Reduced vmPFC-insula functional connectivity in generalized anxiety disorder: a Bayesian confirmation study.

Area of Science:

• Neuroscience
• Psychiatry
• Computational Statistics

Background:

• Generalized anxiety disorder (GAD) is associated with altered brain network activity.
• Traditional statistical methods like null-hypothesis significance testing (NHST) may limit the discovery of disorder-relevant functional connectivity (FC) patterns.
• Bayesian statistical frameworks offer a complementary approach to analyze complex neuroimaging data.

Purpose of the Study:

• To compare Bayesian and frequentist (NHST) statistical approaches for analyzing resting-state functional connectivity (FC) in generalized anxiety disorder (GAD).
• To identify specific brain region FC differences between females with GAD and healthy controls.
• To investigate the utility of Bayesian modeling in uncovering novel FC abnormalities in GAD.

Main Methods:

• Resting-state functional magnetic resonance imaging (fMRI) data were acquired from females with GAD and matched healthy controls.
• Eleven a-priori hypotheses regarding functional connectivity (FC) were tested using both Bayesian multilevel models and frequentist t-tests.
• Statistical analyses were performed comparing FC between predefined brain regions, including the ventromedial prefrontal cortex (vmPFC), posterior-mid insula (PMI), anterior insula, amygdala, and dorsolateral prefrontal cortex (dlPFC).

Main Results:

• Both Bayesian and frequentist methods confirmed reduced FC between the ventromedial prefrontal cortex (vmPFC) and posterior-mid insula (PMI), associated with anxiety sensitivity.
• Frequentist analysis, after multiple comparison correction, did not find significant differences in FC for vmPFC-anterior insula, amygdala-PMI, and amygdala-dlPFC pairs.
• Bayesian modeling provided evidence for decreased FC in these latter region pairs in the GAD group and revealed additional FC abnormalities in previously undescribed regions.

Conclusions:

• Bayesian statistical modeling is valuable for detecting decreased functional connectivity (FC) in brain regions like the vmPFC, insula, amygdala, and dlPFC in females with generalized anxiety disorder (GAD).
• The Bayesian framework identified FC abnormalities missed by frequentist approaches, highlighting its sensitivity in clinical neuroimaging studies.
• This study underscores the utility of Bayesian methods for a more comprehensive understanding of brain connectivity alterations in GAD.