A bayesian approach to examining default mode network functional connectivity and cognitive performance in major...

Area of Science:

• Neuroscience
• Cognitive Science
• Computational Psychiatry

Background:

• Resting-state functional connectivity (RSFC) associations with cognitive performance show high variance in healthy and depressed individuals.
• Existing methods for estimating RSFC may lack the precision needed to uncover subtle relationships with cognition, particularly in clinical populations.

Purpose of the Study:

• To develop and validate a Bayesian spatio-temporal model for precise estimation of RSFC.
• To investigate the correlation between spatially-adjusted functional connectivity (saFC) within the default mode network (DMN) and cognitive performance in depressed and nondepressed participants.
• To compare the efficacy of the Bayesian approach against conventional methods in detecting diagnostic differences in connectivity-cognition relationships.

Main Methods:

• A Bayesian spatio-temporal model was proposed to estimate saFC, focusing on the extended DMN.
• Multiple linear regression analyses examined the relationship between DMN saFC and cognitive performance across four domains, controlling for age, sex, and education.
• The Bayesian saFC estimator was compared with a conventional ROI-based average functional connectivity (AVG-FC) estimator.

Main Results:

• The Bayesian approach successfully identified significant correlations between DMN saFC and cognitive performance in specific ROI pairs (PCC, ACC).
• Only the Bayesian method detected significant diagnostic differences in the connectivity-cognition association within PCC and ACC regions.
• The Bayesian model yielded a saFC estimator with smaller variance, enhancing precision and reliability.

Conclusions:

• A precise and reliable saFC estimator derived from the Bayesian model facilitates scientific discovery in neuroimaging research.
• The Bayesian approach offers advantages over conventional methods for studying RSFC and cognitive performance, especially in clinical populations like depression.
• This methodology can potentially uncover cognitive-connectivity relationships previously obscured by high variance in traditional estimators.