Model-based whole-brain effective connectivity to study distributed cognition in health and disease

Area of Science:

• Neuroimaging
• Cognitive Neuroscience
• Computational Neuroscience

Background:

• Neuroimaging techniques are crucial for studying human cognition and brain networks.
• Functional connectivity is a common proxy for brain activity distribution.
• Limitations in existing connectivity measures necessitate dynamic models for directional interaction assessment.

Purpose of the Study:

• To present a model-based whole-brain effective connectivity framework for fMRI data analysis.
• To discuss the advantages and disadvantages of the proposed approach compared to existing methods.
• To explore the application of dynamic models in understanding brain coordination and cognition.

Main Methods:

• Utilized the multivariate Ornstein-Uhlenbeck (MOU) process for dynamic modeling.
• Developed a framework referred to as MOU-EC for analyzing fMRI data.
• Applied machine-learning tools to address challenges in whole-brain analysis.

Main Results:

• MOU-EC provides directed connectivity estimates reflecting BOLD activity dynamics.
• The approach can extract biomarkers for task-specific brain coordination.
• Changes in MOU-EC connections can be interpreted collectively and model-based, linking to network analysis.

Conclusions:

• The MOU-EC framework offers a comprehensive tool for studying distributed cognition.
• This model-based approach provides new perspectives for analyzing brain network dynamics.
• It holds potential for exploring neuropathologies through brain connectivity analysis.