Hi - it sounds like maybe it's working fine within the limits of slight differences in mathematical precision between the C++ vs matlab parts of the processing - so the main question would be - have you looked in a viewer at the difference image - e.g. are the voxels with large differences isolated or eg at the edge of the brain?
Cheers. > On 28 Sep 2016, at 02:49, Glasser, Matthew <[email protected]> wrote: > > I think this is more of a question for the FSL list, but I don’t know fsl_glm > well enough to say if what you are doing is equivalent or not. > > Peace, > > Matt. > > From: "Ely, Benjamin" <[email protected] <mailto:[email protected]>> > Date: Tuesday, September 27, 2016 at 7:20 PM > To: Matt Glasser <[email protected] <mailto:[email protected]>>, "Burgess, > Gregory" <[email protected] <mailto:[email protected]>>, > "[email protected] <mailto:[email protected]>" > <[email protected] <mailto:[email protected]>> > Subject: Re: [HCP-Users] MELODIC denoising vs. released ICA-FIX datasets > > Hi Matt and Greg, > > Thanks for the feedback! I've looked at the various fix .m files from the > current release; based on fix_3_clean.m, I tried the following for a single > resting-state run: > > # highpass filter; sigma of 1000.08 = FWHM of 2355 per Smith et al 2013 > NeuroImage, also consistent with comments in the fix_3_clean.m script > fslmaths rfMRI_REST1_LR.nii.gz -bptf 1000.08 -1 REST1LR_bp > > # format movement parameters (manually corrected header after paste step, not > shown) > Text2Vest Movement_Regressors.txt Movement_Regressors.mat > Text2Vest Movement_Regressors.txt Movement_Regressors_dt.mat > paste Movement_Regressors.mat Movement_Regressors_dt.mat > > Movement_Regressors_all.mat > > # regress movement parameters out of timeseries and re-add mean > fsl_glm -i REST1LR_bp.nii.gz -d Movement_Regressors_all.mat > --out_res=REST1LR_bp_mc_demeaned.nii.gz --demean > fslmaths REST1LR_bp.nii.gz -Tmean REST1LR_bp_mean > fslmaths REST1LR_bp_mc_demeaned.nii.gz -add REST1LR_bp_mean.nii.gz > REST1LR_bp_mc > > # regress movement parameters out of melodic mix > > fsl_glm -i filtered_func_data.ica/melodic_mix -d Movement_Regressors_all.mat > --out_res=melodic_mix_mc --demean > > # regress unique variance from bad components (taken from .fix file) out of > timeseries > > fsl_regfilt -i REST1LR_bp_mc.nii.gz -d melodic_mix_mc -o > REST1LR_bp_mc_softICA -f "1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, > 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 28, 29, 30, 31, 32, 33, 34, 36, > 38, 39, 40, 41, 42, 43, 44, 45, 46, 49, 50, 51, 52, 53, 54, 55, 56, 57, 61, > 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 81, 82, 83, 84" > > # compare against HCP's released FIX-denoised file > fslmaths rfMRI_REST1_LR_hp2000_clean -sub REST1LR_bp_mc_softICA > diff_REST1LR_bp_mc_softICA > > Visual inspection and fslstats indicate reasonably good agreement between my > denoised file and the HCP's denoised file; the mean difference is about 0.84 > units (compared to a mean signal intensity of around 10,000), and the > "robust" range of the difference is about +/- 72 units. More worryingly, > though, the maximum difference is around 2000 units, and around 6000 voxels > show differences greater than 500 units, so I'm not sure machine precision > can account for the differences. > > Does the above denoising scheme seem consistent with what FIX is doing? I > plan to use FIX going forward, rather than trying to replicate it using the > FSL command-line, but I'd like to understand any discrepancies between the > two. > > Thanks again, > -Ely > > > > The materials in this message are private and may contain Protected > Healthcare Information or other information of a sensitive nature. If you are > not the intended recipient, be advised that any unauthorized use, disclosure, > copying or the taking of any action in reliance on the contents of this > information is strictly prohibited. If you have received this email in error, > please immediately notify the sender via telephone or return mail. > > _______________________________________________ > HCP-Users mailing list > [email protected] <mailto:[email protected]> > http://lists.humanconnectome.org/mailman/listinfo/hcp-users > <http://lists.humanconnectome.org/mailman/listinfo/hcp-users> --------------------------------------------------------------------------- Stephen M. Smith, Professor of Biomedical Engineering Head of Analysis, Oxford University FMRIB Centre FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK +44 (0) 1865 222726 (fax 222717) [email protected] http://www.fmrib.ox.ac.uk/~steve <http://www.fmrib.ox.ac.uk/~steve> --------------------------------------------------------------------------- Stop the cultural destruction of Tibet <http://smithinks.net/> _______________________________________________ HCP-Users mailing list [email protected] http://lists.humanconnectome.org/mailman/listinfo/hcp-users
