Connectome caricatures remove large-amplitude coactivation patterns in resting-state fMRI to emphasize individual...

Area of Science:

• Neuroscience
• Cognitive Science
• Data Science

Background:

• Resting-state functional magnetic resonance imaging (fMRI) captures brain activity during rest.
• High-amplitude coactivation patterns, though sparse, significantly influence resting-state functional connectivity and resemble task-based activation.
• The majority of the resting-state signal remains under-characterized.

Purpose of the Study:

• To develop and validate a novel method, "caricaturing," for analyzing the under-explored portion of resting-state fMRI signals.
• To investigate whether this novel approach can reveal meaningful individual differences beyond dominant coactivation patterns.

Main Methods:

• Caricaturing projects resting-state fMRI data onto a subspace orthogonal to a manifold of task coactivation patterns.
• This method effectively removes linear combinations of dominant coactivation patterns, creating "caricatured connectomes."
• Task fMRI data from two large-scale neuroimaging datasets were used to construct the coactivation manifold.

Main Results:

• Caricatured connectomes demonstrated lower between-individual similarity compared to standard connectomes.
• These caricatured connectomes showed higher identifiability at the individual level.
• The method enabled prediction of phenotypic measures, reflecting individual behavioral differences, often more effectively than standard connectomes.

Conclusions:

• A significant and useful neural signal exists beyond the dominant coactivation patterns in resting-state fMRI.
• Caricaturing offers a novel approach to characterizing the brain's intrinsic functional architecture.
• This method may provide a more sensitive tool for understanding individual differences in brain function.