Combinatorics

Pattern recognition (psychology)

Neuroscience

Mathematics

Resting state fMRI

Betweenness centrality

Default mode network

Cognition

Artificial intelligence

Psychology

Computer science

Independent component analysis

Centrality

Dynamic functional connectivity

Functional interactions among the brain regions define a meaningful representation of human brain activity. Estimation of time-varying interactions using functional MRI (fMRI) becomes challenging due to the complex brain signal dynamics. The evaluation and identification of these dynamical interactions between brain regions is an important component of the brain study. This paper considers Bayesian switching and empirical mode decomposition techniques for identifying connectivity strengths in a resting state fMRI dataset. fMRI data for three categories of subjects has been used for the analysis. Three categories are Alzheimers disease (AD), Control Normal (CN) and Early Mild Cognitive Impairment (EMCI) subjects. Time-varying dynamic connectivity interactions of the default network are analysed using Bayesian switching framework. Empirical Mode Decomposition of the resting state fMRI data is studied to understand the variations in connectivity strengths for each component of the data. Connectivity structure between different regions is determined from fast unfolding community framework. Graph analysis parameters such as modularity, closeness, average degree, global efficiency, radius, Wiener index, eccentricity, betweenness are used to analyse the functional connectivity interaction for three category subjects. The obtained results show significant differences in default network regions for Alzheimer, mild cognitive impairment and normal subjects.

Functional Connectivity Analysis of Default Mode Network for Healthy and Unhealthy Brains