Back to the main page.
Bug 35 - channel selection in ft_connectivityanalysis
| Status | CLOSED FIXED |
| Reported | 2010-02-08 13:02:00 +0100 |
| Modified | 2011-01-06 09:12:19 +0100 |
| Product: | FieldTrip |
| Component: | core |
| Version: | unspecified |
| Hardware: | PC |
| Operating System: | Linux |
| Importance: | P5 normal |
| Assigned to: | Jan-Mathijs Schoffelen |
| URL: | |
| Tags: | |
| Depends on: | |
| Blocks: | |
| See also: |
Andre Bastos - 2010-02-08 13:02:40 +0100
2 things: 1. connectivityanalysis does not support cfg.channelcmb = {'chan1' 'all'}; 2. To do 1 by hand, I manually set cfg.channelcmb = {'chan1' 'chan2'; 'chan1' 'chan3'; 'chan1' 'chan1'; 'chan2' chan2'l 'chan3'; 'chan3'} and call connectivityanalysis, but the channelcmb is inappropriately assigned after calling channelcombination on line 109. Commenting this line out fixes the problem. I am not sure why I was not seeing this problem in my previous analyses, but it could be do to the fact that I am now calculating coherence from chan1 to ALL (instead of just within an ROI). See my full cfg specification in the attachment. Also, could we work on this together? I would like to understand better the channelselection process as it seems to continually come up as a royal pain in the ass... =)
Andre Bastos - 2010-02-08 13:03:01 +0100
Created attachment 2 code
Jan-Mathijs Schoffelen - 2010-02-08 13:15:48 +0100
Don't blame connectivityanalysis, it's probably an unwanted feature of channelcombination. Please come down if you want to look at it together, once you have a vnc'able matlab session open in which the problem can be easily reproduced... (In reply to comment #0) > 2 things: > > 1. connectivityanalysis does not support cfg.channelcmb = {'chan1' 'all'}; > > 2. To do 1 by hand, I manually set cfg.channelcmb = {'chan1' 'chan2'; 'chan1' > 'chan3'; 'chan1' 'chan1'; 'chan2' chan2'l 'chan3'; 'chan3'} and call > connectivityanalysis, but the channelcmb is inappropriately assigned after > calling channelcombination on line 109. Commenting this line out fixes the > problem. I am not sure why I was not seeing this problem in my previous > analyses, but it could be do to the fact that I am now calculating coherence > from chan1 to ALL (instead of just within an ROI). > > See my full cfg specification in the attachment. Also, could we work on this > together? I would like to understand better the channelselection process as it > seems to continually come up as a royal pain in the ass... > > =) >