Fieldtrip analysis of Yokogawa format

Robert Oostenveld R.Oostenveld at FCDONDERS.RU.NL
Wed Apr 26 14:14:15 CEST 2006

Dear Kaoru

On 25 Apr 2006, at 11:12, Kaoru Amano wrote:
> We are using MEG system of Yokogawa, and very much interested in
> source
> estimation of e.g. gamma band activity. According to the Wiki,
> Yokogawa
> format (*.sqe, *.ave, *.con, *.raw) seems to be supported. But the
> following error message appeared and I couldn't load the data.
> "Undefined command/function 'yokogawa2grad'." (Actually, I couldn't
> find
> the file "yokogawa2grad.m").

The file indeed is missing in the ftp version, please find it
atatched. It will be automatically included in future releases on the
ftp server. However, looking in the file I do notice that it is not
yet complete. I suggest that you read
_magnetometers_or_gradiometers_described and that you complete the
missing section of the yokogawa2grad function (indicated with "FIXME").

> Since I can load the data into Matlab, I reformatted the data into
> your
> datastructure (data.label, data.fsample, data.trial, data.time), as
> explained in Wiki "How can I import my own dataformat?" But, of
> course,
> I have to load other data such as sensor position, for the source
> estimation.
> Could you tell me which analysis is possible for Yokogawa's format? If
> we want to do source estimation, how can I import necessary
> information
> to your format?

Almost all analyses can be done on your MEG data. The missing
information on gradiometer locations is (or should be) implemented in
the yokogawa2grad function, see above.

For inverse modelling, you would start with timelockanalysis or
freqanalysys, followed by making topoplots of the resulting data to
see whether the hypothesised effects are visible on the channel
level. Subsequently you would do sourceanalysis. See http://

Sourceanalysis requires a head model, which can be read from various
file formats (but not Yokagawa) or which can be made using the
fieldtrip functions prepare_localspheres or the prepare_singleshell.
Reading anatomical MRIs in yokagawa format is however not yet
supported, so you would have to solve the alignment of the anatomical
MRI and volume model and the sensors yourself.

You could start with a simple single sphere headmodel, which can be
specified in sourceanalysis as
cfg.vol.r = [r]     % radius
cfg.vol.o = [x y z] % origin

best regards,

Robert Oostenveld, PhD
F.C. Donders Centre for Cognitive Neuroimaging
Radboud University Nijmegen
phone: +31-24-3619695

-------------- next part --------------
A non-text attachment was scrubbed...
Name: yokogawa2grad.m
Type: application/octet-stream
Size: 2719 bytes
Desc: not available
URL: <>
-------------- next part --------------

More information about the fieldtrip mailing list