In the healthy brain, large-scale white matter architecture and local neuronal membrane properties facilitate seamless transitions between cognitive states. Patients with schizophrenia display significant white matter abnormalities with disorganized brain activity. However, the degree to which dysfunctional brain activity in psychosis arises from structural or functional pathology remains unknown. The identification of conclusive neuroimaging findings in this cohort has been challenged by 1) inadequate methods to assess distributed multimodal pathological phenotypes, and 2) the significant pathogenetic heterogeneity in schizophrenia. Practically, the latter challenge can be in part addressed by the study of chromosome 22q11.2 deletion syndrome (22qDS), where the prevalence of clinical schizophrenia is 25-fold higher than that of healthy individuals. The former challenge can be addressed by recent advances in network science and machine learning, which have generated insights about structure-function relationships in the healthy brain. Utilizing these methods to study the spectrum of psychotic symptoms in a population with a defined genetic lesion is a promising direction for investigating psychosis pathophysiology. In this proposal, we describe the development of a novel time-point-based analysis of functional neuroimaging data to study structure-function relationships in a sample of patients with 22qDS currently being seen at the Hospital for the University of Pennsylvania. Using neuroimaging data from a large sample of youths (n = 690) acquired through the Philadelphia Neurodevelopmental Cohort, our preliminary analyses demonstrate previously uncharacterized relationships between brain structure, brain activity, and working memory performance. In this proposal, we aim to 1) compare brain state transition dynamics across the psychotic spectrum of 22qDS, 2) compare structure-function coupling in 22qDS to healthy controls and relate structure-function coupling to psychotic symptom severity, and 3) relate state transition dynamics to working memory performance in 22qDS. A better understanding of the underlying mechanism of psychosis-spectrum symptoms would lay the groundwork for the development of targeted therapies for psychosis. Furthermore, utilizing a cohort with a known genetic lesion provides a unique opportunity to bridge our understanding of molecular mechanisms with neuroimaging biomarkers for psychosis-spectrum symptoms.
Targeted, informed therapies in neuropsychiatric disease remain elusive in part due to a lack of understanding of how complex patterns of brain activity arise from stationary white matter architecture. We propose to apply a novel method based on machine learning and network science to provide a new perspective on structure-function relationships in healthy controls as well as a sample of patients with psychosis secondary to chromosome 22q11.2 deletion syndrome. Our study will leverage the presence of a consistent genetic lesion in this patient cohort to generate findings that could lay the groundwork for the development of interventions to restore healthy brain dynamics in individuals with psychosis-spectrum symptoms.