Estimating time-varying brain connectivity networks from functional MRI time series

Area of Science:

• Neuroimaging
• Cognitive Neuroscience
• Network Science

Background:

• Functional brain networks are often assumed to be static in neuroimaging analyses.
• Emerging evidence indicates significant non-stationarity in brain connectivity, even during rest.
• There is a critical need for advanced methodologies to capture the dynamic nature of functional brain networks.

Purpose of the Study:

• To introduce the Smooth Incremental Graphical Lasso Estimation (SINGLE) algorithm for estimating dynamic brain networks from fMRI data.
• To investigate dynamic changes in functional network structure during a cognitive task.
• To identify brain regions whose connectivity profiles change dynamically with task demands.

Main Methods:

• Development and application of the Smooth Incremental Graphical Lasso Estimation (SINGLE) algorithm.
• Analysis of functional magnetic resonance imaging (fMRI) data from 24 healthy participants performing a Choice Reaction Task.
• Utilizing graph theoretic measures to quantify network properties.

Main Results:

• The SINGLE algorithm successfully estimated dynamic changes in brain network structure.
• Significant dynamic alterations were observed in the Right Inferior Frontal Gyrus and Right Inferior Parietal lobe during the task.
• These identified regions are consistent with known roles in cognitive control.

Conclusions:

• The Right Inferior Frontal Gyrus and Right Inferior Parietal lobe play crucial roles in attention and executive function.
• Dynamic network analysis provides insights into the regulation of brain regions during cognitively demanding tasks.
• The SINGLE algorithm offers a valuable tool for studying non-stationary brain connectivity.