When I calculate the the FD with fsl_motion_outliers, I get for the rawData about the same as frmiprep gives me out in the confound_regressors files.
But calculating the FD with fsl_motion_outliers of the fmriprep output (sub-2_ses-2_task-rest_run-1_space-MNI152NLin2009cAsym_desc-preproc_bold.nii.gz) gives me FD values which are a lot lower.
(Below 0.5 so I would not need to despike or motion scrub).
I guess the reason is the motion correction in fmriprep. Do I still need to despike or motion scrub the data then?
And after regressing out ‘trans_x’,‘trans_y’,‘trans_z’,‘rot_x’,‘rot_y’,‘rot_z’,‘global_signal’,‘a_comp_cor_00’,‘a_comp_cor_01’ and bandpassfiltering of the fmriprep output the FD values are very.
Regarding this paper (http://www.ajnr.org/content/38/2/336.short) and https://www.researchgate.net/publication/273577605_Evaluation_of_ICA-AROMA_and_alternative_strategies_for_motion_artifact_removal_in_resting_state_fMRI
" However, both ICA-AROMA and ICA-FIX resulted in significantly improved resting-state network reproducibility and decreased loss in tDoF compared to spike regression and scrubbing." Independent component analysis denoising would be better than scrubbing.
So I could use the aroma tag in fmriprep and do not use despiking or motion scrubbing?
But regress ‘trans_x’,‘trans_y’,‘trans_z’,‘rot_x’,‘rot_y’,‘rot_z’,‘global_signal’,‘a_comp_cor_00’,‘a_comp_cor_01’
out on the fmriprep with the aroma tag output afterwards?