Temporal Dynamics Analysis on converted CIFTI data

Hello,

I'm trying to run an ALFF analysis using the DPABI TDA toolbox on converted cifti data and the returned data is very odd.  Using wb_command -cifti-convert I transformed resting state cifti data to a 32767x3x1 nifti file, ran TDA, then converted back to cifti.  The TDA generated data when projected back to the surface shows right angles in data patterns, large squares of high activation, and has 0 activity in the left cerebellum.  I've tested the cifti->nifti->cifti conversion without any problem so I don't think it's that.  Running this on nifti data produced at an earlier preprocessing step didn't create any issues.  Checking the cifti converted ALFF nifti in matlab shows that the first 7019 voxels in the third column are set to 0.  Any idea what is going wrong?  I've attached images of the ALFF output and the TDA options used.

 

Thank you,

James Kennedy

AttachmentSize
Image icon TempDynamics.png23.74 KB
Image icon PostALFFSurface.png457.26 KB
Image icon PostALFFVol.png83.27 KB

Forums:

It should be nDim1*nDim2*nDim3*nDimTimePoints.

Seems your data is not in this format.

It doesn't look like that's the issue.  fslinfo gives my converted cifti dimensions as 

data_type      FLOAT32
dim1           32767
dim2           3
dim3           1
dim4           383
datatype       16
pixdim1        1.000000
pixdim2        1.000000
pixdim3        1.000000
pixdim4        1.000000
cal_max        0.0000
cal_min        0.0000
file_type      NIFTI-1+,
 
my regular nifti version is 
data_type      FLOAT32
dim1           91
dim2           109
dim3           91
dim4           383
datatype       16
pixdim1        2.000000
pixdim2        2.000000
pixdim3        2.000000
pixdim4        0.800000
cal_max        0.0000
cal_min        0.0000
file_type      NIFTI-1+
 
The string of 0 s turns out to be part of the expected cifti to nifti format.  Looking at the converted cifti rfMRI data and the ALFF output in fslview I noticed that the X dimension was negative in the later, flipping the data matrix in matlab gives results that look more like what I expected.
 
James