decision-making.
The authors note key limitations that temper interpretation. Cluster sizes were small for the two moderate-distress profiles (N=14 and N=11), which reduces statistical power and increases risk of unstable estimates. The developmental window examined (ages 12–16.9 years) is constrained, and longitudinal associations were less robust across small extensions of this range, indicating sensitivity to the sampled age span. The study is a preprint and has not been peer reviewed. The authors call for replication in larger samples and broader developmental ranges to confirm whether the observed profile-specific BrainAGE differences are robust.
In this preprint from the Longitudinal Adolescent Brain Study, distinct mental-health profiles derived from distress, wellbeing and suicidal ideation relate variably to MRI-derived BrainAGE. One profile—moderate distress with lower suicidal ideation—showed higher BrainAGE compared with a low-distress group (B=1.87, SE=0.76, p=.015), while a moderate-distress profile with greater suicidal ideation did not differ. Results are preliminary, limited by small cluster sizes and a narrow adolescent age range, and require replication before clinical application.