Fig. 1

Enrolled population in this study and derivation of DLB-pattern. a A flow chart of enrolled participants in this study. b Cortical thinning in the DLB patients compared with healthy controls showed significant thinning in bilateral temporal, frontal, parietal, and occipital cortices (corrected P < 0.01). c Schematic figure showing the derivation of the DLB-pattern. The normalized cortical thickness matrix was collected in DLB patients and healthy controls followed by the derivation of the subject by subject covariance pattern. The colored boxes corresponding to each ROI of cortical thickness were assigned a label ranging from 1 to 148, as listed in Additional file 1: Table S1. The covariance pattern was fed into the PCA analyses which resulted in principal components of cortical thickness spatial pattern. The best model that differentiated DLB patients from healthy control was chosen and the normalized weights for all ROIs from the best model were defined as the DLB-pattern. The horizontal dotted line denotes 5% explained in the principal component analyses. The receiver operating characteristic curve for differentiation of DLB patients from healthy controls with DLB-pattern is presented as an  inset figure