Integrating Multimodal Neuroimaging and Genetics: A Structurally-Linked Sparse Canonical Correlation Analysis...

Area of Science:

• Neuroscience
• Genetics
• Biostatistics

Background:

• Neuroimaging genetics seeks to understand the relationship between genetic variations and brain imaging data.
• Existing methods like Sparse Canonical Correlation Analysis (SCCA) are limited in handling multiple imaging modalities and identifying structurally linked features.

Purpose of the Study:

• To develop an enhanced SCCA framework for integrating multiple neuroimaging modalities (T1-weighted MRI, fMRI, diffusion tensor imaging) and genetic data (SNPs).
• To simultaneously identify structurally linked brain regions and their genetic underpinnings.

Main Methods:

• A novel structurally linked SCCA framework was developed.
• Incorporated diffusion tensor imaging to identify structurally linked brain regions.
• Leveraged T1-weighted MRI, fMRI, and SNP data for comprehensive analysis.

Main Results:

• The proposed method demonstrated superior performance on simulated data compared to existing approaches.
• Identified interpretable gene-imaging associations in large-scale Human Connectome Project (HCP) data.
• Successfully delineated both structural and functional brain characteristics.

Conclusions:

• The developed methodology provides a robust framework for exploring the genetic basis of brain structure and function.
• Offers novel insights into gene-brain interactions for the field of neuroscience.
• The code is publicly available for further research.