On the importance of modeling fMRI transients when estimating effective connectivity: A dynamic causal modeling study...

Area of Science:

• Neuroimaging
• Systems Neuroscience
• Computational Neuroscience

Background:

• Effective connectivity is vital for understanding brain function.
• Current models often struggle with accurately capturing neuronal and vascular signals.
• Physiologically informed dynamic causal models (P-DCM) offer potential improvements.

Purpose of the Study:

• To evaluate a novel P-DCM against previous dynamic causal models (DCMs) for estimating effective connectivity.
• To assess the impact of modeling hemodynamic response transients on connectivity estimates.
• To investigate the benefits of using arterial spin labeling (ASL) fMRI data (BOLD and CBF) with P-DCM.

Main Methods:

• Comparison of P-DCM with traditional DCM variants using fMRI data from a visuo-motor task.
• Analysis of effective connectivity in a five-region network.
• Application of P-DCM to both BOLD-only data and combined BOLD-CBF data from ASL fMRI.

Main Results:

• Connectivity estimates are sensitive to the specific DCM used, particularly in modeling hemodynamic transients like post-stimulus undershoot.
• P-DCM demonstrates superior estimation of effective connectivity compared to previous DCMs.
• Integrating CBF data with BOLD signals using P-DCM enhances the ability to differentiate neuronal and vascular effects.

Conclusions:

• Accurate modeling of fMRI response transients is essential for reliable effective connectivity estimation.
• P-DCM, incorporating neuronal, neurovascular, and hemodynamic components, represents a significant advancement.
• Additional hemodynamic data, such as from ASL, improves the disambiguation of BOLD signal components.