Reducing individual differences in task fMRI with OGRE (One-step General Registration and Extraction) preprocessing

Mark M McAvoy,Lei Liu,Ruiwen Zhou,Benjamin A Philip
DOI: https://doi.org/10.1101/2023.09.19.558290
2024-06-21
Abstract:Volumetric analysis methods continue to enjoy great popularity in the analysis of task-related functional MRI (fMRI) data. Among these methods, the software package FSL (FMRIB, Oxford, UK) is omnipresent throughout the field. However, it remains unknown what advantages might be gained by integrating FSL with alternative preprocessing tools. Here we developed the One-step General Registration and Extraction (OGRE) pipeline to apply FreeSurfer brain extraction for simultaneous registration and motion correction ("one-step resampling"), for FSL volumetric analysis of fMRI data. We compared three preprocessing approaches (OGRE, FSL, and fMRIPrep) in a dataset wherein adult human volunteers (N=39) performed a precision drawing task during fMRI scanning. The three approaches produced grossly similar results, but OGRE's preprocessing led to lower inter-individual variability across the brain and greater detected activation in primary motor cortex contralateral to movement. This demonstrates that FreeSurfer tools and one-step resampling can improve FSL's volumetric analysis of fMRI data. The OGRE pipeline provides an off-the-shelf method to apply FreeSurfer-based brain extraction and one-step resampling of motion correction and registration for FSL analysis of task fMRI data.
Neuroscience
What problem does this paper attempt to address?
The paper attempts to address the issue of reducing individual differences in task functional magnetic resonance imaging (task fMRI) data analysis. Specifically, the research team developed a new preprocessing pipeline called OGRE (One-step General Registration and Extraction), which aims to improve the volumetric analysis workflow of FSL (FMRIB Software Library) by integrating the FreeSurfer brain extraction tool with a "one-step resampling" method. The main goal of the study is to evaluate the performance of OGRE in reducing inter-individual variability and enhancing activation detection capability compared to existing preprocessing methods such as FSL's default preprocessing and fMRIPrep. The research results indicate that the OGRE preprocessing method significantly reduced inter-individual variability in most regions of interest (ROIs) compared to FSL's default preprocessing and fMRIPrep, and detected higher activation signals in specific task-related brain areas (such as the contralateral primary motor cortex). Additionally, OGRE reduced the size of brain masks in standard space, primarily by removing non-brain voxels. Overall, the OGRE preprocessing method improved the reliability and accuracy of fMRI data analysis while maintaining similar overall results.