Back to the main page.

Bug 1464 - "dimord" field

Reported 2012-05-10 09:34:00 +0200
Modified 2012-08-23 10:49:05 +0200
Product: FieldTrip
Component: core
Version: unspecified
Hardware: PC
Operating System: Windows
Importance: P3 normal
Assigned to: Jan-Mathijs Schoffelen
Depends on:
See also:

Yoni Levy - 2012-05-10 09:34:03 +0200

Please note that if one runs for instance ft_singleplotTFR(cfg,x), and in the 'x' inserts x.dimord=freq_time, as done before -- there is an error message returning "index exceeds matrix dimensions". I think it should be noted to the users that a change to the code has been made regarding the field "dimord".

Jan-Mathijs Schoffelen - 2012-05-10 20:29:13 +0200

Hi Yoni, Could you be a bit more specific, and upload some data that reproduces the problem?

Yoni Levy - 2012-05-11 10:23:33 +0200

Created attachment 264 stat file

Yoni Levy - 2012-05-11 10:24:33 +0200

Hi Jan-Mathijs, I just noticed that my previous script, which ran: " . . . stat=ft_freqstatistics(cfg,freq) stat.dimord='freq_time'; . . . ft_singleplotTFR(cfg,stat) " - is returning an error message: "index exceeds matrix dimensions". This error message is not returned if I omit stat.dimord='freq_time'; So I think that somewhere the code for one of the two functions was changed? but that it is not clear to the user what to do with the dimord field I attached the "stat" file in the previous comment. (In reply to comment #1)

Jan-Mathijs Schoffelen - 2012-07-17 09:40:50 +0200

I don't know about the old behavior, but indeed the dimord should be 'chan_freq_time', because the dimensionality of the numeric data in this case is: 1xNfreqxNtime. Had the dimensionality been NfreqxNtime, the dimord should have been 'freq_time'. I don't know what was going on with your data, but to me this just seems to work fine. Agreed?

Jan-Mathijs Schoffelen - 2012-07-17 09:41:13 +0200


Yoni Levy - 2012-07-17 10:14:31 +0200

(In reply to comment #4) Agreed. Yoni