tag:blogger.com,1999:blog-5737874959005852552.post8957150121276544264..comments2023-10-18T16:31:41.080-05:00Comments on MVPA Meanderings: HCP: extracting timecourses from the surface filesJo Etzelhttp://www.blogger.com/profile/04277620767760987432noreply@blogger.comBlogger13125tag:blogger.com,1999:blog-5737874959005852552.post-10932095535585666092017-09-10T23:15:05.587-05:002017-09-10T23:15:05.587-05:00Hi Jo,
Ah~ , I just read Tim Coalson's comment...Hi Jo,<br />Ah~ , I just read Tim Coalson's comment. Does that mean after -cifti-separate command, the obtained out.func.gii file actually contains all the surface vertices data for each hemispheres? If so, how can I relate their indices with the corresponding 3d coordinates?<br />Thank you!<br />Best,<br />Ray.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-5737874959005852552.post-64837560264345671142017-09-10T23:02:49.208-05:002017-09-10T23:02:49.208-05:00Hi Jo,
I just started my brain research project. Y...Hi Jo,<br />I just started my brain research project. Your tutorials have helped a lot! <br />I have a question here. I want to do cortex parcellation. May I know if this post is the right thing to start with in order to extract the surface time courses for HCP data?<br />If so, does that mean that I should look for a MNI brain as the input for roi.nii.gz inside the volume-to-surface-mapping call?<br /><br />Thank you very much!<br />Best,<br />Ray.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-5737874959005852552.post-78619040756159098382016-12-19T16:50:55.652-06:002016-12-19T16:50:55.652-06:00If you're using matlab, Guillaume Flandin'...If you're using matlab, Guillaume Flandin's library (http://www.artefact.tk/software/matlab/gifti/) should write giftis; his page shows an example.Jo Etzelhttps://www.blogger.com/profile/04277620767760987432noreply@blogger.comtag:blogger.com,1999:blog-5737874959005852552.post-27955855031940472682016-12-16T11:00:30.010-06:002016-12-16T11:00:30.010-06:00Hi, this is a great tutorial on HCP data. I really...Hi, this is a great tutorial on HCP data. I really joined it.<br />I have a question about writing the results. Maybe you can help me.<br />I have extracted the timecourses from my regions of interest on the surface and performed statistical analysis on them. Then, I tried to write the results back to the surface. But I could not find the right command to do it. <br />Do you know anyways to write the statistics back to the surface vertex in MATLAB?<br /><br />Thank you very much.<br />Best,<br /> Yu.Anonymoushttps://www.blogger.com/profile/04782651760685478901noreply@blogger.comtag:blogger.com,1999:blog-5737874959005852552.post-72619803652493626122015-12-08T03:06:12.168-06:002015-12-08T03:06:12.168-06:00Thank you a lot! Your suggestion helps.
I have an...Thank you a lot! Your suggestion helps.<br /><br />I have another quick question. <br /><br />Do you know a simple way to draw some points on the brain? (lets see I have a list of XYZ points, how to draw them on the cortex with different colour?). Would prefer to use HCP Workbench.<br /><br />Thank you a lot.Anonymoushttps://www.blogger.com/profile/06065336394216134906noreply@blogger.comtag:blogger.com,1999:blog-5737874959005852552.post-26264291155778569272015-08-27T08:27:06.843-05:002015-08-27T08:27:06.843-05:00Have you tried using wb_command -surface-closest-v...Have you tried using wb_command -surface-closest-vertex to get the vertex closest to your MNI coordinate? In the HCP data there isn't a 1-to-1 correspondence between vertices and voxels: sometimes multiple vertices might fall into one voxel, other times, just one or none.Jo Etzelhttps://www.blogger.com/profile/04277620767760987432noreply@blogger.comtag:blogger.com,1999:blog-5737874959005852552.post-89475232568658209712015-08-25T05:09:51.045-05:002015-08-25T05:09:51.045-05:00This comment has been removed by a blog administrator.Anonymoushttps://www.blogger.com/profile/06065336394216134906noreply@blogger.comtag:blogger.com,1999:blog-5737874959005852552.post-92075466970724045712015-08-25T05:07:07.865-05:002015-08-25T05:07:07.865-05:00Oh Sorry for the stupid question previously.
Tha...Oh Sorry for the stupid question previously. <br /><br />Thank you very much for answering my question! I truly appreciate it. I am an undergraduate student and have just started doing research on HCP.<br /><br />What I really want to ask is: I now have some some MNI coordinated volume voxels(like x=-25, y=-98, z=-12), and I am also able to translate it into IJK.) --- How do I extract time course for a volume index from *.dtseries.nii?<br /><br />The method you described here only works for surface data. I tried your way by doing %wb_command -cifti-separate tfMRI_EMOTION_LR_Atlas.dtseries.nii COLUMN -volume-all out1.func.gii, and use gifti to read the output, however the file was always too big and I could not read in Matlab.<br /><br />Then I tried another way by using this:<br />data = ft_read_cifti('E:\FYP\Workbench\workbench\bin_windows64\temp\tfMRI_EMOTION_LR_Atlas.dtseries.nii');<br /><br />The data I get is data = <br /><br /> dimord: 'pos_time'<br /> data: [96854x176 double]<br /> time: [1x176 double]<br /> hdr: [1x1 struct]<br /> unit: 'mm'<br /> brainstructure: [96854x1 double]<br /> brainstructurelabel: {1x21 cell}<br /> dim: [91 109 91]<br /> pos: [96854x3 double]<br /> transform: [4x4 double]<br /> dtseries: [96854x176 double]<br /><br />It seems very easy for me since I only need to match my xyz with data.pos and get the dtseries of that line.<br /><br />However, this does not work since the data.pos was not correct. (the result I got does not match with what I got in the workbench).<br /><br />Thank you for your answer again. And I would deeply appreciate if you could help me out. Anonymoushttps://www.blogger.com/profile/06065336394216134906noreply@blogger.comtag:blogger.com,1999:blog-5737874959005852552.post-18095159408303232162015-08-20T10:16:59.729-05:002015-08-20T10:16:59.729-05:00Oh, just a string giving the path to where the out...Oh, just a string giving the path to where the out.func.gii file is located on your computer. On a windows machine, the syntax will be something like inpath = 'd:/temp/';.Jo Etzelhttps://www.blogger.com/profile/04277620767760987432noreply@blogger.comtag:blogger.com,1999:blog-5737874959005852552.post-85240631438923057022015-08-20T02:53:24.785-05:002015-08-20T02:53:24.785-05:00Hi,
What is the inpath variable here in this line...Hi,<br /><br />What is the inpath variable here in this line?<br /><br />wm = gifti([inpath 'out.func.gii']); % load the functional GIFTI<br /><br />I am using gifti1.5 and get this error - Undefined function or variable 'inpath'.<br /><br />Thank you.Anonymoushttps://www.blogger.com/profile/06065336394216134906noreply@blogger.comtag:blogger.com,1999:blog-5737874959005852552.post-52948762466637427782015-01-05T15:58:58.519-06:002015-01-05T15:58:58.519-06:00I'm not sure what you meant by "surface p...I'm not sure what you meant by "surface parcellations", but if you are referring to the out.func.gii file from the -cifti-separate example, that contains only surface data for one hemisphere. The -cifti-separate command also has a -volume option (for one structure), and a -volume-all (for all structures represented as voxels) option, which will output the volume data as nifti volume files. Similarly, it also has an -roi option to tell you which voxels have data. The -crop option is a notable difference, in that it saves space by using the bounding box of valid data, but if you need to use it with something of the original volume space, you'll need to be careful with your indices (the nifti origin information will align the cropped volume with the original volume space, so you can work out the change in indices).<br /><br />For reference, you can run any workbench command with no arguments (i.e., "wb_command -cifti-separate") to get its usage information, or browse them here:<br /><br />http://www.humanconnectome.org/software/workbench-command.php<br />Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-5737874959005852552.post-58153282633858108012015-01-05T11:37:57.025-06:002015-01-05T11:37:57.025-06:00Hi. Thanks so much for your helpful website! Made ...Hi. Thanks so much for your helpful website! Made extracting surface time series from the workbench really simple!<br /><br />This may be a really simple question, but do you know if there is an easy way to extract subcortical and cerebellar time series data from the same file that you use for the surface parcellations (i.e. 'dtseries.gii')?<br /><br />Thanks!<br /><br />Macmacshinehttps://www.blogger.com/profile/04885306649207985282noreply@blogger.comtag:blogger.com,1999:blog-5737874959005852552.post-74835601706187888152014-05-20T16:25:28.943-05:002014-05-20T16:25:28.943-05:00For the -cifti-separate command (and similarly for...For the -cifti-separate command (and similarly for several other cifti commands), the COLUMN specifier is actually to tell it which dimension to split along - it needs to be of type BRAIN_MODELS, because that's what contains structures. The timeseries are actually along the row.<br /><br />Some people find this "along column" notation to be unintuitive, but one reason it is that way is that CIFTI-2 can contain 3 dimensions, and future versions may allow even more, and "across" doesn't generalize to the third dimension, as the only remaining direction is along the third dimension.Anonymousnoreply@blogger.com