It has proved hard to identify any reproducible and diagnostically-specific biomarkers in psychiatry – despite significant investment and a growing arsenal of tools for probing the human brain. Our lab has been concerned with two major roots to the problem. First, behaviorally-defined psychiatric diagnoses display profound overlap and heterogeneity at epidemiological, clinical, neurobiological and genetic levels of analysis. Second, psychiatric disorders emerge in the context of human brain development, which is exceptionally complex, protracted and experimentally inaccessible. Chromosomal aneuploidies and recurrent sub-chromosomal copy number variations (henceforth collectively “CNVs”) offer a potential passage through this impasse because they represent genetically-defined and “quasi-experimental” models of developmental risk for neuropsychiatric morbidity. This talk will present a series of studies that harness deep-phenotypic data in multiple CNV disorders to resolve pathways of biological risk for psychopathology in humans. Particular emphasis will be placed on the value of integrative analytic approaches that: (i) identify convergence and specificity by parallel analysis of different CNV groups, (ii) hone patient studies using information from large-scale models of typical brain development, and (iii) decode in vivo imaging data using postmortem maps of human brain cytoarchitecture and gene expression.