fMRat: an extension of SPM for a fully automatic analysis of rodent brain functional magnetic resonance series

• Published in: Medical & biological engineering & computing Vol. 54; no. 5; pp. 743 - 752
• Main Authors: Chavarrías, Cristina, García-Vázquez, Verónica, Alemán-Gómez, Yasser, Montesinos, Paula, Pascau, Javier, Desco, Manuel
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01-05-2016; Springer Nature B.V
• Subjects: Algorithms; Analysis; Animals; Automation; Bioengineering; Biomedical and Life Sciences; Biomedical Engineering and Bioengineering; Biomedicine; Brain; Brain - physiology; Computer Applications; Computer Graphics; Computer programs; Design; Hemodynamics; Human Physiology; Imaging; Kernels; Magnetic Resonance Imaging - methods; Matlab; Medical imaging; Neuroimaging; NMR; Nuclear magnetic resonance; Original Article; Pipelines; Radiology; Rats, Sprague-Dawley; Rats, Wistar; Rodents; Smoothing; Software; Statistical analysis; Studies; User interface; User-Computer Interface; fMRI; SPM; Rat; Automatic analysis
• ISSN: 0140-0118; 1741-0444
• DOI: 10.1007/s11517-015-1365-9

The purpose of this study was to develop a multi-platform automatic software tool for full processing of fMRI rodent studies. Existing tools require the usage of several different plug-ins, a significant user interaction and/or programming skills. Based on a user-friendly interface, the tool provides statistical parametric brain maps (
t
and
Z
) and percentage of signal change for user-provided regions of interest. The tool is coded in MATLAB (MathWorks
®
) and implemented as a plug-in for SPM (
Statistical Parametric Mapping
, the Wellcome Trust Centre for Neuroimaging). The automatic pipeline loads default parameters that are appropriate for preclinical studies and processes multiple subjects in batch mode (from images in either Nifti or raw Bruker format). In advanced mode, all processing steps can be selected or deselected and executed independently. Processing parameters and workflow were optimized for rat studies and assessed using 460 male-rat fMRI series on which we tested five smoothing kernel sizes and three different hemodynamic models. A smoothing kernel of FWHM = 1.2 mm (four times the voxel size) yielded the highest
t
values at the somatosensorial primary cortex, and a boxcar response function provided the lowest residual variance after fitting. fMRat offers the features of a thorough SPM-based analysis combined with the functionality of several SPM extensions in a single automatic pipeline with a user-friendly interface. The code and sample images can be downloaded from
https://github.com/HGGM-LIM/fmrat
.