Back to the main page.

Bug 2108 - error_in ft_sourceinterpolate

Reported 2013-04-11 14:33:00 +0200
Modified 2014-03-06 15:38:36 +0100
Product: FieldTrip
Component: inverse
Version: unspecified
Hardware: PC
Operating System: Windows
Importance: P3 normal
Assigned to: Jan-Mathijs Schoffelen
Depends on:
See also:

Nadia Kalogianni - 2013-04-11 14:33:33 +0200

Error in running ft_sourceinterpolate as depicted below, after using ft_sourceanalysis with mne method. Error using interpn (line 154) Wrong number of input arguments or some dimension of V is less than 2. Error in ft_sourceinterpolate>my_interpn (line 495) av(sel) = interpn(fv, ax(sel), ay(sel), az(sel), interpmethod); Error in ft_sourceinterpolate (line 398) av( sel) = my_interpn(fv, ax(sel), ay(sel), az(sel), cfg.interpmethod,;

Roemer van der Meij - 2013-04-11 15:13:39 +0200

*** Bug 2107 has been marked as a duplicate of this bug. ***

Jan-Mathijs Schoffelen - 2013-05-01 13:13:57 +0200

Please add a bit more information here. You report only the low level error. It would be very useful if you upload the script, i.e. the code that generates the cfg for the call to ft_sourceinterpolate. Also, please upload some small example data that allows us to reproduce.

Nadia Kalogianni - 2013-05-01 13:35:53 +0200

Created attachment 472 input for source interpolate.

Nadia Kalogianni - 2013-05-01 13:43:02 +0200

(In reply to comment #2) I've added as an attachment the input to ft_sourceinterpolate that produced the error. I used the tutorial for the Minimum Norm Estimate with my data(EEG, 128 channels). I am sorry, I thought I've added the attachment at the first post.

Jan-Mathijs Schoffelen - 2013-05-01 15:02:43 +0200

OK. This is better. At present ft_sourceinterpolate does not allow for 4D data (i.e. data defined on a 3D grid + a time dimension) to be interpolated onto a high resolution anatomy. This is partly to protect you against yourself. Note that by doing the interpolation, you will end up with a matrix that has dimensionality: 256^3 x 287 (287 being your number of time points) I doubt whether you want that.

Jan-Mathijs Schoffelen - 2013-05-01 15:07:55 +0200

added to this, there shouldn't be a principled reason why it should not work. I noticed a bug in the code that prevents it to work in the first place. Yet, I'd recommend downsampling the anatomy first.

Jan-Mathijs Schoffelen - 2013-10-09 16:56:41 +0200

Hi Nadia, I think this is still an issue, but I am not sure. Does it work when you downsample the data, i.e. read in the mri and call ft_volumedownsample on this mri?

Jan-Mathijs Schoffelen - 2014-01-13 11:39:43 +0100

Hi Nadia, See my previous comment. Should we still address this?

Jan-Mathijs Schoffelen - 2014-01-15 16:12:56 +0100

Given the lack of response, I suggest to change status for now: we can always re-open if needed.