Voxel-wise intermodal coupling analysis of two or more modalities using local covariance decomposition

Area of Science:

• Neuroimaging
• Computational Neuroscience
• Biostatistics

Background:

• Multi-modal brain imaging captures information within and between modalities.
• Intermodal coupling (IMCo) quantifies voxel-wise covariation between modalities.
• Existing IMCo methods are limited to two modalities and lack symmetry.

Purpose of the Study:

• To generalize IMCo estimation for two or more modalities using a symmetric, voxel-wise coupling coefficient.
• To introduce principal-component-based IMCo (pIMCo) for analyzing multi-modal brain data.
• To investigate neurodevelopmental patterns in coupling between cerebral blood flow, amplitude of low frequency fluctuations, and local connectivity.

Main Methods:

• Developed a generalized IMCo estimation using local covariance decompositions.
• Defined a symmetric, voxel-wise coupling coefficient applicable to multiple modalities.
• Applied the pIMCo method to multi-modal neuroimaging data from 803 subjects (ages 8-22).

Main Results:

• Demonstrated that pIMCo reveals spatially heterogeneous coupling patterns.
• Showed that coupling varies with age and sex during neurodevelopment.
• Identified novel patterns in multi-modal brain data not apparent in individual modalities.

Conclusions:

• pIMCo offers a powerful, symmetric approach for summarizing relationships across multiple brain imaging modalities.
• The method is valuable for analyzing the increasing availability of multi-modal brain data.
• pIMCo can uncover complex neurodevelopmental patterns and disease-related alterations.