Dear all,
Hereby I would like to announce that ICA-AROMA is now available on GitHub (https://github.com/rhr-pruim/ICA-AROMA) or via http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/OtherSoftware. ICA-AROMA is a data-driven method to identify motion-related ICA components from FMRI data. Identified components are subsequently removed from FMRI data using fsl_regfilt. Unlike FSL-FIX, ICA-AROMA is not aimed at more generic denoising but focuses on removal of motion artifacts, therefore allowing a simple and robust classifier that does not require classifier re-training. The ICA-AROMA strategy as well as an extensive evaluation in the context of alternative methods for correction for secondary motion artifacts (motion parameter regression, spike regression, scrubbing, compcor, SOCK, and ICA-FIX) are discussed in:
1) Pruim, R.H.R., Mennes, M., van Rooij, D., Llera Arenas, A., Buitelaar, J.K., Beckmann, C.F., 2015. ICA-AROMA: A robust ICA-based strategy for removing motion artifact from fMRI data. Neuroimage 112, 267–277. doi:10.1016/j.neuroimage.2015.02.064
2) Pruim, R.H.R., Mennes, M., Buitelaar, J.K., Beckmann, C.F., 2015. Evaluation of ICA-AROMA and alternative strategies for motion artifact removal in resting-state fMRI. Neuroimage 112, 278–287. doi:10.1016/j.neuroimage.2015.02.063
! NOTE: Previous beta-versions of the python versions of ICA-AROMA as uploaded on GitHub (v0.1-beta & v0.2-beta) contained an error in removing identified components from the data. This issue is solved in version v0.3-beta. See the GitHub page for a log-report on the implemented changes. We recommend that users reprocess their data using the latest version of ICA-AROMA. This error was not present in the scripts used for the publications.
Feel free to contact us if you have any questions!
Happy denoising,
Best Regards,
Raimon Pruim
|