Hi Phil:

So I am now successful at tracking in 2 and 3-tensor reconstructed data, but the variable "multitensor" data seems to have some trouble in that "track" complains it's running out of voxels.

In short, I first run

voxelclassify -inputdatatype short -bgthresh 200.0 -schemefile ${scheme} -order 6 -ftest 1E-20 1E-6 1E-6 > ${classify_map}

then

multitenfit -voxclassmap ${classify_map} -schemefile ${scheme}

and finally

cat tensors.Bdouble | track -inputmodel multitensor -interpolate -anisthresh 0.17 -ipthresh 0.8 -seedfile ${sources}

Scheme file and seed file are the same for multitensor as they are for twotensor and threetensor. Indeed, the same "track" command is used for all three. Feeding the twotensor and threetensor tensor files works, feeding the multitensor file gives me

SEVERE: Exception in thread "main" class data.DataSourceException: No more voxels in data source.
at data.VoxelOrderDataSource.readNextVoxel(VoxelOrderDataSource.java:231)
at data.VoxelOrderDataSource.nextVoxel(VoxelOrderDataSource.java:190)
at tractography.DT_TractographyImage.(DT_TractographyImage.java:81)
at tractography.DT_TractographyImage.getDT_TractographyImage(DT_TractographyImage.java:277)
at apps.StreamlineTractography.execute(StreamlineTractography.java:972)
at apps.EntryPoint.main(EntryPoint.java:226)

Do I need to pass any additional options to "tract" for "multitensor" data?

By the way - I am also wondering if there is a way to get a usable FA map out of multitensor data? Simply running the tensor data through "fa" seems to produce variable numbers of FA values per pixel, depending on the number of tensors reconstructed at each.

Thanks again!

TR

PS: the command I use for reconstructing the multitensor data is

multitenfit -voxclassmap ${classify_map} -schemefile ${scheme}