The study workflow.

The analysis pipeline incorporated three sequential components. (a) Calculation of adiposity accrual velocity using four anthropometric indicators (zBMI, WC, BRI, WHtR) and brain developmental patterns via symmetrical percentage change (SPC) from longitudinal neuroimaging. (b) A two-stage analytical approach employing linear mixed-effects models to examine both cross-sectional associations (baseline adiposity metrics vs. cognition/brain morphology) and longitudinal relationships (adiposity accrual velocity vs. follow-up outcomes), with mediation analysis of neural pathways. (3) Subgroup analysis of obese children between those showing fat accumulat versus reduction, in terms of neurocognitive trajectories, cognitive and neurodevelopmental outcomes. respectively. Abbreviation: zBMI, body mass index z-score, calculated according to the WHO Child Growth Standards; WC, waist circumference; BRI, body roundness index; WHtR, waist-to-height ratio; Δ, rate of change, calculated as the difference in indicators divided by the difference in age; CT, cortical thickness; CV, cortical volume.

The demographic, adiposity and lifestyle statistics of the study population.

Associations between adiposity indicators, and neuroimaging features, and neurocognitive characteristics.

(a) The associations between adiposity and cognitive task scores. * p<0.05, **p<0.01, ***p<0.001 (FDR-adjusted). (b) The associations between adiposity and brain morphology. The colored brain areas represented the significant associations after FDR correction, and the value indicated the z-value. (c) Significant mediating brain regions for a representative cognitive task (picture vocabulary score). The colored brain areas represented the significant mediation effects, and the value indicated the average casual mediation effects (ACME). Abbreviation: BMI, body mass index; WC, waist circumference; BRI, body roundness index; WHtR, waist-to-height ratio; CT, cortical thickness; CV, cortical volume.

Association between adiposity indicators and cognitive task scores

Differential contributions of 4-year follow-up adiposity and adiposity accrual rate on follow-up cognitive performance in joint models.

Association between adiposity accrual rate and cognitive development.

Relations between adiposity accrual rate, neurocognitive characteristics, and brain development.

(a) The associations between adiposity accrual rate and SPC of cortical CT or CV. The colored brain areas represented the significant associations after FDR correction, and the value indicated the β-value. (b) Representative mediation models illustrating the indirect effects of ΔBRI on flanker performance via brain development. Abbreviation: BMI, body mass index; BRI, body roundness index; CT, cortical thickness; CV, cortical volume; SPC, symmetrical percentage change; ACME, average casual mediation effect; ADE, average direct effect; Prop_mediated, proportion mediated (ratio of ACME to total effect).

Neurodevelopmental trajectories across central fat change groups.

This figure compares brain and cognitive development among obese adolescents with distinct central fat trajectories: increasing (orange), decreasing (yellow), and stable (green) BRI groups. (a) shows group separation in BRI changes over time. (b) reveals differential cognitive changes, with significant group differences in flanker task slopes (*p<0.05, **p<0.01). (c) shows global cortical thickness trajectories across adiposity groups, demonstrating distinct growth pattern among all three groups. (d) presents region specific cortical thickness change patterns in obese versus control groups, with color intensity reflecting statistical significance (−log10[p]).