Thanks Matt! But now how do I create a dscalar file from a text file of labels (one label ID per line, line number = vertex ID)? Looking at the help for "wb_command -cifti-create-dense-scalar” it seems I need a "metric” file for each surface – but what is the format of a metric file?
Cheers, M@ On 18/5/16 13:10 , "Glasser, Matthew" <[email protected]<mailto:[email protected]>> wrote: You should be running the import command on a dscalar file that has your parcellation in it with values that match the keys in your text file. Peace, Matt. From: <[email protected]<mailto:[email protected]>> on behalf of Matthew George Liptrot <[email protected]<mailto:[email protected]>> Date: Wednesday, May 18, 2016 at 4:39 AM To: HCP Listserv <[email protected]<mailto:[email protected]>> Subject: [HCP-Users] Generating a dlabel.nii file from a new parcellation Hi, I'm trying to create a CIFTI dlabel.nii file from a novel parcellation we have generated from a structural connectivity matrix. However, I seem to hit a problem when the number of new clusters is less than those present in the template CIFTI file I am using for its header-info. In this example below, I want to generate a dlabel file with 68 clusters using the Desikan-Killiany dlabel file as a skeleton. Here are the commands I am using: wb_command -cifti-label-import 100307.aparc.32k_fs_LR.dlabel.nii CIFTI_LUT_ModelA_Clusters68.txt 100307.aparc.32k_fs_LR_ButWithClusterLUT.dlabel.nii -drop-unused-labels wb_command -cifti-convert -from-text ClusterIDs_ModelA_Clusters68.txt 100307.aparc.32k_fs_LR_ButWithClusterLUT.dlabel.nii ModelA_Clusters68.32k_fs_LR.dlabel.nii wb_command -set-map-names ModelA_Clusters68.32k_fs_LR.dlabel.nii -map 1 ModelA_Clusters68 Where the two textfiles used are defined thus: # CIFTI_LUT_ModelA_Clusters68.txt A plain text file with lines: Cluster_0001 1 85 85 0 255 Cluster_0002 2 85 170 0 255 Cluster_0003 3 85 255 0 255 etc # ClusterIDs_ModelA_Clusters68.txt A plain text file with lines: 48 23 34 19 etc However, running the command: wb_command -file-information <CIFTI_ModelClusters_dlabel> generates the following output: Name: ModelA_Clusters68.32k_fs_LR.dlabel.nii Type: Connectivity - Dense Label Structure: CortexLeft CortexRight Data Size: 237.65 Kilobytes Maps to Surface: true Maps to Volume: false Maps with LabelTable: true Maps with Palette: false Number of Maps: 1 Number of Rows: 59412 Number of Columns: 1 Volume Dim[0]: 0 Volume Dim[1]: 0 Volume Dim[2]: 0 Palette Type: None CIFTI Dim[0]: 1 CIFTI Dim[1]: 59412 ALONG_ROW map type: LABELS ALONG_COLUMN map type: BRAIN_MODELS Has Volume Data: false CortexLeft: 29696 out of 32492 vertices CortexRight: 29716 out of 32492 vertices Map Map Name 1 ModelA_Clusters68 Label table for ALL maps KEY NAME RED GREEN BLUE ALPHA 0 ??? 1.000 1.000 1.000 0.000 1 Cluster_0001 0.333 0.333 0.000 1.000 2 Cluster_0002 0.333 0.667 0.000 1.000 ... 68 Cluster_0068 1.000 1.000 1.000 1.000 69 LABEL_69 0.776 0.055 0.686 1.000 70 LABEL_70 0.196 0.569 0.169 1.000 As you can see, the original number of labels is somehow retained. And even though label values 69 and 70 are not present in the vertex:label mapping. (If I use the 100307.aparc.a2009s.32k_fs_LR.dlabel.nii file as a template, then I get a label table with 150 entries, again with the first 68 being the new ones). This is despite passing the '-drop-unused-labels' option to "wb_command -cifti-label-import”. (Not using ‘-drop-unused-labels’ gives the same result). So somehow I need to fully purge the CIFTI template file of its original LUT when the number of clusters that I want to insert is less than was originally present in the file. Any ideas? Is there a better way to do this? Cheers, M@ -- -- Matthew George Liptrot <http://about.me/matthewliptrot> Assistant Professor Image Group Department of Computer Science University of Copenhagen & Section for Cognitive Systems Department of Applied Mathematics and Computer Science Technical University of Denmark http://about.me/matthewliptrot _______________________________________________ HCP-Users mailing list [email protected] http://lists.humanconnectome.org/mailman/listinfo/hcp-users
