Task-Based Functional Connectomes Predict Cognitive Phenotypes Across Psychiatric Disease

Importance We show that three common approaches to clinical deficits (cognitive phenotype, disease group, disease severity) each offer useful and perhaps complimentary explanations for the brain’s underlying functional architecture as affected by psychiatric disease. Objective To understand how different clinical frameworks are represented in the brain’s functional connectome. Design We use an openly available dataset to create predictive models based on multiple connectomes built from task-based functional MRI data. We use these models to predict individual traits corresponding to multiple cognitive constructs across disease category. We also show that these same connectomes statistically differ depending on disease category and symptom burden. Setting This was a population-based study with data collected in UCLA. Participants Healthy adults were recruited by community advertisements from the Los Angeles area. Participants with adult ADHD, bipolar disorder, and schizophrenia were recruited using a patient-oriented strategy involving outreach to local clinics and online portals (separate from the methods used to recruit healthy volunteers)