A brain-region-based meta-analysis method utilizing the Apriori algorithm

Area of Science:

• Neuroscience
• Cognitive Science
• Data Mining

Background:

• Brain network connectivity modeling is essential for understanding cognitive functions.
• Meta-analyses aggregate findings from individual studies for reliable results.
• Existing meta-analytic connectivity modeling often relies on predefined regions of interest (ROIs).

Purpose of the Study:

• To propose a new meta-analysis method for deriving brain network connectivity models.
• To enable the identification of connectivity models directly from literature activation data, without requiring ROIs.
• To leverage the Apriori algorithm for mining co-activation patterns.

Main Methods:

• Extract activation information from experimental studies using similar cognitive tasks.
• Map activation data to brain areas using an automatic anatomical labeling atlas.
• Calculate brain area activation rates and apply the Apriori algorithm to derive connectivity models.

Main Results:

• Successfully applied the method to a language review article's citations.
• The derived network connectivity model showed consistency with previously reported findings.
• Demonstrated the potential to mine brain network connectivity from published activation data.

Conclusions:

• The proposed method effectively identifies brain network connectivity by analyzing co-activation relationships.
• Results can serve as prior knowledge for dynamic causal modeling, potentially reducing dimensionality and increasing efficiency.
• Offers a novel, ROI-independent approach to brain connectivity meta-analysis.