Identifying Transdiagnostic Functional Connectivity Biomarkers for Cognitive Health and Psychopathology

Project Abstract
Psychiatric disorders are among the most common illnesses across the lifespan, with more than 75% of
individuals developing symptoms beginning in adolescence. Most psychiatric disorders include aspects of
cognitive dysfunctions that have been suggested to predispose individuals to develop the psychiatric conditions
and may serve as early markers of subsequent illness. Cognitive deficits and behavioral disturbances were
indicated to be related to broad-based functional impairments across disorders, which traditional case-control
studies are hard to capture. Following the Research Domain Criteria (RDoC) initiative launched by NIMH, there
is an urgent need for developing biomarkers that can cross current diagnostic boundaries and drive new ways
of defining psychiatric disorders based on dimensions of behavioral and neurobiological measures. In this project,
we will quantify functional connectivity biomarkers that capture brain dysfunctions spanning multiple psychiatric
disorders for an improved understanding of cognitive deficits and psychopathology. With high-density
electroencephalography (EEG), we will quantify connectivity biomarkers predictive of individual cognitive
behavior across the diagnostic spectrum (Aim 1). We will build a robust prediction model by combining relevance
vector machine and connectome-based predictive modeling to identify transdiagnostic neural circuits that map
the connectivity features to individual cognitive deficits. In Aim 2, we will design a dimensional approach based
on multiway canonical correlation analysis to robustly reveal neural circuit-correlated dimensions of
psychopathology. This approach allows us to jointly identify brain dysfunctions and dimensional behavioral
phenotypes. We will evaluate these tools and compare the obtained results between EEG and fMRI using a
large-scale transdiagnostic database from Healthy Brain Network. The proposed research will lead to an
innovative and generalizable solution for the robust quantification of transdiagnostic EEG connectivity
biomarkers that predict individual cognitive ability and delineate dimensions of psychopathological behavior
across psychiatric disorders. Successful outcomes of the project will produce translatable biomarkers crossing
current diagnostic boundaries in line with the goals of RDoC and provide a new avenue for EEG connectivity-
based transdiagnostic study of psychopathology, thereby representing an important step towards the
development of personalized therapeutics for improved mental health. We will release the developed tools to be
publicly available to facilitate other transdiagnostic neuroimaging studies in psychiatry.

Grant Number: 5R21MH130956-02
NIH Institute/Center: NIH
Principal Investigator: YEVGENY BERDICHEVSKY