An integrative computational interrogation of circuit dysfunction inschizophrenia via neural timescales

SUMMARY/ABSTRACT Schizophrenia is a devastating and burdensome illness the mechanisms of which remain elusive. Contributing to their elusiveness are a highly complex set of genetic factors, proposed etiological and pathophysiological pathways, and phenotypic manifestations. To address this complexity, we propose a hybrid method combining data-driven approaches to large-scale multimodal datasets and theory-driven computational approaches in order to provide a theoretically constrained framework bridging genetics, development, circuit function, cognition, and phenomenology of schizophrenia. To that end, and in response to ‘Notice of Special Interest regarding the Use of Human Connectome Data (HCP) for Secondary Analysis’, we will use data from up to 64,000 individuals, including healthy individuals and patients with schizophrenia and other disorders, from various HCP-related projects as well as the UK Biobank. We specifically propose measuring intrinsic neural timescales (INT) from resting-state fMRI data as a theory-driven index of excitation/inhibition (E/I) imbalance in cortical microcircuits. First, extending our prior work we aim to confirm and further characterize INT alterations in schizophrenia (widespread trait-like INT reductions and local hierarchy-dependent INT modulations in relation to psychotic symptoms) and to test their specificity relative to other disorders. Second, we will evaluate the developmental trajectories of INT and characterize the genetic profile of this fMRI measure and its overlap with the genetic profile for schizophrenia risk. Third, given the role of E/I ratio in cortical microcircuits in supporting working-memory computations, we will examine the relationship between INT and working-memory activation and performance. We will further seek to establish INT as a circuit-level mediator of polygenic risk for schizophrenia on cognitive deficits. Throughout, we will use well-powered, rigorous, state-of-the-art fMRI and statistical data-driven methods suitable for large-scale studies and HCP-style fMRI sequences, including cross- validation and tests of generalizability. Together with a strong theoretical foundation and using biophysical modeling to complement fMRI analyses, this hybrid—theory- and data-driven—approach will facilitate an integrated understanding of the circuit-level mechanisms bridging distal genetic-risk factors and proximal manifestations of schizophrenia. In particular, the combination of cutting-edge cell-type enrichment analyses of GWAS (which in schizophrenia have suggested converging enrichment in excitatory and inhibitory cortical cells) and biophysical modeling at the level of cortical microcircuits of interacting excitatory and inhibitory cellular populations will provide an interpretation of disparate data in terms of convergent cell- and circuit-level pathways. In doing so, this project will validate a theoretically informative, interpretable, translatable, and scalable resting-state fMRI measure—INT—that may be relevant across several disorders and, which additionally owing to its high reliability and ease of acquisition, has high potential as a candidate biomarker.