First establishment of population-specific brain volume reference standards for Chinese children and adolescents: an AI-based segmentation study.

Objective: To establish the first AI-based, population-specific reference standards for brain structural volumes in healthy Chinese children and adolescents, and to elucidate unique developmental trajectories that may differ from existing Western norms. Methods: In this cross-sectional study, 1,100 healthy participants aged 6-18 years were enrolled. T1-weighted images were acquired using a 3.0T MRI scanner, and regional brain volumes were quantified using an AI-based automated segmentation tool. Multiple comparisons were corrected using the false discovery rate (FDR) method. Differences in brain volume by sex, age, and hemisphere, as well as correlations with age, were analyzed. Results: After controlling for age, males exhibited larger total intracranial volume (P < 0.001) and left cerebral white matter volume (P = 0.015), whereas females showed larger volumes in the bilateral parahippocampal gyri (left: P = 0.023; right: P = 0.011) and left fusiform gyrus (P = 0.026). Significant hemispheric asymmetries were observed in multiple regions. Cerebrospinal fluid (CSF) and bilateral white matter volumes were positively correlated with age (P < 0.001), while volumes in certain frontal and temporal lobe regions, as well as subcortical structures, were negatively correlated with age (P < 0.05). Conclusion: This study is the first to establish reference values for brain volume in healthy Chinese children and adolescents aged 6-18 years using AI-based automated segmentation. It reveals sex- and age-related differences in brain structural development, thereby providing a valuable quantitative reference for future research on brain development deviations. These population-specific standards may provide a theoretical foundation for improving diagnostic accuracy in Chinese children, though their actual utility in reducing misclassification requires future empirical validation.