Multiband Artifact Regression in Simultaneous Slices (MARSS)

This is a Python version of a pipeline developed for use in simultaneous multi-slice (multiband; MB) fMRI data.
MARSS is a regression-based method that mitigates an artifactual shared signal between simultaneously acquired slices in unprocessed MB fMRI.
Software Authors: Philip N. Tubiolo, John C. Williams, Ashley Zhao, and Jared X. Van Snellenberg

Accompanies the following manuscript:
Tubiolo PN, Williams JC, Van Snellenberg JX.
Characterization and Mitigation of a Simultaneous Multi-Slice fMRI Artifact: Multiband Artifact Regression in Simultaneous Slices.
Hum Brain Mapp. 2024 Nov;45(16):e70066. doi: 10.1002/hbm.70066. PMID: 39501896; PMCID: PMC11538719.

Software Requirements

FSL must be installed and included in the system's PATH prior to use. For more information, visit https://fsl.fmrib.ox.ac.uk/fsl/docs/#/
This software has been tested on the following operating systems, but should be compatible with MacOS as well:
Linux: Red Hat Enterprise Linux 7.9
Windows: Windows 10 Home 64-bit

Hardware Requirements

MARSS should only require the minimum RAM to handle a single fMRI timeseries (approximately 2GB). However, it has been tested with these minimum specifications:
RAM: 16 GB
Processor: Intel(R) Core(TM) i7-10750H CPU @ 2.60GHz

With the above specifications, the total time taken for MARSS to complete on a single fMRI timeseries of 563 volumes is approximately 10 minutes.

Usage

Installation

To install this package, run the following command:

pip install MARSS

MARSS_main

MARSS_main is the main function of this pipeline, and the only function that must be directly run in order to perform MARSS on a single timeseries.

Syntax

MARSS_main(timeseriesFile, MB, workingDir,*args) performs MARSS artifact correction on a single unprocessed, MB fMRI timeseries.

Input Arguments

timeseriesFile (string): Full path to unprocessed, MB fMRI timeseries
MB (double): Multiband Acceleration Factor used during image acquisition
workingDir (string): Parent directory for all MARSS outputs. MARSS will create a separate folder within this folder named after timeseriesFile.
*args (string): Optional argument specifying a path to motion parameters.

Outputs in workingDir

za_.nii: this NIFTI is the MARSS corrected timeseries.
_slcart.nii: this NIFTI is the timeseries of MARSS-estimated artifact signal that was subtracted from timeseriesFile to produce za*.nii
_AVGslcart.nii: this NIFTI is the mean absolute value across timepoints of slcart.nii (shown as a single 3D volume).
corrMatrix*.png: slice correlation matrix of pre-MARSS data, along with the average difference in pearson correlation between simultaneously acquired slices and adjacent-to-simultaneous slices.
corrMatrixza*.png: slice correlation matrix of MARSS-corrected data.

Citation

When using MARSS, please cite the following:
Tubiolo PN, Williams JC, Van Snellenberg JX.
Characterization and Mitigation of a Simultaneous Multi-Slice fMRI Artifact: Multiband Artifact Regression in Simultaneous Slices.
Hum Brain Mapp. 2024 Nov;45(16):e70066. doi: 10.1002/hbm.70066. PMID: 39501896; PMCID: PMC11538719.

License

This software is released under the GNU General Public License Version 3.