Some amendments to this: 1. If there is an oblique sform in the input image header to FreeSurfer, there can be more than a translation. The HCP MRI Pipelines always avoid this situation. The MEG pipelines may not. 2. For human data, there is not volume space faking by the HCP Pipelines, rather the data are down sampled to 1mm and then highres data are incorporated in white surface placement and pial surface placement (at their full resolution and with correct headers). The faking that Tim is talking about occurs in an unreleased pipeline for processing non-human primate data.
Matt. From: <hcp-users-boun...@humanconnectome.org<mailto:hcp-users-boun...@humanconnectome.org>> on behalf of Timothy Coalson <tsc...@mst.edu<mailto:tsc...@mst.edu>> Date: Tuesday, January 31, 2017 at 7:44 PM To: K Jeffrey Eriksen <eriks...@ohsu.edu<mailto:eriks...@ohsu.edu>> Cc: "hcp-users@humanconnectome.org<mailto:hcp-users@humanconnectome.org>" <hcp-users@humanconnectome.org<mailto:hcp-users@humanconnectome.org>> Subject: Re: [HCP-Users] importing HCP files into Brainstorm Freesurfer has a coordinate displacement between their surfaces and volumes, which depends on the volume space that was used. That might be part of the picture. However, this is purely a shift, no rotation or flip, and I think is typically about half the brain size, so something else must also be going on. Another thing that that the HCP pipelines do is that they pull some tricks to avoid downsampling the 0.7mm structural images when putting them through freesurfer. This involves a specific cropping and faking a new volume space that pretends to be 1mm isotropic. The HCP pipelines correct for this change in coordinate space, and also correct for the shift that freesurfer puts into its surface coordinates. The end result is that all of the anatomical gifti surface outputs from the HCP pipelines have raw coordinates that are already aligned with the relevant 0.7mm T1 voxel coordinates (NOT the modified volumes that it runs freesurfer on), per the nifti standard coordinate conventions (via the sform method). Thus, wb_view and wb_command do not need to apply any shifts when reading, displaying, or computing in order to align them. Hopefully this is useful information for understanding what the issue might be. Here are some relevant code snippets as to how we interpret the coordinate information in the file types in question, which might be useful to the Brainstorm developers: Note that we do not apply any of the transform matrices that are in the gifti file, because freesurfer does something strange with them: https://github.com/Washington-University/workbench/blob/master/src/Files/SurfaceFile.cxx#L263 So, we use the coordinate array exactly as it comes out of the base64/gzip encoding. Using nifti sform info: https://github.com/Washington-University/workbench/blob/master/src/FilesBase/VolumeSpace.h#L145 m_sform is just a copy of the sform fields from the nifti header (as long as sform_code is nonzero, see https://github.com/Washington-University/workbench/blob/master/src/Nifti/NiftiHeader.cxx#L220 - otherwise it constructs the equivalent sform from whatever spacing information is available) Some computation code showing direct comparison of voxel and surface coordinates: https://github.com/Washington-University/workbench/blob/master/src/Algorithms/AlgorithmVolumeToSurfaceMapping.cxx#L306 Tim On Tue, Jan 31, 2017 at 4:33 PM, K Jeffrey Eriksen <eriks...@ohsu.edu<mailto:eriks...@ohsu.edu>> wrote: Hello, I am a very new HCP Connectome WB user trying to import nifti and gifti files into Brainstorm (Bst) for EEG modeling. I am having a strange problem and am looking for some direction. In the past I was able to import T1.nii volume files and lh.pial/rh.pial files from FreeSurfer with no problem into Brainstorm. Now I am trying to import similar files proceed by the HCP pipeline. Let me call these HCP-T1 and HCP-cortsurf, and the original FreeSurfer files FS-T1 and FS-cortsurf. When I import the FS files into Bst they can be plotted in 3D, and align well. However, when I import the HCP-T1 and HCP-cortsurf, they do not align. In fact, assuming the T1 is correctly displayed, the surface ends up displaced about two head diameters worth and is also inverted L-R and upside down. This is very odd since these same two files align perfectly when viewed with the Connectome Workbench. I need to point out that the FS-cortsurf is a full 250,000 vertex file for a particular subject, while the HCP-cortsurf is the 64,000 vertex version using the HCP standard. Please provide me with some suggestions of how to begin to address this problem. The Brainstorm support personnel are unable to help me on this. Thanks, -Jeff _______________________________________________ HCP-Users mailing list HCP-Users@humanconnectome.org<mailto:HCP-Users@humanconnectome.org> http://lists.humanconnectome.org/mailman/listinfo/hcp-users _______________________________________________ HCP-Users mailing list HCP-Users@humanconnectome.org<mailto:HCP-Users@humanconnectome.org> http://lists.humanconnectome.org/mailman/listinfo/hcp-users ________________________________ 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 HCP-Users@humanconnectome.org http://lists.humanconnectome.org/mailman/listinfo/hcp-users
