A unified framework for multimodal structure-function mapping based on eigenmodes.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Network Science

Background:

• The relationship between brain structure and function is crucial for understanding behavior.
• White matter network structure significantly influences functional connectivity.
• Existing methods for mapping structure to function are fragmented and lack comparative analysis.

Purpose of the Study:

• To develop a unified computational framework for structure-function mapping in the brain.
• To generalize and compare existing methods based on eigendecompositions.
• To identify limitations and propose future directions for improving brain structure-function predictions.

Main Methods:

• Developed a generalized computational framework based on eigenmodes to unify structure-function mapping models.
• Applied the framework to 50 subjects from the Human Connectome Project.
• Systematically compared existing and newly proposed models within the unified framework.

Main Results:

• Successfully reproduced six previously published structure-function mapping results.
• Introduced two novel models within the generalized framework.
• Provided a direct, comparative analysis of various eigendecomposition-based mapping methods.
• Identified a performance ceiling for current eigenmode-based mapping approaches.

Conclusions:

• The unified framework effectively generalizes and compares existing structure-function mapping methods.
• Current eigenmode-based approaches may have reached their performance limit.
• Further research is needed to explore alternative methods to overcome performance limitations in predicting functional connectivity from structural networks.