Hi, I was trying to run the pestica on the sample dataset, before I move on to my own data, but I keep getting an error message in the 2nd step:
FATAL ERROR: can't open --source dataset  .../resp_mean_mni.nii' :-( and
FATAL ERROR: can't open --source dataset '.../card_mean_mni.nii' :-(

(I'm not including the full path here)

I was trying to look through the code to see where the resp_mean_mni.nii and card_mean_mni.nii sre created but seem to find it.
I would really appreciate any help or suggestions
Ekaterina

Hi Ekaterina,
did you first edit the setup script and run it from the PESTICA directory?  There is a script "setup_pestica.sh" that you have to modify to contain the correct MATLAB_AFNI_DIR (the location of the AFNI MATLAB distribution, unzipped) and PESTICA_DIR (the location where you downloaded and unzipped pestica), and then you have to select the right PESTICA_VOL - the default is usually the right one (16mm^3), and then you have to run this script once from the pestica directory.  Doing this will setup softlinks to the appropriate resp/card_mean_mni.nii files.  Actually, it looks like you have done this, but possibly with an incorrect PESTICA_DIR setup.  The 3dAllineate command that is called from run_pestica_est.sh looks in $PESTICA_DIR/ for the resp_mean_mni.nii file and it looks like you have PESTICA_DIR set to "..".  Can you tell me what this variable is set to?  It should be the location of your pestica distribution.  I would also like to know, do you also have a resp_mean_mni.nii file in your pestica distribution?
Erik

Hi Erik, Thank you for your response. I did edit and sourced the setup file, I also got the first step run_pestica_ica.sh to run without an error. Do I understand correctly that the  resp_mean_mni.nii is part of the distribution? it wasn't part of the package I downloaded. Here's the list of files that was in the pestica_afni_v1_2:
apply_PESTICA2.m          
meanepi_mni.nii
assemble_slices_to_timeseries.m   
optimize_pmu.m
card_mean_mni.nii             
prepare_ICA_decomp.m
disassemble_timeseries_to_slices.m
README                   

resp_mean_mni.nii
epi_132vols.nii                  
run_irfret.sh
epi_mask.nii                    
run_pestica_est.sh
example_physio.asv     
 run_pestica_ica.sh
example_physio.m                 
setup_pestica.sh
filter_pestica_est.sh           
tfilter_fft.m
get_correlation_threshold.m       
view_and_correct_estimator.m
get_slice_timing.m

Also, here's the full error including the complete path:



And the  PESTICA_DIR is set to: PESTICA_DIR=/net/synapse/nt/ekat/test_pestica/pestica_afni_v1_2
 FATAL ERROR: can't open --source dataset '/net/synapse/nt/ekat/test_pestica/pestica_afni_v1_2/resp_mean_mni.nii' :-(
FATAL ERROR: can't open --source dataset '/net/synapse/nt/ekat/test_pestica/pestica_afni_v1_2/card_mean_mni.nii' :-(

thank you for your help
Ekaterina

Hi Ekaterina,
okay, good, but now I'm in a quandary - you listed resp_mean_mni.nii in the list of files inside pestica_afni_v1_2, but 3dAllineate cannot load it.  Can you post the output of "ls -l resp_mean_mni.nii" from your pestica directory?  It should be a softlink to a file that does exist (most likely pestica_volumes/resp_pestica_16mm3_avg_mni.nii).  I see you have a couple of other files in there - I recommend you don't run any PESTICA tools from within the PESTICA directory, but instead make a new directory somewhere else in your filesystem with the example data moved there.

erik

Hi Erik, thank you for being patient.
Here's the response to the command you mentioned:
ls -l resp_mean_mni.nii
lrwxrwxrwx 1 ekat bmacgrp 98 2012-07-19 15:43 resp_mean_mni.nii -> /net/synapse/nt/ekat/test_pestica/pestica_afni_v1_2/pestica_volumes/resp_pestica_64mm3_avg_mni.nii

The softlink does seem to be there but I still ge the same error when running the program on a sample database from the directory different from the PESTICA_DIR. Do you think it could be a problem with the link?

Thats even weirder, that link looks fine.  In run_pestica_est.sh, it actually runs three separate 3dAllineate commands - the first one is the coregistration between the MNI EPI template and your data (and which should produce a file named epi_mask.crg2mni.nii, assuming your masked EPI data is named epi_mask.nii), and the second is applying this transform to the respiratory noise template and the third is doing the same for the cardiac noise template.  Do any of these following commands give an error or do they complete normally?
3dinfo /net/synapse/nt/ekat/test_pestica/pestica_afni_v1_2/meanepi_mni.nii
3dinfo /net/synapse/nt/ekat/test_pestica/pestica_afni_v1_2/resp_mean_mni.nii
3dinfo /net/synapse/nt/ekat/test_pestica/pestica_afni_v1_2/card_mean_mni.nii
3dinfo /net/synapse/nt/ekat/test_pestica/pestica_afni_v1_2/pestica_volumes/resp_pestica_64mm3_avg_mni.nii

Finally, can you also check the coregistration of your EPI to MNI space using (if the above worked):
afni -dset /net/synapse/nt/ekat/test_pestica/pestica_afni_v1_2/meanepi_mni.nii  epi_mask.crg2mni.nii   epi_mask.nii    

Basically, the epi_mask.crg2mni.nii should line up reasonably well over the meanepi_mni.nii when you put one as the overlay.  Again, I've assumed the name of your filename, so you may need to modify that.
Erik

Out of the 4 3dinfo commands, only the first runs without errors. The others had an error similar to
3dinfo /net/synapse/nt/ekat/test_pestica/pestica_afni_v1_2/card_mean_mni.nii
++ 3dinfo: AFNI version=AFNI_2011_05_26_1457 (Jun 22 2011) [64-bit]
** FATAL ERROR: Can't open dataset /net/synapse/nt/ekat/test_pestica/pestica_afni_v1_2/card_mean_mni.nii
** Program compile date = Jun 22 2011

and the registration seems to be doing a pretty good job
Kat

Okay, theres definitely something wrong with the links.  I'd like you to re-run the setup script, but to do so, you have to delete the link files and the .firsttime file that tells the setup script whether to generate the links or not.  To do this, you will need to run (from the PESTICA distribution directory) "rm .firsttime resp_mean_mni.nii card_mean_mni.nii".  That will delete the two link files and the .firsttime file.  Then run ./setup_pestica.sh to re-make the links.  From this directory, test the link by running "3dinfo resp_mean_mni.nii".  That ought to work, but if it doesn't there may be some permissions issue specific to your computing setup.  If that doesn't work, post the output of "ls -l" on the PESTICA dir and the pestica_volumes subdirectory.
erik