[FieldTrip] We struggle with an obsolete add-on-toolbox.

[Schoffelen, J.M. (Jan Mathijs)]

Dear Rasmus,

This type of questions are best asked on the fieldtrip discussion list, more info about this is on: https://www.fieldtriptoolbox.org/discussion_list/

The reasons for me forwarding this message are:
- there are multiple people reading along, so many more brains can help out to provide an answer
- rather than dealing with problems on a one-to-one basis, the time I spend on answering questions/fixing problems is much better spent, because 1) other people will learn from it (so they don’t need to ask the question again), and 2) the list’s archive serves as reference documention, which people can search themselves, and 3) I still have the somewhat vain hope that other people might also occasionally pick up the challenge to answer some of the questions, or to chime in into the discussion.

Just to already point you into the direction of an answer, but perhaps other people might want to chime in: I would recommend against using the DMLT toolbox, because it has not actively been supported for a long time. It could be that the basic functionality is still there, but it is not guaranteed. I would assume that the functionality that is highlighted in the tutorial that you refer to should still work, provided your overall MATLAB environment has correctly been set up.


In relation to the MATLAB environment, and specific to the error message that you pasted, it seems that your MATLAB-path is a bit messed up. This resulted in a higher level fieldtrip function (ft_checkdata) to use a copy of a mex-file for the nanmean function, which is located in the dmlt toolbox. Since this version of the file is out of the control of the fieldtrip maintainers, it is as of yet unclear whether or not the dmlt toolbox will provide you with the functionality that you need. I notice that in the tutorial an addpath(genpath('/your-path-to-fieldtrip/external/dmlt’), is executed, which is probably the culprit here, because it adds dmlt with all its subfolders to the matlab path, including the folder that contains the mex-file that causes the error message that you reported.

You could do 2 things at this point: either debug your path settings, and see whether you can get the dmlt tutorial to run, or step away from the dmlt toolbox altogether. If you want to apply multivariate analysis in fieldtrip, I recommend to start with: https://www.fieldtriptoolbox.org/tutorial/mvpa_light/. Note that this also relies on an external toolbox (so long term support is not guaranteed), but at the moment it seems to work.

Best wishes,

Jan-Mathijs Schoffelen




On 20 Sep 2022, at 11:50, Rasmus Ahmt Hansen <sfk384 at alumni.ku.dk<mailto:sfk384 at alumni.ku.dk>> wrote:

Dear Dr J.M. Schoffel,

My name is Rasmus Ahmt Hansen and I am a master's student at the University of Copenhagen.
My supervisor is d Dr. M. Topor.
My supervisor and I work in the fieldtrip-toolbox to analyze our EEG data, and want to perform a Multivariate analysis.

We follow this guide
" https://www.fieldtriptoolbox.org/tutorial/multivariateanalysis/<https://urldefense.com/v3/__https://www.fieldtriptoolbox.org/tutorial/multivariateanalysis/__;!!HJOPV4FYYWzcc1jazlU!6OXnWR3a6VfPRB-WI0_5advQdRDCsJ0PjX5U63gjPQe5IFduIeG7dxi8sQT5ndrAWiSpaS72hCjDqIpUSgoh3SVvYhvd$> "

We encounter problems as the guide uses the DMLT toolbox, which is obsolete, this is confirmed by Marcel van Gerven who was kind enough to help me download it 😊

And the toolbox is developed to work on MATLAB distribution 7.6.0.324 (R2008a)

We encounter this error massage:

[cid:image001.png at 01D8CCE7.2FED1D50]

The reason I'm writing is because I'm hoping you might be able to give me/us some advice to make the tutorial work or if we have to find an alternative.?


Best wishes Rasmus Ahmt Hansen ^_^

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20220920/cbab49df/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 170825 bytes
Desc: image001.png
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20220920/cbab49df/attachment-0001.png>