Neurometric Encoding and Decoding: Using Multivariate Functional Connectivity Methods to Describe Cognitive States, Traits and Clinical Endophenotypes

Abstract

This research was undertaken for the purpose of demonstrating the neurometric utility of functional connectivity methods by combining metrics that utilize information derived from independent component analyses (ICAs) with traditional fMRI and graph theory analyses. The combination of these methodologies was used to establish traits and evaluate cognitive states from a behavioral genetics perspective, as well as to posit connectivity endophenotypes related to psychiatric and neurological diseases. The studies described below demonstrate that the metrics used to study intrinsic connectivity networks (ICNs) are useful tools for studying the in vivo brain in states of normalcy and disease. For instance, by examining ICNs across tasks and monozygotic twins, it was possible to establish these brain networks as traits. The ICNs were stable across cognitive states, while still exhibiting sensitivity to specific demands. In addition, the state- dependent modulation of these ICNs, as well as their other characteristics, was shown to be influenced by genetic factors in two separate twin samples. In the second twin sample, and a study of connectivity phenotypes related to schizophrenia, ICNs were useful for establishing the relationships between ICNs and tasks in both cases. The task-related characteristics and resting state profiles of ICNs were also useful for establishing novel endophenotypes of the disease states of schizophrenia and Parkinson's disease. Overall, this research serves to establish the study of the brain's intrinsic connectivity across the domains of both cognitive and clinical neuroscience and this work serves a contribution to the understanding of the dimensions along which normal and abnormal neurobiological functioning lie, and how intrinsic connectivity networks can be examined in both spheres.