Related Experiment Video
Updated: Dec 4, 2025
08:51
Statistical Modelling of Cortical Connectivity Using Non-invasive Electroencephalograms
Published on: November 1, 2019
5.9K
Characterising group-level brain connectivity: A framework using Bayesian exponential random graph models
B C L Lehmann1, R N Henson2, L Geerligs3
1MRC Biostatistics Unit, University of Cambridge, UK; Big Data Institute, University of Oxford, UK; Department of Statistics, University of Oxford, UK.
Neuroimage
|October 25, 2020
Summary
This study introduces a new Bayesian framework to analyze brain connectivity networks across groups. The method effectively characterizes and compares functional brain networks in young versus old individuals, advancing network neuroscience.
Area of Science:
- Neuroscience
- Network Science
- Statistical Modeling
Background:
- The brain is modeled as a network where nodes represent brain regions and edges represent interactions.
- Analyzing group-level brain connectivity requires understanding common network features while accounting for individual variability.
- Existing methods either oversimplify by creating a group-representative network (GRN) or miss shared information by analyzing individuals independently.
Purpose of the Study:
- To develop a novel Bayesian framework for characterizing the distribution of an entire population of brain networks.
- To extend exponential random graph models (ERGM) to handle multiple networks simultaneously.
- To apply the framework to compare functional brain connectivity structures between young and old individuals.
Main Methods:
- Utilized a Bayesian framework based on exponential random graph models (ERGM) extended for multiple networks.
- Applied the method to resting-state functional magnetic resonance imaging (fMRI) data from the Cam-CAN project.
- Compared functional connectivity structures between a group of young and a group of old healthy adults.
Main Results:
- The proposed Bayesian framework successfully characterized the distribution of brain networks within the studied population.
- The method allowed for reliable characterization and comparison of functional connectivity structures across age groups.
- Demonstrated the framework's utility in identifying age-related differences in brain network properties.
Conclusions:
- The developed Bayesian multi-network ERGM framework provides a robust approach to analyzing group-level brain connectivity.
- This method effectively captures both commonalities and individual variations in network structure across populations.
- The findings highlight the potential of this framework for studying brain network changes in aging and other conditions.
