Rodent Automated Bold Improvement of EPI Sequences (RABIES): A standardized image processing and data quality platform for rodent fMRI

Abstract

Functional magnetic resonance imaging (fMRI) in rodents holds great potential for advancing our understanding of brain networks. However, rodent fMRI lags the human community with regards to the availability of image processing softwares to support reproducibility, a significant challenge facing neuroimaging. To address these shortcomings, we introduce RABIES, a novel pipeline designed along best practices for reproducibility and transparency handling processing stages from image registration to data quality assessment during analysis. First, we present a robust registration workflow validated across both acquisition sites and species. Leveraging insights from the human literature, we introduce a novel framework for data quality assessment of single scans and during group-level network analysis. This framework will support the development of standardized high-quality acquisition protocols and harmonization of findings between studies. Taken together, the RABIES software will allow the emerging community to adopt reproducible practices and foster progress in translational neuroscience.

A consensus protocol for functional connectivity analysis in the rat brain

“…Overall, we found that the aggregated datasets represent current rodent fMRI acquisition trends 9 . Moreover, given the low exclusion rate due to mis-registration, we concluded that the RABIES toolbox can be effectively employed to pre-process rat datasets despite widely varying acquisition parameters 14 . This paves the way for reproducible and interoperable data processing across sites.…”

“…Scans were organized according to the BIDS format 27 . Pre-processing was performed on each scan session separately using a reproducible containerized software environment for RABIES 0.3.5 (Singularity 3.7.3-1.el7, Sylabs) 14 . The pre-processing was performed using autobox 28 , N4 inhomogeneity correction 29 , motion correction 29 , a rigid registration between functional and anatomical scans 29 , non-linear registration between anatomical scan and template and a common space resampling to 0.3 × 0.3 × 0.3 mm 3 .…”

Multimodal measures of spontaneous brain activity reveals both common and divergent patterns of cortical functional organization

“…We preprocessed BOLD data using the RABIES package [53], which has been successfully applied to previous mouse [18] and rat [19] fMRI datasets. This included a simple high-pass filtering step [6] which resulted in a frequency range of 0.01 to 0.5 Hz for BOLD (acquisition rate was 1 Hz; see Methods#).…”

“…Head motion is known to systematically alter the correlation structure across brain regions [78], which could in turn alter derived overlap scores. To address this, we applied a conservative threshold to remove high motion frames inferred from BOLD data [53] (the same set of frames removed from both modalities; see Methods). Finally, we applied a statistical thresholding method to eliminate small membership values that were not significantly different than zero (see Methods).…”

“…For fMRI preprocessing, we used RABIES (Rodent automated BOLD improvement of EPI sequences) v0.4.2 [71]. We applied functional inhomogeneity correction N3 (nonparametric nonuniform intensity normalization; [102, 103]), motion correction (ANTS, Advanced normalization tools; [99]), and slice time correction, all in native space.…”