Using a computational and network neuroscience framework to study pharmacological manipulations of state representation processes in early psychosis

University Of Minnesota, Minneapolis MN

Investigators

Linked publications & trials

Abstract

PROJECT 5 - PROJECT SUMMARY The purpose of PROJECT 5 is to manipulate the computational parameters being studied in PROJECTS 1-4 to reveal mechanistic information relevant to the expression and treatment of cognitive and motivational deficits in psychosis. Based on our analyses of computational parameters derived from state representation processes in a cognitive control and a reward-based decision-making task, we have observed dimensions of computational variation (subprofiles) in early psychosis that show meaningful preliminary clinical, cognitive, and neurophysiologic correlates. These subprofiles are characterized by various combinations of impairments in state estimation, maintenance, and/or learning. Based on parallel findings in our Center’s mouse and macaque projects, we hypothesize that some of these subprofiles may more predominantly reflect cortical NMDAR hypofunction vs. striato-cortical (dopamine-glutamate) dysfunction, with differential aberrations in large-scale neural network dynamics across the central executive, salience, and default mode networks. We hypothesize that these subprofiles may deepen our understanding of psychosis illness heterogeneity and inform personalized interventions for cognition and motivation. Participants will be drawn from PROJECT 4, where they will have been assessed with clinical, behavioral and EEG measures. In Aim 1, we investigate computational parameter and EEG changes in healthy individuals after psychosis-relevant pharmacological manipulations of NMDAR and dopaminergic function. We will randomly assign 75 healthy volunteers to 1 of 5 single-dose conditions: NMDAR-antagonism alone (ketamine); increased striatal dopamine (DA) tone alone (modafinil); both probes together; DA antagonism (risperidone) and placebo. We test hypotheses that NMDAR-antagonism, increased DA, their combination, and D2 receptor blockade each affect state representation processes ways consistent with observations from animal models. In Aim 2, we perform a within-subject study of changes in computational parameters and precision functional connectomics in psychosis after pharmacologic manipulations that enhance either NMDAR or dopaminergic function. We will stratify 20 participants with early psychosis on two of our identified computational subprofiles that exhibit validity with prior research, and randomly assign the order of receiving the NMDAR co-agonist d-serine, the stimulant modafinil, and placebo. Using highly innovative Precision Functional Mapping methods, we will determine the within-subject changes after each probe, with a focus on central executive, salience, and default mode network dynamics. Our goal is to use the pharmacological manipulations as mechanistic probes to test for specific differential effects on microcognitive processes and network dynamics across the 2 subprofiles, with the goal of setting the stage for future precision psychiatry research.