Back to the main page.
Bug 2663 - ft_rejectvisual sorting of minimum values
| Status | CLOSED FIXED |
| Reported | 2014-07-29 16:56:00 +0200 |
| Modified | 2019-08-10 12:28:34 +0200 |
| Product: | FieldTrip |
| Component: | core |
| Version: | unspecified |
| Hardware: | PC |
| Operating System: | Linux |
| Importance: | P5 normal |
| Assigned to: | |
| URL: | |
| Tags: | |
| Depends on: | |
| Blocks: | |
| See also: |
Gio Piantoni - 2014-07-29 16:56:20 +0200
Hi, ft_rejectvisual/summary works really well for most metrics but it does not plot the correct values when you want the minimum values across trials and channels. With metric = 'min', rejectvisual_summary first gets the lowest value for each channel [1], and then it takes the maximum value across trials [2]. It should take the minimum value across trials too. Example data: ChanA = [-5, -2, 3]; ChanB = [-10, -5, 0]; ChanC = [5, 2, -2]; If the metric is min, you get: ChanA = -5, ChanB = -10, ChanC = -2. Then it takes the max value, which is -2, while it should be -10. I suggest to flip the sign before computing the max. So to insert at line 214: if strcmp(info.metric, 'min') level = level * -1; end [maxperchan maxpertrl maxperchan_all maxpertrl_all] = set_maxper(level, info.chansel, info.trlsel); if strcmp(info.metric, 'min') maxperchan = maxperchan * -1; maxpertrl = maxpertrl * -1; maxperchan_all = maxperchan_all * -1; maxpertrl_all = maxpertrl_all * -1; level = level * -1; % see point below end The second part is necessary to have the correct sign in the plots. I'm not sure about the second "level = level * -1", because by flipping the sign again the color palette is the opposite of the color palette for the other metrics (red -> dangerously high values) and there is no colorbar with interpretable values. I can commit the changes if you agree. -g [1] https://github.com/fieldtrip/fieldtrip/blob/master/private/rejectvisual_summary.m#L172 [2] https://github.com/fieldtrip/fieldtrip/blob/master/private/rejectvisual_summary.m#L214
Jan-Mathijs Schoffelen - 2014-09-10 15:47:41 +0200
(In reply to Gio Piantoni from comment #0) You have obviously give this a lot of thought ;-) If you trust this to work well, I would say please commit!
Gio Piantoni - 2014-09-10 17:52:00 +0200
(In reply to Jan-Mathijs Schoffelen from comment #1) Sure, I'll commit the changes (I'll be on vacation for two weeks now, so don't expect it soon). I think I waited at the time to see if somebody else had a better overview than me of the steps in ft_rejectvisual.
Gio Piantoni - 2014-09-24 23:47:13 +0200
fix applied: Sending private/rejectvisual_summary.m Transmitting file data . Committed revision 9839.