Back to the main page.

Bug 2084 - ft_sourceanalysis should give warning if crsspctrm is missing

Reported 2013-03-30 08:09:00 +0100
Modified 2016-05-09 08:48:59 +0200
Product: FieldTrip
Component: core
Version: unspecified
Hardware: PC
Operating System: Mac OS
Importance: P3 normal
Assigned to: Jan-Mathijs Schoffelen
Depends on:
See also:

Robert Oostenveld - 2013-03-30 08:09:29 +0100

when using ft_sourceanalysis for frequency domain beamforming, it does not object against freqanalysis input that was obtained with cfg.output=’pow’. It just ignores the absent crsspctrm. I know the user has the major responsibility here, but it would be easy to at least give a warning.

Robert Oostenveld - 2013-03-30 08:15:55 +0100

I suggest you go ahead and help out with creating a test script for the specific situation. Probably you have the data right at your fingertips at this moment (otherwise you would not have noticed), for me it requires at least 15 minutes to re-create the situation that is required for testing. Please provide a test_bug2048.m script and (when needed) a bug2048.mat file. The mat file can contain some example data (please keep it small) and the test script the cfg and pipeline of function calls.

Robert Oostenveld - 2013-03-30 08:16:25 +0100

(In reply to comment #1) sorry, bug2048 should have read bug2084.

Jan-Mathijs Schoffelen - 2016-02-19 14:21:26 +0100

With the current code, the low-level function prepare_freq_matrices throws an error that states that the cad matrix is incomplete.