Imaging Genetics and Genomics in Psychiatry: A Critical Review of Progress and Potential

Abstract

Imaging genetics and genomics research has begun to provide insight into the molecular and genetic architecture of neural phenotypes and the neural mechanisms through which genetic risk for psychopathology may emerge. As it approaches its third decade, imaging genetics is confronted by many challenges, including the proliferation of studies using small sample sizes and diverse designs, limited replication, problems with harmonization of neural phenotypes for meta-analysis, unclear mechanisms, and evidence that effect sizes may be more modest than originally posited, with increasing evidence of polygenicity. These concerns have encouraged the field to grow in many new directions, including the development of consortia and large-scale data collection projects and the use of novel methods (e.g., polygenic approaches, machine learning) that enhance the quality of imaging genetic studies but also introduce new challenges. We critically review progress in imaging genetics and offer suggestions and highlight potential pitfalls of novel approaches. Ultimately, the strength of imaging genetics and genomics lies in their translational and integrative potential with other research approaches (e.g., nonhuman animal models, psychiatric genetics, pharmacologic challenge) to elucidate brain-based pathways that give rise to the vast individual differences in behavior as well as risk for psychopathology.

Keywords: Candidate; Genetics; Genomics; Imaging; MRI; Neurogenetics; Polygenic.

Figure 1. Imaging Genetics and Genomics Models

(A) The traditional imaging genetics and genomics model suggests that genetic variation confers risk for psychopathology indirectly through its influence on the brain. This theoretical model is well suited for traditional mediation models estimating indirect associations (demarcated with the dashed line), through which genetic background is linked to behavior through neural phenotypes. (B) Imaging genetics and imaging genomics redux: In the future, as imaging genetics and genomics expand to include larger and longitudinal samples it will be possible to evaluate a more complete interactive model in which bidirectional relationships between the genome, brain, and behavior may be investigated in the context of environmental experience and peripheral biological markers. For example, socioeconomic status has been associated with epigenetic modifications that are, in turn, related to psychiatrically-relevant brain function (145). Moreover, environmental experience (e.g., trauma experienced during early life) moderates genetic associations with neural phenotypes and associations between neural phenotypes and behavior (20,146). Further, genetic background influences peripheral indices such as gut microbiome (147), which in turn has been linked to neural phenotypes and psychopathology (148). As a result, a more complete mechanistic understanding requires multiple levels of analyses in the context of longitudinal and convergent data. Currently, convergence across multiple methods and studies testing legs separately is attainable. Informed by such studies, in the future, as large multimodal longitudinal studies develop, it is plausible that more complete pathways could be tested in the framework of a single study.