Related Experiment Video
Updated: Jul 11, 2025
08:51
Statistical Modelling of Cortical Connectivity Using Non-invasive Electroencephalograms
Published on: November 1, 2019
5.7K
Improved cerebrovascular reactivity mapping using coherence weighted general linear model in the frequency domain
Botian Xu1, Chau Vu1, Matthew Borzage2
1Department of Biomedical Engineering, University of Southern California, Los Angeles, CA, United States; Department of Pediatrics and Radiology, Children's Hospital Los Angeles, Los Angeles, CA, United States.
Neuroimage
|November 12, 2023
Summary
A new coherence weighted general linear model (CW-GLM) improves cerebrovascular reactivity (CVR) estimation using end-tidal CO2 and MRI signals. This method enhances accuracy and simplifies experiments by reducing noise and eliminating the need for signal alignment.
Area of Science:
- Neuroimaging
- Medical Physics
- Biomedical Engineering
Background:
- Cerebrovascular reactivity (CVR) is a key indicator of cerebrovascular health.
- Estimating CVR from endogenous end-tidal carbon dioxide (CO2) and MRI signals is challenging due to low signal-to-noise ratio (SNR).
- Existing methods like time-domain general linear model (TD-GLM) and frequency-domain general linear model (FD-GLM) have limitations in accuracy and experimental simplicity.
Purpose of the Study:
- To develop an improved method for estimating CVR from end-tidal CO2 and MRI signals.
- To enhance the accuracy and robustness of CVR estimation, particularly under resting-state conditions.
- To simplify the experimental requirements for CVR assessment.
Main Methods:
- Proposed a novel coherence weighted general linear model (CW-GLM) for CVR estimation.
- CW-GLM utilizes Fourier coefficients weighted by signal coherence in the frequency domain.
- Compared CW-GLM performance against TD-GLM and FD-GLM using synthetic and in-vivo data under resting and sinusoidal CO2 stimulus conditions.
Main Results:
- CW-GLM demonstrated superior performance in synthetic data, with lower mean-absolute errors for gray matter (0.7%) and white matter (1.2%) compared to other methods.
- For resting-state data, CW-GLM achieved mean-absolute errors of 4.1% (gray matter) and 8% (white matter).
- In-vivo measurements showed CW-GLM improved the limits of agreement between resting and sinusoidal CO2 paradigms by 12%-209% compared to time-domain methods.
Conclusions:
- The CW-GLM method effectively enhances CVR estimation robustness by acting as a self-adaptive band-pass filter.
- CW-GLM significantly improves accuracy and reduces noise in CVR measurements.
- The method eliminates the need for signal temporal alignment, simplifying experimental procedures for CVR assessment.
