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

Reply via email to