Investigation of the Effect of Physiological Artifacts on Task-based Functional Connectivity: A Simulation Study

Area of Science:

• Neuroimaging
• Cognitive Neuroscience
• Biomedical Engineering

Background:

• Functional connectivity analysis reveals brain region interactions using blood-oxygen-level-dependent (BOLD) functional magnetic resonance imaging (fMRI).
• Physiological artifacts and noise significantly contaminate BOLD fMRI signals, potentially compromising the accuracy of functional connectivity estimations.
• Understanding and mitigating these artifacts is crucial for reliable interpretation of brain network dynamics.

Purpose of the Study:

• To quantitatively assess the impact of physiological artifacts on functional connectivity patterns derived from BOLD fMRI signals.
• To compare the performance of different artifact mitigation strategies, including resting-state versus task-based analysis and full versus partial correlation methods.

Main Methods:

• Simulated physiological and BOLD fMRI signals under resting and task conditions.
• Quantification of functional connectivity patterns using established methods: full correlation and partial correlation.
• Analysis of artifact-induced errors in functional connectivity estimations.

Main Results:

• Physiological artifacts disproportionately affected brain regions with similar physiological response functions.
• Functional connectivity computed during task execution showed lower error rates compared to resting-state computations.
• Partial correlation analysis consistently demonstrated lower errors than full correlation analysis in the presence of artifacts.

Conclusions:

• Physiological artifacts introduce significant distortions in BOLD fMRI-based functional connectivity.
• Task-based fMRI analysis and the use of partial correlation are effective strategies for reducing artifact-related errors.
• This study provides quantitative insights into artifact impact, guiding the selection of robust methods for neuroimaging analysis.