In order to better understand how schizophrenia risk genetics operate at the cognitive and neural systems level, our recent work under this project has focused on (1) identifying novel phenotypes that either (a) distinguish both people with schizophrenia and unaffected family members from unrelated healthy individuals or (b) identify meaningful subgroups of individuals with schizophrenia, and (2) discovering associations between specific risk genes and schizophrenia-linked brain phenotypes. Ongoing studies in these areas permit better understanding of heritable, trait-related abnormalities in schizophrenia, of the underlying molecular biology responsible for such abnormalities, and of strategies for resolving some of the illness heterogeneity that makes biological research so challenging. With respect to identifying illness-related phenotypes, in the largest MRS study of its kind to date, we recently found that one measure of in vivo levels of GABA, a critical inhibitory neurotransmitter, in the dorsal anterior cingulate cortex (dACC) was reduced in patients with schizophrenia and their unaffected siblings as compared to healthy volunteers, was moderately heritable, and was nominally associated with positive symptom ratings. This work follows hypotheses of GABAergic disruption in schizophrenia primarily founded on post-mortem brain tissue investigations, and is in contrast to an important negative finding in which we confirmed an absence of MRS measured glutamate levels in the same brain region. Efforts to elaborate on these observations and, in the case of GABA, determine underlying genetic mechanisms are ongoing. Another strategy for identifying-illness related phenotypes is to employ hypothesis- and data-driven parcellations of multidimensional data to generate meaningful subgroupings of clinical populations. For example, we completed analyses based on stable negative and distress symptom dimensions, identifying deficit, distress, and low symptom subgroups with starkly different demographic, diagnostic, clinical, cognitive, and personality characteristics, as well as distinct working-memory related brain activation patterns. Other current subgrouping analyses examine cluster analysis-derived patterns of current vs. pre-diagnosis cognitive performance. These subgrouping schemes are now being tested as phenotypes for further studies of clinical, brain and genetic differences, which may ultimately prove relevant in assessing treatment needs. In so doing, we have now expanded our portfolio of genetic discovery. For example, by taking advantage of genome-wide typing, these cognitive trajectory subgroups in fact show quite distinct patterns of genetic association with two broad genetic indexes summarizing, respectively, the genetics of schizophrenia risk and of general cognitive performance. Thus, we found a convergence between stratification based on common cognitive measures and stratification based on schizophrenia and cognitive genetics. For most individuals in our schizophrenia cohort, summary scores indexing common genetic risk for the condition were predictive of cognitive impairment. Strikingly, for one subgroup, that risk was entirely decoupled from cognitive performance. Future work is underway to further test the naturally following hypothesis that different patterns of genetic risk identify subgroups with distinct illness etiologies and accompanying different trajectories of cognitive development through adolescence. These results provide a fascinating complement to recent targeted investigations of cognitive genetics, in which we provided evidence that a single genetic variant related to sodium channel biology affects gene expression, contributes to measured cognitive abilities in individuals with schizophrenia, unaffected family members, and healthy cohorts, and is tied both to prefrontal inefficiency during working memory and interregional functional correspondence within the frontal cortex during rest.
Showing the most recent 10 out of 29 publications
