Hybrid ICA-Seed-Based Methods for fMRI Functional Connectivity Assessment: A Feasibility Study

Area of Science:

• Neuroimaging
• Cognitive Neuroscience
• Data Analysis

Background:

• Functional connectivity (FC) analysis of fMRI data is crucial for understanding brain networks.
• Traditional seed-based methods and Independent Component Analysis (ICA) have limitations in reproducibility.
• ICA is data-driven but group inferences from its maps can be challenging due to reproducibility issues.

Purpose of the Study:

• To introduce and evaluate hybrid ICA-seed-based FC methods.
• To address reproducibility issues in group-level FC analysis.
• To compare the performance of hybrid methods against standard approaches.

Main Methods:

• Developed five hybrid ICA-seed-based FC methods.
• Compared hybrid methods with standard regression against a priori time courses and group ICA back-reconstruction.
• Utilized ROI-based (single, few, many-voxel seeds) and dual-regression-based (single, multiple ICA map seeds) approaches.

Main Results:

• Demonstrated potential advantages of hybrid ICA-seed-based FC methods across five experiments.
• Hybrid methods offer a viable solution to circumvent reproducibility issues in ICA-derived FC maps.
• Results suggest improved reliability for group inferences in brain connectivity.

Conclusions:

• Hybrid ICA-seed-based FC methods enhance the reliability of brain connectivity analysis.
• These methods provide a robust alternative for researchers investigating functional brain networks.
• The study highlights the potential of integrating ICA with seed-based approaches for reproducible neuroimaging research.