Fig. 1

Timeline of methodological approaches common in neuroimaging-genomics studies of neuropsychological disorders. The field of neuroimaging genomics was initiated in the early 2000s using a hypothesis-driven candidate-gene approach to investigate brain and behavior phenotypes [2, 3]. Towards the end of the decade, other candidate-gene approaches, investigating alternative genetic models, began to emerge. These included gene–gene interactions [172], gene–environment interactions [7], and epigenetic effects [6]. Simultaneously, hypothesis-free approaches such as genome-wide association studies (GWAS) were initiated [173] and the need for increased statistical power to detect variants of small individual effects soon led to the formation of large-scale consortia and collaborations [36, 37]. The emergence of the “big data” era presented many statistical challenges and drove the development of multivariate approaches to account for these [174]. GWAS of neuropsychological disorders soon identified significant associations with genetic variants with unknown biological roles, resulting in candidate neuroimaging genomics studies to investigate and validate the genetic effects on brain phenotypes [175]. The emergent polygenic nature of these traits encouraged the development of polygenic models and strategies to leverage this for increased power in genetic-overlap studies between clinical and brain phenotypes [114]. Most recently, hypothesis-free approaches are starting to extend to alternative genetic models, such as gene–gene interactions [70]