Meeting Banner
Abstract #2253

Network Diffusion Models for Functional Brain Connectivity Networks

Farras Abdelnour1, Henning Voss1, Ashish Raj1

1Radiology, Weill Cornell Medical College, New York, NY, United States

We present a simple and intuitive network diffusion model which produces an accurate mathematical description of the structure-function relationship. We hypothesize that resting state functional relationships between brain regions result from this diffusion process applied to the structural network during rest. The network diffusion model applied to the structural networks closely predicts both the spatial and temporal correlation structures seen in the functional networks. We compare our work with published models. The proposed model is simple yet offers improved estimates of functional connectivity.

Keywords

accurate activation active adults allows almost although anatomic applied approaches arbitrary better biophysical bold brain captured clockwise closed closely coding coefficient collected college color comparing completely configuration connected connectivities connectivity constant correlated correlation covariance critical deep description details determined develop diffusion discover dynamic dynamics either elucidated empirical entire established evaluated evaluating experimentally explanation expressions form fully function functional generalized generating generative goal gray healthy henning honey hypothesis hypothesize importantly initial intuitive involving isolated linear linearity located long major manner maps match mathematical matrix measured mechanistic medical model models need neocortex network networks neural node nodes nonlinear obviates outperforms overestimates pair performances phenomena pipelines predicted prediction predicts preprocessing previous process processing produces proposed published quantitative radiology realistic related relationship relationships represents requiring rest resting scale similarity similarly simple simplicity simply simulated solution spatial stat straightforward strongly structural structure structures subjects substrates sustain table temporal tend testable topology true underlying whether