An information-theoretic analysis of resting-state versus task fMRI

Area of Science:

• Neuroimaging
• Cognitive Neuroscience
• Computational Neuroscience

Background:

• Resting-state functional magnetic resonance imaging (fMRI) is a popular neuroimaging technique.
• A quantitative comparison of information content between resting-state and task-based fMRI is lacking.
• Understanding neural responses requires precise data analysis.

Purpose of the Study:

• To systematically compare the information provided by resting-state fMRI versus task-based fMRI.
• To quantify the quality of inferences about neural responses from both fMRI paradigms.
• To assess the information gain regarding effective connectivity parameters.

Main Methods:

• Bayesian Data Comparison was employed to quantify information-theoretic data quality.
• Dynamic Causal Modelling (DCM) estimated effective connectivity parameters from cross-spectral densities.
• Data from 50 individuals from the Human Connectome Project were analyzed, comparing resting-state and Theory-of-Mind task conditions.

Main Results:

• The Theory-of-Mind task fMRI provided significantly more information (exceeding 10 bits) than resting-state fMRI.
• This superior information gain was linked to stronger effective connectivity during the active task.
• Bayesian Data Comparison formally quantified the higher precision of task-based fMRI.

Conclusions:

• Task-based fMRI, specifically the Theory-of-Mind task, offers superior informative value over resting-state fMRI for analyzing neural responses.
• Active cognitive tasks enhance effective connectivity, leading to greater information gain in fMRI.
• Further research is needed to determine if this trend is generalizable across cognitive systems.