Back to the main page.
Bug 930 - test_ft_preprocessing fails for yokogawa160
| Status | CLOSED FIXED |
| Reported | 2011-09-03 18:16:00 +0200 |
| Modified | 2011-11-11 11:19:40 +0100 |
| Product: | FieldTrip |
| Component: | core |
| Version: | unspecified |
| Hardware: | PC |
| Operating System: | Mac OS |
| Importance: | P1 normal |
| Assigned to: | Stephen Whitmarsh |
| URL: | |
| Tags: | |
| Depends on: | |
| Blocks: | |
| See also: |
Robert Oostenveld - 2011-09-03 18:16:03 +0200
Tilmann Sander-Thommes - 2011-10-17 17:32:51 +0200
Please attach "test_ft_preprocessing.m" here. I work presently to fix Yokogawa stuff. Yours, Till
Jan-Mathijs Schoffelen - 2011-10-17 19:41:40 +0200
Hi Till, The idea would be that upon the call to ft_read_header, where the gradiometer description is created by reading in the file's header info, the description of the gradiometer array is according to the new convention. See: http://fieldtrip.fcdonders.nl/faq/how_are_electrodes_magnetometers_or_gradiometers_described At present, the low-level reading functions still represent the gradiometer array in the old-fashioned way, and rely on fixsens and channelposition (both private functions) to convert this representation. My suggestion would be to avoid this altogether, and represent the hdr.grad correct to begin with. This should be done in yokogawa2grad. Furthermore, in order to support backward compatibility, i.e. allowing to use data which is already on disk, we could consider writing a separate 'elsif' clause in channelposition for the yokogawa system.
Tilmann Sander-Thommes - 2011-10-18 11:28:22 +0200
Hi Jan-Mathijs, I will change yokogawa2grad to reflect the new convention. I do not understand your remark "In order to support backward compatibility, we could consider writing a separate 'elsif' clause in channelposition for the yokogawa system." Where exactly in channelposition is this 'elseif' needed ? Yours, Till
Tilmann Sander-Thommes - 2011-10-20 12:07:59 +0200
Hi Robert and Jan-Mathijs, Please respond to this comment soon, if possible. I have read the content of "how_are_electrodes_magnetometers_or_gradiometers_described" and fully agree, very useful approach. Before implementing this for yokogawa I need some more detail: - Definition of chanpos = mean of coil positions ? - Definition of chanori = orientation of coil closer to head ? These definitions yield for magnetometers with a single coil the following: coilpos = chanpos and coilori = chanori. Is that correct ? Yours, Till
Jan-Mathijs Schoffelen - 2011-10-20 12:15:43 +0200
Dear Till, For the ctf-system we use the convention that the chanpos is the location of the bottom coil, i.e. the coil closest to the head. Same goes for the chanori. Indeed for magnetometers the chanpos and chanori are the same as the corresponding coilpos and coilori. BW, JM
Tilmann Sander-Thommes - 2011-10-21 17:21:26 +0200
I paste an e-mail from Robert here relating to this bug. So bug can be closed. **** Hi Till, This should now be fixed twofold: - channelposition now already work fine with with mixed gradiometers and magnetometers - the grad does not contain the refs any more, i.e. it only consists of gradiometers Robert