Childhood trauma acts as an independent risk factor for mental disorders. This study utilized multi-modal MRI techniques to integrate cortical structural measurements and functional connectivity (FC) to identify neurobiological markers of trauma. This study recruited 215 participants, divided into a trauma group (n = 57) and a well-matched non-trauma group (n = 158) based on the childhood trauma questionnaire. We compared differences in cortical volume (CV) and surface area (SA) between the groups and performed a seed-based FC analysis using seeds at identified structural clusters. A machine learning approach with logistic regression was employed to classify individuals with and without childhood trauma. The childhood trauma group showed uniformly lower SA and CV. Reduced SA was found in a cluster consisting of the left precentral gyrus, postcentral gyrus, and paracentral lobule, and decreased CV was observed in a cluster involving the left postcentral gyrus. Seed-based FC analyses revealed decreased FC between the CV-cluster and regions of the bank of the superior temporal sulcus, inferior parietal gyrus, and supramarginal gyrus. In contrast, SA-related clusters showed increased FC with the left postcentral gyrus, superior parietal gyrus, and supramarginal gyrus. The logistic regression model, based on these structural and functional features, achieved a statistically significant classification accuracy of 78 % (p <
0.001) in distinguishing groups with and without childhood trauma. The childhood trauma group exhibits abnormalities in cortical structure and FC which are related to aberrant emotional and cognitive functions. These findings may serve as neuroimaging biomarkers of childhood trauma.