DESCRIPTION (provided by applicant): Psychosis is ranked as one of the most disabling conditions worldwide. Functional deterioration occurs at the fastest rate during the early phase of psychosis, which is both a "critical period" of neuronal and psychosocial plasticity and a "window of opportunity" during which treatment may confer disproportionately favorable outcomes. There is a large variability in the recovery trajectory and outcome of first episode of psychosis [FEP] patients. However, current algorithms do not allow differences in biological background or specific neurobiological profile to guide individualized treatments. Many neurobiological alterations, including neurocognitive deficits, subtle structural brain changes, electroencephalographic (EEG) measures, and clinical variables, collectively contribute to variance in functioning in FEP. We have used K-means multivariate classification analyses and identified three distinct "Bio-classes" of psychosis independent of diagnosis. Within each class, patients shared a similar neurobiological profile that uniquely distinguished among the groups. These data present a diagnosis-free approach to integrate information across biomarkers, yielding neurobiologically distinct subgroups and provide strong evidence supporting the superiority of neurobiological vs. clinical classification in differentiating psychotic disorders. Building on our prior work, the goal of this project is to stratify FEP patients into homogeneous subgroups based on patients' unique neurobiological profiles and relate these profiles to later functional outcome. This project will address four questions: 1) Are bio-classes characterized in the FEP similar to those observed in chronic psychosis patients? [Aim 1a]; 2) Are there structural neuroanatomical features associated with distinct bio-classes [Aim 1b]; 3) are bio-classes characterized at baseline predictive of later functional outcomes two years later, controlling for baseline severity and other confounding factors? [Aim 2a]; and 4) are bio-classes superior to DSM diagnoses in predicting functional outcomes? [Aim 2b]. This project will recruit a total of one hundred (n=120) FEP patients at two FEP clinics and 100 demographic matched controls. We will follow up FEP patients every 6 months for two years and collect an expanded biomarker panel as well as symptomatic and functional outcome measures at each time point. Patients will be bio-classified into one of the three bio-classes at baseline using a constellation of cognitive and EEG biomarkers. A comprehensive battery of clinical, premorbid, treatment, biomarker, and functional variables will be obtained at each time point. Brain structural measures, not used in forming the bio-classes, will be used to validate the distinctness of the bio-classes. We will examine the relationships between bio-classes and functional outcome and between DSM diagnoses and functional outcome at each time point. We will compare the two mo...