PROJECT TITLE :

Causality Analysis of fMRI Data Based on the Directed Information Theory Framework

ABSTRACT:

This paper aims to conduct fMRI-primarily based causality analysis in brain connectivity by exploiting the directed data (DI) theory framework. Not like the well-known Granger causality (GC) analysis, that relies on the linear prediction technique, the DI theory framework does not have any modeling constraints on the sequences to be evaluated and ensures estimation convergence. Moreover, it will be used to generate the GC graphs. During this paper, first, we have a tendency to introduce the core ideas within the DI framework. Second, we tend to gift a way to conduct causality analysis using DI measures between 2 time series. We tend to provide the detailed procedure on how to calculate the DI for 2 finite-time series. The 2 major steps concerned here are optimal bin size choice for knowledge digitization and chance estimation. Finally, we tend to demonstrate the applicability of DI-based mostly causality analysis using both the simulated knowledge and experimental fMRI data, and compare the results with that of the GC analysis. Our analysis indicates that GC analysis is effective in detecting linear or nearly linear causal relationship, however might have difficulty in capturing nonlinear causal relationships. On the opposite hand, DI-primarily based causality analysis is additional effective in capturing both linear and nonlinear causal relationships. Moreover, it is observed that brain connectivity among totally different regions generally involves dynamic two-way information transmissions between them. Our results show that when bidirectional data flow is gift, DI is additional effective than GC to quantify the general causal relationship.