Replicability of time-varying connectivity patterns in large resting state fMRI samples
Anees Abrol1, Eswar Damaraju1, Robyn L Miller2
1The Mind Research Network, Albuquerque, NM, USA; Department of Electrical and Computer Engineering, University of New Mexico, Albuquerque, NM, USA.
Neuroimage
|September 17, 2017
Summary
This study confirms that dynamic functional network connectivity (dFNC) patterns in the human brain are reproducible across different analysis methods and datasets. These findings support the reliability of the chronnectome approach for understanding brain dynamics.
Area of Science:
- Neuroscience
- Cognitive Science
- Data Science
Background:
- Recent advancements utilize temporal changes in brain connectivity (chronnectome) for analysis.
- Dynamic functional network connectivity (dFNC) frameworks are emerging to study brain activity over time.
Purpose of the Study:
- To investigate the replicability of human brain inter-regional coupling dynamics during rest.
- To evaluate two distinct dFNC analysis frameworks using a large fMRI dataset.
- To quantify the reproducibility of emergent functional connectivity (FC) patterns.
Main Methods:
- Utilized 7,500 functional magnetic resonance imaging (fMRI) datasets.
- Characterized temporal dynamics by deriving summary measures of FC patterns.
- Applied analysis frameworks to large, independent, age-matched samples.
- Evaluated reproducibility using statistically stationary, linear, and Gaussian surrogate datasets.
Main Results:
- Demonstrated reproducibility through the existence of basic connectivity patterns (FC states).
- Found that some state summary measures were statistically significant.
- Surrogate data analysis indicated that null models did not fully explain the fMRI data.
- Estimated FC states showed robustness against variations in data quality, analysis, and methods.
Conclusions:
- The chronnectome approach and dFNC methods offer reproducible insights into brain connectivity dynamics.
- Estimated FC states are robust across various analytical and data quality factors.
- Future research should prioritize investigating the functional and neurophysiological relevance of time-varying connectivity.
