Fig. 3: Transcriptomic decoding of the intersubject variability of white-matter functional connectivity (WMFC).

a Weighted gene expression map of the first component of partial least squares (PLS1). The combination of genes, the most explained variance of intersubject variability in WMFC, was obtained by the PLS correlation method. b Scatterplot of the relationship between PLS1 scores and intersubject variability in WMFC. c Ranked PLS1 loading. Genes that were strongly positively weighted on PLS1 (e.g., GPR22) correlated positively with intersubject variability, whereas genes that were strongly negatively weighted on PLS1 (e.g., FRYL) correlated negatively with intersubject variability. Functional enrichment analysis of top ranked genes (from c) with PLS1 + (d) and PLS1− (e) weights. All terms were retained after P < 0.05, false discovery ratio-corrected. nDensity, density estimate, scaled to a maximum of 1. Source data provided as Source Data file.