Petr why do you keep trying to name your input files something.mgz, when you were in wisconsin i told you the inputs always need to be 001.mgz

at this point i think you need to get invited to upload one of your data sets to ugh and see if there is a problem with the data or is it a problem with your computer.

Can someone there at martinos invite Petr to upload the dataset and tell him the mechanism for doing this ?

On 01/18/13, zach.petr@post.cz wrote:

I tried on another standar PC (4 core processor) with XP Win recon-all in VirtualBox (latest edition) running Ubuntu.

I attach file folder with patient brain - first in VirtualBox as described and then in command line of terminal Ubuntu window:

sudo mount - vboxsf folder /home/virtualuser/freesurfer/subjects/folder

Then I convert original TrioTim scanner files using mri_convert -it IMA -ot mgz filename foldername.mgz

• here is slight problem - recon log shows that mri_convert recognizes itself that although my brain files are labeled as IMA that they are not old Siemens files, so that it unpacks them as DICOM and suggests that if it fails I should run mri_convert as DICOM

But after all mri_convert succesfully writes foldername.mgz file as a result of conversion in foldername directory

Ok, then I run recon-all -subjid foldername -autorecon1 -cw256

Now -cw256 was suggested in another error run because files do not fit in 256 matrix

Recon stops and sayz no 001.mgz files in /mri/orig directory

So I copy converted foldername.mgz from foldername directory into /mri/orig directory and rename it to 001.mgz

*now here is my question: do all people get 001.mgz file the same way as me?

Then when I run recon-all -subjid foldername -autorecon1 -cw256 it halts with error at nu_correct as you can see in attached recon-all.log file

Petr