HELLO expert:
i am trying to analyse VLSM with NPM  (the version: Chris Rorden's NPM :: 6 June 2013 CacheMB = 512; Threads used = 4),   it could successfully run in the tutorial datasets，but in my datasets，it always showed "unsupported compressed data type 64", how can i solve this problem?
thank you

Hello,

I suspect this is related to:
  https://github.com/nipy/nibabel/issues/1046
if you create these images with Python, please choose an appropriate precision for your data (e.g. dtype=np.uint8 for binary data)


Assuming your data is binary lesion maps (where each voxel is either lesioned [1], or unlesioned [0]), the 64-bit precision is not required, and you can dramatically reduce image size by storing them as UINT8, e.g.
  fslmaths ~/T1 -add 0 ~/T1c -odt char
or to save as 32-bit float:
 fslmaths ~/T1 -add 0 ~/T1f -odt float


 I am glad that my legacy popular remains popular. I have moved my own development on to NiiStat.
  https://www.nitrc.org/plugins/mwiki/index.php/niistat:MainPage
 Feel free to maintain and exend the NPM source code and project.