A probabilistic framework to infer brain functional connectivity from anatomical connections.

Area of Science:

• Neuroscience
• Computational Biology
• Statistical Modeling

Background:

• Brain function relies on complex interactions between structure and activity.
• Predicting functional brain activity from anatomical connectivity is a challenging problem.
• Existing methods often fail to account for the inherent correlations in functional connectivity data.

Purpose of the Study:

• To develop a novel probabilistic framework for mapping brain anatomical connectivity to functional connectivity.
• To establish a rigorous statistical basis for understanding the relationship between brain structure and activity.
• To introduce a robust model selection approach for structured-output learning in neuroimaging.

Main Methods:

• Formulated the prediction problem as a structured-output learning task.
• Introduced a cross-validation framework with a parametrization-independent loss function for covariance matrices.
• Constrained functional activity's conditional independence structure using anatomical connectivity.
• Employed a stationary multivariate autoregressive model for functional connectivity prediction.

Main Results:

• Demonstrated that anatomical connectivity can predict functional connectivity across multiple subjects.
• Showcased the effectiveness of the proposed probabilistic framework in modeling brain activity.
• Validated the importance of accurate functional connectivity modeling for assessing structure-function links.

Conclusions:

• Functional brain connectivity can be statistically explained by anatomical connectivity.
• The developed framework provides a rigorous method for linking brain structure and function.
• Accurate modeling of functional connectivity is essential for understanding its relationship with anatomical connectivity.