A comparison of statistical methods for detecting context-modulated functional connectivity in fMRI

Area of Science:

• Cognitive Neuroscience
• Clinical Neuroscience
• Neuroimaging

Background:

• Functional connectivity (FC) analysis in fMRI investigates how contextual factors modulate cognitive processes.
• Context-modulated FC is often tested by comparing connectivity between regions A and B under task conditions X and Y.
• Existing methods include beta-series and psychophysiological interaction (PPI) analyses, but their comparative power is unclear.

Purpose of the Study:

• To compare the statistical power of generalized PPI (gPPI), standard PPI (sPPI), and beta-series methods for detecting context-modulated FC.
• To evaluate method performance under various experimental parameters, such as trial number and duration.
• To assess the sensitivity of these methods in a real-world fMRI dataset.

Main Methods:

• Simulations were used to compare the power of sPPI, gPPI, and beta-series methods.
• Analyses were conducted across different experimental designs (block vs. event-related) and parameters.
• A real fMRI dataset from adolescent girls was analyzed to validate simulation findings.

Main Results:

• gPPI and beta-series methods generally showed higher power than sPPI.
• gPPI was more powerful for block designs, while beta-series excelled with more trial repetitions and hemodynamic response function variability.
• In a real dataset, PPI methods showed greater sensitivity in block designs, and beta-series in event-related designs with many trials.

Conclusions:

• The choice between gPPI and beta-series methods for context-modulated FC analysis depends on specific experimental design parameters.
• Understanding these performance differences is crucial for optimizing fMRI study design and interpretation.
• These findings inform researchers on selecting the most powerful statistical approach for their context-modulated FC research questions.