Back to the main page.
Bug 3120 - ft_apply_montage breaks on different trial-lengths
| Status | CLOSED FIXED |
| Reported | 2016-05-04 13:17:00 +0200 |
| Modified | 2016-06-14 16:14:53 +0200 |
| Product: | FieldTrip |
| Component: | forward |
| Version: | unspecified |
| Hardware: | PC |
| Operating System: | Windows |
| Importance: | P5 normal |
| Assigned to: | Robert Oostenveld |
| URL: | |
| Tags: | |
| Depends on: | |
| Blocks: | |
| See also: |
Stephen Whitmarsh - 2016-05-04 13:17:29 +0200
Hi there, The last commit (https://github.com/fieldtrip/fieldtrip/commit/20e645fb7068b429a2c2f001a4787e3ac7c02016) added a check for NaNs. However, the use of a 'cat' supposes similar trial lengths, therefor breaking when this is not the case: ****************** Error using cat Dimensions of matrices being concatenated are not consistent. Error in ft_apply_montage (line 245) tmp = cat(1, data_unused.trial{:}); Error in ft_rejectcomponent (line 185) data = ft_apply_montage(data, montage, 'keepunused', keepunused, 'feedback', cfg.feedback); ****************** I think a solution could be to replace line 245&246: tmp = cat(1, data_unused.trial{:}); if any(isnan(tmp(:))) simply with: any(isnan([data_unused.trial{:}])); This works for me. Cheers, Stephen
Robert Oostenveld - 2016-05-04 13:46:31 +0200
The "cat" by eelke is along the wrong direction, it should be along dimension 2. That is also what you are doing it with the []. I am worried about the memory impact of either solution. A copy of the data is made and for large data that is inefficient. I have changed it into % use an anonymous function to test for the presence of NaNs in the input data hasnan = @(x) any(isnan(x(:))); if any(cellfun(hasnan, data_unused.trial)) ... end mac011> git commit fileio/private/ft_apply_montage.m [master b1d9fbc] FIX - see http://bugzilla.fieldtriptoolbox.org/show_bug.cgi?id=3120 1 file changed, 3 insertions(+), 2 deletions(-)