A new article provides a comprehensive review of this important area where an evolutionary perspective offers great promise:
Keller, M. C. (2018). Evolutionary Perspectives on Genetic and Environmental Risk Factors for Psychiatric Disorders. Annual Review of Clinical Psychology, 14(1), https://doi.org/10.1146/annurev-clinpsy-050817-084854
ABSTRACT: Evolutionary medicine uses evolutionary theory to help elucidate why humans are vulnerable to disease and disorders. I discuss two different types of evolutionary explanations that have been used to help understand human psychiatric disorders. First, a consistent finding is that psychiatric disorders are moderately to highly heritable, and many, such as schizophrenia, are also highly disabling and appear to decrease Darwinian fitness. Models used in evolutionary genetics to understand why genetic variation exists in fitness-related traits can be used to understand why risk alleles for psychiatric disorders persist in the population. The usual explanation for species-typical adaptations—natural selection—is less useful for understanding individual differences in genetic risk to disorders. Rather, two other types of models, mutation-selection-drift and balancing selection, offer frameworks for understanding why genetic variation in risk to psychiatric (and other) disorders exists, and each makes predictions that are now testable using whole-genome data. Second, species-typical capacities to mount reactions to negative events are likely to have been crafted by natural selection to minimize fitness loss. The pain reaction to tissue damage is almost certainly such an example, but it has been argued that the capacity to experience depressive symptoms such as sadness, anhedonia, crying, and fatigue in the face of adverse life situations may have been crafted by natural selection as well. I review the rationale and strength of evidence for this hypothesis. Evolutionary hypotheses of psychiatric disorders are important not only for offering explanations for why psychiatric disorders exist, but also for generating new, testable hypotheses and understanding how best to design studies and analyze data.