Back to the main page.

Bug 2767 - improve support for averaged neuroscan fif files

Status CLOSED INVALID
Reported 2014-11-25 12:43:00 +0100
Modified 2015-02-11 10:46:48 +0100
Product: FieldTrip
Component: fileio
Version: unspecified
Hardware: PC
Operating System: Mac OS
Importance: P5 enhancement
Assigned to: Robert Oostenveld
URL:
Tags:
Depends on:
Blocks:
See also:

Robert Oostenveld - 2014-11-25 12:43:32 +0100

Woojin Choi (CC) approached me by email regarding a fif file that has been processed (in this case averaged over trials) by Elekta software, asking whether it is possible to process this data further in FieldTrip. That should be possible, but I realise that 1) the file causes problem when reading 2) we don't have documentation about reading processed data. Re 2: we do this a lot for BESA, but usually not for other systems, although it should also work for neuroscan *.avg etcetera.


Robert Oostenveld - 2014-11-25 12:49:04 +0100

I made a test script which in principle should just work (but does not). mac011> svn commit test_bug2767.m Adding test_bug2767.m Transmitting file data . Committed revision 9986.


Robert Oostenveld - 2015-01-19 14:44:20 +0100

roboos@mentat001> svn commit test/test_bug* Sending test/test_bug2767.m Transmitting file data . Committed revision 10108. I fixed the path to the test file in the test script. No progress on the issue itself,


Robert Oostenveld - 2015-01-26 10:56:07 +0100

I had a close look at the fif file. The problem arises in fieldtrip/external/mne/fiff_read_evoked_all.m around line 200, where if aspect_kind == FIFF.FIFFV_ASPECT_AVERAGE does not hold. Subsequently the variable "data" does not get assigned. This is actually caught in ft_read_header, causing hdr.nSamples to become zero. So the fieldtrip representation of the evoked data (having 0 samples) is correct. I made the warning and error reporting a bit more informative. But there is nothing in the code I can fix to resolve it. I suggest you look into teh averaging procedure once more. Note that the external/mne code is not maintained by us, so if you believe that to be incorrect, you have to contact the MNE developers. best regards, Robert roboos@mentat001> svn commit Sending fileio/ft_read_data.m Sending fileio/ft_read_header.m Sending test/test_bug2767.m Transmitting file data ... Committed revision 10120.


Robert Oostenveld - 2015-02-11 10:44:55 +0100

Closed several bugs that were recently resolved. If you don't like the resolution, please reopen and explain.


Robert Oostenveld - 2015-02-11 10:46:48 +0100

Closed several bugs that were recently resolved. If you don't like the resolution, please reopen and explain.