Effect of Spatial Smoothing on Task fMRI ICA and Functional Connectivity

Area of Science:

• Neuroimaging
• Data Analysis
• Brain Connectivity

Background:

• Spatial smoothing is a standard preprocessing step in functional magnetic resonance imaging (fMRI) data analysis.
• Understanding its effects on task-evoked functional mapping and connectivity is crucial for accurate interpretation.
• Independent Component Analysis (ICA) is a common technique for decomposing fMRI data into distinct brain networks.

Purpose of the Study:

• To investigate the impact of varying spatial smoothing levels on task-evoked fMRI brain functional mapping and connectivity.
• To compare the effects of spatial smoothing on single-subject versus multi-subject ICA analyses.
• To provide recommendations for optimal spatial smoothing parameters in fMRI studies.

Main Methods:

• Task-fMRI data were decomposed using Independent Component Analysis (ICA).
• Task-modulated ICA components were identified using the task paradigm.
• Spatial smoothing was applied with Gaussian kernels of varying Full Width at Half Maximum (FWHM) (1-35 mm).
• Task activation volume and inter-component functional connectivity (FC) were measured.

Main Results:

• Spatial smoothing decreased task extraction performance in single-subject ICA more than multi-subject ICA.
• Spatial smoothing increased task volume and strengthened functional connectivity in single-subject ICA.
• The positive-negative imbalance of single-subject ICA was more affected by smoothing than multi-subject ICA.

Conclusions:

• A spatial smoothing of 2-3 voxel FWHM is suggested for single-subject ICA to balance functional connectivity.
• A broader range of 2-5 voxel FWHM is recommended for multi-subject ICA.
• Spatial smoothing significantly influences fMRI analysis outcomes, necessitating careful parameter selection.