You can resample cifti data to a different resolution without splitting it up first, use wb_command -cifti-resample . You will need to specify both -left-spheres and -right-spheres.
Tim On Fri, Nov 3, 2017 at 1:26 PM, Jan Mathijs Schoffelen < [email protected]> wrote: > Dear Julia, > > If you want to parcellate the low-resolution (i.e. ~4k vertices per > hemisphere) MEG source level time series, you first need to create the > appropriate low-resolution dlabel file, sampled at 4k. The one you provided > in your e-mail represents the parcellation at 32k vertices. This can be > done using some clever combination of wb_commands, which involves splitting > the dlabel.nii into two files (one for each cortical hemisphere), followed > by a downsampling (using wb_command -label-resample). The appropriate > low-level 4k spherical meshes are in the megconnectome software (in the > template directory). > > The MEG data does not have to be transformed in any way. Although the > source time series have been computed in native space, due to the > one-to-one vertex-wise correspondence across the different surface > representations, individual vertices can be directly compared across > subjects. > > Best wishes, > > Jan-Mathijs > > PS: alternatively, you can have a look here http://www. > fieldtriptoolbox.org/tutorial/networkanalysis. This tutorial uses some > data (available through FTP) which contains the parcellation you are > looking for at the appropriate resolution (if memory serves me well). It > is, however, represented as a matlab .mat file. > > > > > > > On 3 Nov 2017, at 18:58, Julia Hamilton <[email protected]> > wrote: > > Hello, > > > We are working on parcellating HCP’s MEG data using the Glasser 2016 > parcels (inputFile and parcellationFile attached). The MEG data was > downloaded from WU-Minn HCP Data-1200 Subjects Study, Source-Level > Processed, Resting State (dtseries) (ex. > “100307_MEG_3-Restin_bfblpenv_betahigh.power.dtseries.nii”; > each file having ~8000 surface vertices). > > > Although we had previously used the -cifti-parcellate function (below, via > Matlab) to parcellate fMRI data, we are receiving errors when attempting to > do so with the MEG data. The following error occurred when using the > 100307_MEG_3-Restin_bfblpenv_betahigh.power.dtseries.nii as the inputFile. > > >> eval([‘!wb_command -cifti-parcellate ’ inputFile ‘ ’ parcellationFile ‘ > COLUMN ’ outputFile]) > ERROR: no parcels found, output file would be empty, aborting > > We have access to the surface locations via the corresponding gifti files > (ex. “100307.L.midthickness.4k_fs_LR.surf.gii”, attached), however, we > are unsure how to incorporate this surface location data into the dtseries > file. > > We’re also unclear whether the MEG data needs to be transformed before > applying the -cifti-parcellate function e.g. from native surface space > (assumed to be the original space of the MEG dtseries files) to MNI space. > > Any input or advice would be much appreciated! > > Thank you, > Julia Hamilton > > 100307_MEG_3-Restin_bfblpenv_betahigh.power.dts... > <https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fdrive.google.com%2Ffile%2Fd%2F1qXGWCsFXB11F3c6mEJ2eyQOYX-AR-WPj%2Fview%3Fusp%3Ddrive_web&data=02%7C01%7Cjulia.hamilton%40rutgers.edu%7C767af945945e4c2654ba08d522e42667%7Cb92d2b234d35447093ff69aca6632ffe%7C1%7C0%7C636453285659453379&sdata=DPoVj0G%2FzDRbMF9ST548wlpGM3vRKCB832II6TdKljU%3D&reserved=0> > > _______________________________________________ > HCP-Users mailing list > [email protected] > http://lists.humanconnectome.org/mailman/listinfo/hcp-users > <100307.L.midthickness.4k_fs_LR.surf.gii><Q1-Q6_RelatedParcellation210.L. > CorticalAreas_dil_Colors.32k_fs_LR.dlabel.nii> > > > > > _______________________________________________ > HCP-Users mailing list > [email protected] > http://lists.humanconnectome.org/mailman/listinfo/hcp-users > _______________________________________________ HCP-Users mailing list [email protected] http://lists.humanconnectome.org/mailman/listinfo/hcp-users
