White-Matter BOLD Encoding Beyond Marginal Connectivity

Area of Science:

• Neuroimaging
• Functional MRI Analysis
• White Matter Neuroscience

Background:

• Traditionally, functional MRI (fMRI) research prioritized gray matter, often dismissing white matter BOLD signals as noise.
• Emerging evidence indicates white matter BOLD fluctuations hold significant functional information.
• Existing gray-to-white matter analyses using marginal functional connectivity struggle to differentiate direct coupling from shared variance within complex neural systems.

Purpose of the Study:

• To investigate if spontaneous white matter BOLD activity can be predicted by distributed gray matter activity using a multivariate cortical encoding framework.
• To determine if this predictive structure offers insights into white matter organization beyond conventional marginal connectivity measures.
• To explore the functional organization of white matter revealed by predicting its spontaneous activity from cortical networks.

Main Methods:

• Utilized resting-state fMRI data from 81 participants in the Human Connectome Project.
• Employed a strict white matter mask, ensuring no overlap with cortical predictors.
• Applied nested leave-one-run-out ridge regression, using time series from 400 cortical parcels to predict white matter BOLD signals in held-out data.

Main Results:

• Cortical activity demonstrated a modest yet reliable ability to predict white matter BOLD dynamics across extensive white matter regions.
• Ridge regression 'fingerprints' largely mirrored marginal functional connectivity, suggesting a common functional foundation.
• A reproducible divergence between the predictive structure and marginal connectivity was identified, organizing prediction residuals and linked to specific white matter tracts.

Conclusions:

• Multivariate cortical encoding successfully predicts white matter BOLD signals from gray matter activity.
• This approach uncovers a tract-organized dimension of white matter functional coupling previously uncaptured by pairwise connectivity.
• The findings highlight the functional relevance of white matter BOLD signals and offer a novel method for their analysis.