Surface-based analysis increases the specificity of cortical activation patterns and connectivity results

Area of Science:

• Neuroimaging
• Neuroscience
• Medical Imaging Analysis

Background:

• Spatial smoothing is a common preprocessing step in functional magnetic resonance imaging (fMRI) analysis.
• Current standard practice often uses volume-based smoothing, which may lead to signal contamination between adjacent cortical areas.
• Surface-based approaches (SBA) offer a theoretical advantage for separating signals on the unfolded cortex.

Purpose of the Study:

• To evaluate the benefits of contemporary surface-based smoothing (SBA) compared to volume-based smoothing in fMRI data.
• To assess the impact of smoothing methods on signal contamination between the primary motor cortex (M1) and primary somatosensory cortex (S1).
• To determine how signal contamination affects activity and connectivity analyses in these regions.

Main Methods:

• fMRI data acquired from 19 subjects during a tactile stimulation task.
• Application of both volume-based smoothing and surface-based smoothing (SBA) techniques.
• Analysis of simulated Blood-Oxygen-Level-Dependent (BOLD) responses to assess signal contamination and its effects.

Main Results:

• Volume-based smoothing resulted in significant signal contamination of M1 by S1 responses, leading to false positive motor activation.
• Surface-based smoothing (SBA) with appropriate kernel sizes effectively mitigated these false positive activations.
• Volume-based smoothing exaggerated connectivity estimates between M1 and S1, an effect not observed with SBA.

Conclusions:

• Surface-based smoothing (SBA) considerably reduces signal contamination between neighboring functional brain regions.
• SBA enhances the validity and reliability of fMRI-based activity and connectivity analyses.
• The findings advocate for the adoption of surface-based smoothing in routine neuroimaging preprocessing pipelines.