mne_do_forward solution fails reading tag directory

✉️ Mailing List Archive (read-only)
1

When attempting to run mne_do_forward_solution on a subject, forward
computation fails with the message "Could not read tag directory!". Could
anyone tell me what this error message is referring to?

Best wishes,

Vincent Rupp Jr
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20131010/2f7c52e9/attachment.html

2

hi Vincent,

any chance a fif file was corrupted?

Alex

3

I have some old notes from Neuromag on this, which are quoted below. As
Alex suggested, it may be that a copy failed. That was the case for our
subject -- I've never had to actually follow the instructions below. Hope
this is helpful.

Jon

This symptom usually comes when the tail of the file is missing. This
typically happens when somebody copies a file and disk comes full or the
copy is stopped in the middle.

To fix the file: in principle one needs to set the pointer to directory to
-1 to prevent software to try to access the directory, remove all jumps to
nonexisting tail data, add all "closing parenthesis" (end of block markers)
to end of file to make its structure consistent again (graph can open files
that are "open" at the end, so the last step is not absolutely necessary).
If the missing tail is not very short, you are likely to loose digitization,
HPI measurement, patient info and some other pieces that are glued to the
end when the file is saved.

4

Hi Vincent,

These instruction do not apply here but rather to raw data files, which will be then missing part of the data after the fix which is not so bad.

All the files needed by mne_do_forward_solution need to be intact. What is the exact output you get up to the error about fif file reading?

- Matti