Cognition Trajectories in Cognitive Training and Early Intervention Treatment Programs in Schizophrenia

点击翻译按钮获取中文摘要

PROJECT SUMMARY The Recovery After an Initial Schizophrenia Episode (RAISE) Early Treatment Program (ETP) is an exciting new evidence-based intervention for schizophrenia spectrum disorders. However, several questions arise from this study. Why are benefits in Quality of Life seen primarily during the first 6 months of treatment, and only for individuals whose duration of untreated psychosis is <74 weeks? Does a deeper understanding of treatment trajectories lead to new insights about prognosis or about novel leverage points to refine and optimize the treatment approach? The purpose of this project is to answer these questions by performing state-of-the-art computational analyses on data from RAISE-ETP and from our own cognitive training trial. Our interdisciplinary team of clinical investigators and data scientists will use data-driven predictive, causal clustering, and causal discovery analyses to examine clinical and cognitive response patterns in the RAISE-ETP data set. We will then validate the baseline causal model(s), describing the mechanistic relationship among patients' baseline characteristics derived from the RAISE-ETP data, against our own early psychosis data set, and explore the effects of cognitive training on key predictive variables identified in the RAISE-ETP analyses. We seek to understand whether data from our experimental neuroscience-informed cognitive training studies in recent-onset schizophrenia can suggest ways of enriching the RAISE-ETP approach in order to optimize outcomes for as many individuals as possible, not just those with a shorter duration of untreated psychosis. One of our main areas of focus will be cognition, since cognitive deficits predict psychosocial dysfunction even when symptoms are in remission and are unresponsive to current medications. In an exploratory Causal Graph Analysis performed on RAISE-ETP data, we find that baseline Cognition Composite performance is causally associated with Quality of Life (QLS) Total Score at 6 months, which in turn predicts the two-year QLS trajectory. Baseline Cognition Composite is also related to Positive and Negative Syndrome Scale (PANSS) Total Symptoms score at 6 months, via an unidentified latent variable linking QLS and PANSS Total Symptoms. Consistent with these causal relationships, exploratory growth curve modeling in our cognitive training data show that individuals assigned to training demonstrate a linear improvement in PANSS Total Symptoms 6 months after the intervention, while those control condition subjects show no such change. Our goal is to examine these findings in depth in order to inform the design of the next generation of “first episode” treatments.