Related Experiment Video
Updated: Sep 30, 2025
06:02
Topographical Estimation of Visual Population Receptive Fields by fMRI
Published on: February 3, 2015
9.4K
Real-time and Recursive Estimators for Functional MRI Quality Assessment
Nikita Davydov1,2,3, Lucas Peek4, Tibor Auer5
1Aligned Research Group, Los Gatos, USA.
Neuroinformatics
|March 17, 2022
Summary
This study introduces a new real-time quality assessment (rtQA) for functional MRI (fMRI) to automatically detect artifacts. This method enhances data reliability for neuroimaging research and clinical use.
Area of Science:
- Neuroimaging
- Biomedical Engineering
- Data Science
Background:
- Real-time quality assessment (rtQA) is crucial for functional magnetic resonance imaging (fMRI) to ensure data integrity.
- Technical and physiological noise significantly degrade blood oxygen level-dependent (BOLD) sensitivity, leading to fMRI artifacts.
- Subjective visual inspection of fMRI data during acquisition is insufficient for detecting subtle distortions.
Purpose of the Study:
- To develop and validate a comprehensive, automated real-time quality assessment (rtQA) system for fMRI data.
- To enable rapid identification and mitigation of fMRI artifacts during data acquisition.
- To improve the robustness and reliability of fMRI studies and neurofeedback applications.
Main Methods:
- Applied real-time and recursive methods to assess whole-brain fMRI volumes and time-series.
- Estimated recursive temporal signal-to-noise ratio (rtSNR) and contrast-to-noise ratio (rtCNR).
- Calculated real-time head motion parameters, framewise displacement (FD), micro-displacement (MD), and derivative of root mean squared variance over voxels (DVARS).
- Utilized a modified Kalman filter to detect spikes and filtered noise in target regions and networks.
- Implemented incremental general linear modeling (GLM) to assess nuisance regressor contributions.
Main Results:
- Demonstrated the proposed rtQA system in real-time fMRI neurofeedback and resting-state simulations, including scenarios with excessive head motion.
- The rtQA system was successfully implemented as an extension of the OpenNFT software and a unified Python library.
- Flexible estimation and visualization of rtQA metrics facilitated efficient data quality control.
Conclusions:
- The developed automated rtQA system provides efficient and reliable assessment of fMRI data quality in real-time.
- This tool supports informed decisions regarding experiment interruption or restart, enhancing fMRI acquisition robustness.
- Increased confidence in neural estimates is achieved through improved fMRI data quality and artifact detection.
