Abstract: Prediction of risk for mental disorders is the cornerstone of “preventive psychiatry”. Emerging data report that up to 45% of subjects at familial high risk for schizophrenia (FHR) develop broad psychosis spectrum psychopathology and cognitive impairments although conversion to psychosis is less than 15%. Long term prospective examination of FHR adolescents/young adults to characterize early predictors of risk for psychosis, broad psychosis spectrum psychopathology and cognitive impairments is critical for prevention and early intervention. Our well characterized unique cohort of FHR adolescents/young adults that were extensively phenotyped between 2003 and 2008 is uniquely suited to address these goals. We published one predictive model each for the emergence of psychopathology with 80% accuracy (sensitivity 0.5, specificity 0.92, positive likelihood ratio 6.25, negative likelihood ratio 0.54), and for psychosis with 88% accuracy (sensitivity 0.17, specificity 0.99, positive and negative likelihood ratios 17 and 0.83, respectively). Our goal is to call back 75 FHR subjects from this cohort (≈75%) and examine the validity of these predictions with real-world outcomes of psychosis, broad psychosis spectrum psychopathology and cognitive impairments by including more refined baseline variables and genetic vulnerability using genomewide polygenic risk score and variants of complement gene C4. Independent lines of evidence recently showed association of C4 variants with schizophrenia risk at genomewide significance and increased synaptic pruning. The broad range of outcomes are examined within the Research Domain Criteria (RDoC) framework using dimensional changes in working memory and executive function constructs. We will also build a predictive model and a long range multimodal risk calculator for outcome incorporating the real world outcomes to improve sensitivity and positive likelihood ratio while preserving higher specificity and prediction accuracy, elucidate dimensional changes in working memory and executive function, and test these predictions in a replicate sample of FHR from the Edinburgh High Risk Study, which is the largest in the world (aim 1). Using a “top-down” approach, we will examine how the psychopathological outcomes and dimensional changes in working memory and executive function map on to neural circuitry with hippocampus as a hub. Ultra-high field (7 Tesla) structural MRI, diffusion tensor imaging (DTI) and blood oxygenation level dependent (BOLD) fMRI will examine hippocampal subfields and subdivisions, anatomical circuitry, and functional changes to the working memory and executive function tasks (aim 2). We will also derive preliminary outcome biotypes that may be biologically more meaningful for future validation. On successful completion of the project, our data is likely to provide useful leads to long-term biologically-based dimensional phenotypic outcomes of FHR subjects as well as for broad psychosis spectrum psychopathology and cognitive impairments. Identification of such outcomes may provide fundamental biobehavioral constructs for novel treatment designs. Such a long-term follow up of a richly phenotyped FHR individuals has not been attempted, to our knowledge.

Public Health Significance Schizophrenia is a chronic debilitating disorder with poor long term outcome. Predicting the onset of psychosis, early detection and intervention can reduce long term morbidity of the illness. Moreover, predicting the onset of psychosis along with a broad spectrum psychopathology related to psychosis is equally important because of significant morbidity associated with such sub-clinical symptoms. Even individuals with sub-clinical symptoms demonstrate significant social function deficits that increases the cost of caring for these individuals. Since using DSM-defined disorders do not carry the same biological connotation as other medical disorders, using biologically-based dimensional measures proposed in the Research Domain Criteria (RDoC) to characterize dimensional predictors of risk for psychosis and broad psychosis spectrum psychopathology may be more important to develop reliable tests to predict psychosis and design novel interventions.