Mirror effect of genomic deletions and duplications on cognitive ability across the human cerebral cortex

Cognitive deficits are common across many neurodevelopmental and psychiatric conditions, including those studied in the current set of PGC-CNV papers. How changes in regional gene expression across the cerebral cortex influence cognitive ability remains unknown. Population variation in gene dosage-which significantly impacts gene expression-represents a unique paradigm to address this question. We developed a cerebral-cortex gene-set burden analysis (CC-GSBA) to associate a trait with genomic deletions and duplications that disrupt genes with similar expression profiles across 180 cortical regions. We performed CC-GSBA across 180 cortical regions to test associations with cognitive ability in 260,000 individuals from general population cohorts. Most cortical gene sets were associated with a decrease in cognitive ability when deleted or duplicated, and this novel approach revealed opposing cortical patterns for the effect sizes of deletions and duplications. These cortical patterns of effect sizes followed the cortical gradient previously characterized at the molecular, cellular, and functional levels. We show that genes with preferential expression in sensorimotor regions demonstrated the largest effect on cognition when deleted. At the opposing end of the cortical gradient, genes with preferential expression in multimodal association regions affected cognition the most when duplicated. These two gene dosage cortical patterns could not be explained by particular cell types, developmental epochs, or genetic constraints, highlighting the fact that the macroscopic network organization of the cerebral cortex is key to understanding the effects of gene dosage on cognitive traits.