Independence-based causal discovery analysis reveals statistically non-significant regions to be functionally...

Area of Science:

• Neuroimaging
• Computational Neuroscience
• Psychiatry

Background:

• Traditional fMRI analysis often overlooks statistically non-significant brain regions, assuming biological insignificance.
• This study challenges this assumption by investigating causal interactions between significant (active network, AN) and non-significant (silent network, SN) brain regions.

Purpose of the Study:

• To test the causal interactions between AN and SN.
• To determine if these interactions influence psychopathology severity and working memory performance.

Main Methods:

• Examined AN and SN during the N-BACK task in 25 individuals with familial risk for psychosis (FHR) and 37 controls.
• Utilized the PC algorithm for causal discovery and analyzed connectivity of regions with highest alpha-centrality (HAC).

Main Results:

• Identified causal connectivity between SN and AN, indicating mutual influence.
• Found that specific HAC regions in both groups formed reciprocal circuits that causally increased magical ideation severity.

Conclusions:

• Statistically non-significant brain regions causally interact with significant regions, suggesting they are not biologically unimportant.
• Findings question the exclusive inclusion of significant regions in pathophysiological models and highlight the importance of causality analysis.