gii header - NIFTI_INTENT _LABEL

114 views
Skip to first unread message

Erika Raven

unread,
Apr 12, 2021, 12:30:36 PM4/12/21
to HCP-Users
Hello, 

I am trying to view a .func.gii file and it appears I have some file header issues. The file represents cortical thickness and should be a continuous range from approximately 0~4. The file does load, but bins values more like a label at 1, 2, and 3. 

When checking the header, there does indeed seem to be something weird (see below). Is it possible to overwrite this with wb_command? 

And as a second q, I saw in the examples these data are saved differently - .dscalar.nii - is there a major difference here in usability?

Here is the example,

wb_command -file-information {subjid}_native_rms_tlink_12mm_left.func.gii

This is the output, 

WARNING: Metric File: T1_48114-18_native_rms_tlink_12mm_left.func.gii contains data array with NIFTI_INTENT_LABEL !!!

Name:                     T1_48114-18_native_rms_tlink_12mm_left.func.gii
Type:                     Metric
Structure:                CortexLeft
Data Size:                163.85 Kilobytes
Maps to Surface:          true
Maps to Volume:           false
Maps with LabelTable:     false
Maps with Palette:        true
All Map Palettes Equal:   true
Map Interval Units:       NIFTI_UNITS_UNKNOWN
Number of Maps:           1
Number of Vertices:       40962
Map   Minimum   Maximum    Mean   Sample Dev   % Positive   % Negative   Inf/NaN   Map Name   

  1     0.000     3.000   1.989        0.600       97.359        0.000         0   #1         


Thanks so much, 
Erika 


Coalson, Timothy Scott (S&T-Student)

unread,
Apr 12, 2021, 6:58:09 PM4/12/21
to HCP-Users
Where did you get this file from?  I don't actually expect workbench to go out of its way to round input values when it isn't being used/displayed as labels, so there may be additional things wrong with the file.

Gifti is an XML format, so if nothing else, you could edit it with a text editor to "NIFTI_INTENT_NORMAL" (yes, this means something statistical in nifti, just roll with it).  If the DataType is some integer type and not "NIFTI_TYPE_FLOAT32", then the damage was already done when the file was created, and it needs to be regenerated (do not edit the DataType attribute, or you will get garbage and/or errors).  Even if it says float32 already, it is still possible that only 4 unique values are used in the array.

Tim


From: Erika Raven <erika....@gmail.com>
Sent: Monday, April 12, 2021 11:30 AM
To: HCP-Users <hcp-...@humanconnectome.org>
Subject: [hcp-users] gii header - NIFTI_INTENT _LABEL
 
--
You received this message because you are subscribed to the Google Groups "HCP-Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hcp-users+...@humanconnectome.org.
To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/8021699e-1862-4ba0-bc64-0d74407f4689n%40humanconnectome.org.

Erika Raven

unread,
Apr 13, 2021, 10:57:27 AM4/13/21
to hcp-...@humanconnectome.org
Hi Tim, 

This is very useful information. The DataType was Float 32, so that worked, but after editing Intent type to Normal, I still got binned values. I think you are correct in that the file needs to be regenerated. This comes from the CIVET pipeline and I am hopeful these outputs can eventually be compatible with workbench for future analyses. We will try again, thanks!! 

Best wishes, 
Erika 

You received this message because you are subscribed to a topic in the Google Groups "HCP-Users" group.
To unsubscribe from this topic, visit https://groups.google.com/a/humanconnectome.org/d/topic/hcp-users/noVdNmL603U/unsubscribe.
To unsubscribe from this group and all its topics, send an email to hcp-users+...@humanconnectome.org.
To view this discussion on the web visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/SN6PR01MB49910B634C08D1E68E1C18D481709%40SN6PR01MB4991.prod.exchangelabs.com.
Reply all
Reply to author
Forward
0 new messages