Related Experiment Video
Updated: Aug 3, 2026
10:06
High-resolution Functional Magnetic Resonance Imaging Methods for Human Midbrain
Published on: May 10, 2012
A quantitative comparison of motion detection algorithms in fMRI
B A Ardekani1, A H Bachman, J A Helpern
1Center for Advanced Brain Imaging, Nathan Kline Institute, 140 Old Orangeburg Road, Orangeburg, NY 10962, USA. ardekani@nki.rfmh.org
Magnetic Resonance Imaging
|October 12, 2001
Summary
Accurate fMRI motion detection is crucial. SPM99 excelled in accuracy, while AFNI98 offered a balanced speed and accuracy, proving robust against noise for reliable fMRI analysis.
Area of Science:
- Neuroimaging
- Biomedical Engineering
- Data Analysis
Background:
- Subject motion is a significant challenge in functional Magnetic Resonance Imaging (fMRI) time-series analysis.
- Accurate motion detection and correction are vital for reliable fMRI data interpretation.
- Several open-source algorithms exist for addressing motion artifacts in fMRI data.
Purpose of the Study:
- To compare the performance of four widely used fMRI motion detection algorithms: AIR 3.08, SPM99, AFNI98, and the Thévenaz, Ruttimann, and Unser (TRU) pyramid method.
- To evaluate algorithm efficacy in correcting simulated motions across varying noise levels.
- To identify the most accurate, fastest, and robust motion correction tool for fMRI analysis.
Main Methods:
- Simulated fMRI data with known motion parameters were generated.
- Four distinct motion correction algorithms (AIR 3.08, SPM99, AFNI98, TRU) were applied to the simulated data.
- Algorithm performance was assessed based on accuracy in motion parameter estimation and robustness to noise.
Main Results:
- SPM99 demonstrated the highest accuracy in motion detection among the evaluated algorithms.
- AFNI98 offered comparable accuracy to SPM99 but was significantly faster and demonstrated superior robustness in noisy conditions.
- TRU's performance was comparable to SPM99 and AFNI98 for minor misalignments but degraded substantially with larger ones.
- AIR was found to be the least accurate algorithm in this comparison.
Conclusions:
- AFNI98 presents a favorable balance of speed and accuracy, making it a practical choice for fMRI motion correction.
- SPM99 is the most accurate but may be slower, while AFNI98 offers robustness and speed, especially in low signal-to-noise ratio environments.
- The choice of algorithm depends on specific research needs, balancing accuracy, speed, and tolerance to noise in fMRI data processing.
