Here is what I get from using "-debug 3"
[The-Ohio-State-University-Server:data2/coalition07/JUNE13_2007J] scnlab% Dimon -GERT_RECO -gert_filename save_EPI -dicom_org -quit -gert_nz 35 -gert_to3d_prefix EPI -debug 3 -infile_pattern "*.IMA.gz"
** MRD: mri_possibly_dicom() failure
** failed to read file 'WAC2006-007.*.IMA.gz' as dicom
** failed to read DICOM file 963 of 967,
... 100%
++ found 0 DICOM files
** found no DICOM files to process
-------
I also tried the following two to3d options:
scnlab% to3d -time:zt 32 160 2s alt+z -assume_dicom_mosaic "*.IMA.gz"
++ to3d: AFNI version=AFNI_2009_12_31_1431 (Jan 26 2010) [32-bit]
++ Authored by: RW Cox
++ Counting images: total=160 2D slices
** Number of slices on command line = 160
** Number of slices needed for -time: = 5120
** Something is wrong with your command line!
I thought that the "-assume_dicom_mosaic" option would make the command assume each .IMA file was a 3d dataset. I also tried
scnlab% to3d -epan 3Dd:0:0:64:64:32:"*IMA.gz"
++ to3d: AFNI version=AFNI_2009_12_31_1431 (Jan 26 2010) [32-bit]
++ Authored by: RW Cox
++ Counting images: total=5120 2D slices
** ERROR: image file WAC2006-007.*.IMA.gz is 172447 bytes long but must be at least 1048576 bytes long
for hglobal=0 himage=0 nx=64 ny=64 nz=32 and voxel=8 bytes
couldn't open image file
** Program compile date = Jan 26 2010
But here again I am unsure of how many bytes I should skip at the beginning and end for the "hglobal" and "himage" options.
Thanks!
-Andrew