State-transition dynamics of resting-state functional magnetic resonance imaging data: model comparison and
Saiful Islam1, Pitambar Khanra2, Johan Nakuci3
1Institute for Artificial Intelligence and Data Science, University at Buffalo, State University of New York at Buffalo, 215 Lockwood Hall, Buffalo, 14260, NY, USA.
BMC Neuroscience
|March 4, 2024
Summary
We adapted electroencephalogram (EEG) microstate analysis for functional magnetic resonance imaging (fMRI) data. This method reveals reliable brain state dynamics unique to individuals, showing potential for brain fingerprinting.
Area of Science:
- Neuroscience
- Computational Neuroscience
- Brain Imaging Analysis
Background:
- Electroencephalogram (EEG) microstate analysis identifies quasi-stable brain states from time series data.
- Application of microstate analysis to functional magnetic resonance imaging (fMRI) data is limited due to fMRI's slower temporal resolution.
- Existing methods for EEG microstate analysis need adaptation for fMRI data.
Purpose of the Study:
- To extend EEG microstate analysis techniques to resting-state fMRI data.
- To investigate the state-transition dynamics of brain activity in humans using fMRI.
- To assess the reliability and individual discriminability of fMRI-derived brain state dynamics.
Main Methods:
- Applied data clustering methods from EEG microstate analysis to resting-state fMRI data.
- Developed a novel method to evaluate the test-retest reliability of fMRI state-transition dynamics.
- Compared within-participant and between-participant reliability of identified brain states.
Main Results:
- Clustering quality for fMRI data was comparable to that achieved with EEG microstate analysis.
- Within-participant test-retest reliability of state-transition dynamics was higher than between-participant reliability.
- The analysis demonstrated potential for discriminating individuals based on their brain state dynamics.
Conclusions:
- EEG microstate analysis methods can be successfully adapted for fMRI data.
- fMRI state-transition dynamics exhibit high within-individual reliability, suggesting potential for brain fingerprinting.
- This approach offers a promising tool for identifying unique individual brain dynamics from fMRI data.
