Effects of spatial smoothing on functional brain networks.

Area of Science:

• Neuroimaging
• Network Neuroscience
• Computational Neuroscience

Background:

• Graph-theoretical methods are standard for analyzing human brain structure and function.
• Constructing functional brain networks from functional magnetic resonance imaging (fMRI) data involves preprocessing steps that can influence network properties.
• Systematic studies on the impact of preprocessing, particularly spatial smoothing, on fMRI-derived brain networks are lacking.

Purpose of the Study:

• To investigate the effects of spatial smoothing, a common fMRI preprocessing technique, on the structure of functional brain networks.
• To quantify how different levels of spatial smoothing alter network metrics and topology.

Main Methods:

• Applied varying levels of spatial smoothing to resting-state fMRI data.
• Analyzed changes in functional network properties, including node centrality measures and network component composition.
• Examined the impact on network similarity across subjects.

Main Results:

• The degree of spatial smoothing systematically and non-uniformly affects node centrality measures in functional brain networks, with effects dependent on brain geometry.
• Spatial smoothing influences the composition of the largest connected network component.
• Smoothing artificially increases the similarity between functional brain networks of different individuals.

Conclusions:

• Spatial smoothing significantly impacts the topological properties of functional brain networks derived from fMRI data.
• The observed changes can lead to artefactual increases in inter-subject network similarity.
• Avoidance of spatial smoothing is recommended during fMRI data preprocessing for network analysis whenever feasible.