Statistical Perspective on Functional and Causal Neural Connectomics: A Comparative Study.

Area of Science:

• Neuroscience
• Statistics
• Computational Biology

Background:

• Understanding brain network interactions is crucial for linking neural structure to function.
• Current functional connectomics methods often rely on statistical associations, which do not guarantee causality.
• Existing causal approaches for functional connectomics are limited in scope.

Purpose of the Study:

• To develop a systematic statistical framework for causal functional connectomics.
• To define foundational aspects of causality applicable to neural interaction analysis.
• To guide the development and comparison of causal connectomics methodologies.

Main Methods:

• Consolidation of notions of association and representation in neural interaction.
• Description of causal modeling from statistical literature.
• Introduction of directed Markov graphical models and the Directed Markov Property.

Main Results:

• Establishment of a framework for defining and examining causality in functional connectomes.
• Demonstration of a comparative study of existing causal functional connectivity approaches.
• Identification of properties for future causal connectomics methodologies.

Conclusions:

• The proposed framework provides a systematic approach to causal functional connectomics.
• It enables a clearer understanding and comparison of different causal inference methods.
• This work lays the groundwork for more comprehensive causal analyses of brain networks.