CfMRIPrep: A MATLAB toolbox for integrated preprocessing of complex-valued fMRI data

• Published in: International Conference on Information Science and Technology pp. 167 - 171
• Main Authors: Li, Si-Qi, Ma, Ming-Jun, Lin, Qiu-Hua, Calhoun, Vince D.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 08.12.2023
• Subjects: Complex-valued fMRI data; Functional magnetic resonance imaging; Magnetic heads; Manuals; MATLAB toolbox; preprocessing; Smoothing methods; Software libraries; Spatial databases; SPM; Timing
• ISSN: 2573-3311
• DOI: 10.1109/ICIST59754.2023.10367157

Functional magnetic resonance imaging (fMRI) has been widely used in the study of brain function due to its noninvasiveness and high spatial resolution. Preprocessing of fMRI data is essential to remove the noise caused by head motion and scanners, and to enable comparison across different subjects and various studies. To date, a number of toolboxes such as Statistical Parametric Mapping (SPM) and FMRIB Software Library (FSL) have been developed for preprocessing magnitude-only fMRI data. In fact, complete fMRI data are complex-valued including both magnitude and phase data. Complex-valued fMRI data can detect additional and meaningful brain activities beyond the magnitude-only fMRI data. However, there is a lack of a toolbox for preprocessing complex-valued fMRI data. As such, we design a new MATLAB toolbox named CfMRIPrep to perform this role. CfMRIPrep includes phase unwrapping of phase fMRI data, motion correction, spatial normalization and spatial smoothing for both magnitude and phase fMRI data. To do this, the parameters of motion correction and spatial normalization are computed from the magnitude data and applied to the phase data. Experimental results show that CfMRIPrep provides the same preprocessing results as a manual approach and is easy-to-use allowing for automated processing and also can be used for preprocessing of magnitude-only fMRI data.