[FieldTrip] Problem with Beamformer demo

[Zoltan Juhasz]

Sophie,
Magic. It works perfectly with the new version, and finished about 10 times
faster. Perhaps it reused data still in the Matlab workplace, but the
important thing is that it gives correct results.

Thanks again,
Zoltan

On Wed, Mar 4, 2020 at 1:48 PM Zoltan Juhasz <z65.juhasz at gmail.com> wrote:

> Dear Sophie,
> Thanks fort the update. I'll try again shortly with the latest version of
> fieldtrip. Will let you know about the outcome.
> Zoltan
>
> On Wed, 4 Mar 2020, 13:44 Arana, S.L. (Sophie), <S.Arana at donders.ru.nl>
> wrote:
>
>> ​​​Dear Zoltan,
>>
>>
>> I have indeed recently updated the beamformer tutorial. I just ran it
>> with Matlab 2019a and the most recent Fieldtrip version (available on
>> Github) and the figures look fine to me.
>>
>> Could you update your Fieldtrip version and try again?
>>
>>
>> Best,
>>
>> Sophie
>> ------------------------------
>> *From:* fieldtrip <fieldtrip-bounces at science.ru.nl> on behalf of Zoltan
>> Juhasz <z65.juhasz at gmail.com>
>> *Sent:* Wednesday, March 4, 2020 11:38 AM
>> *To:* FieldTrip discussion list
>> *Subject:* Re: [FieldTrip] Problem with Beamformer demo
>>
>> Hi Jan-Mathijs,
>>
>> Thanks for getting back to me. I noticed that the documentation page has
>> changed since my first email, so I created a new script and ran this again.
>> This time no data is loaded from disk during computation, I am using what
>> is produced by the script. The result is very disappointing, all figures
>> are wrong. I have uploaded the script, the execution log and the output
>> figures. Could you please check what could be wrong here?
>>
>> The link to the files is the same as before:
>> https://drive.google.com/drive/folders/1XPZd5l3og1WAEt8dK6Bx7a9W8xgSzCMF?usp=sharing
>> Matlab version: r2019a
>> Fieldtrip version: 20190121
>>
>> Thanks,
>> Zoltan
>>
>>
>> On Thu, Feb 27, 2020 at 10:00 AM Schoffelen, J.M. (Jan Mathijs) <
>> jan.schoffelen at donders.ru.nl> wrote:
>>
>>> Hi Zoltan,
>>>
>>>
>>> Have you checked the values of the numbers in the functional data prior
>>> to calling ft_sourceinterpolate? This would give an indication of whether
>>> it is an interpolation issue or not.
>>> I suspect that latter, because it seems that the interpolation is
>>> pulling ‘activity values’ into areas in the volume for which no functional
>>> estimate was provided.
>>> This could happen if the specification of the ‘inside’ field in the
>>> source variable, prior to interpolation, behaves unpredictably.
>>> Whether or not this is specific to the Matlab version used, I don’t
>>> know. Also, it is unclear what version of fieldtrip you are using.
>>>
>>> What I can see in your script, though, is that the sourcePost_nocon you
>>> use to generate the first ‘correct’ 2 figures is loaded from disk, whereas
>>> the other source structure (with the ‘incorrect’ interpolation) is not.
>>>
>>> Best wishes,
>>> Jan-Mathijs
>>>
>>>
>>> On 25 Feb 2020, at 13:46, Zoltan Juhasz <z65.juhasz at gmail.com> wrote:
>>>
>>> I have recently run into a problem while executing the Beamformer
>>> Fieldtrip demo at http://www.fieldtriptoolbox.org/tutorial/beamformer/
>>>
>>> I am using the sample data set and the source code from this page but
>>> the picture I get for the section "Source Analysis: Contrast activity to
>>> another interval" is rather strange. I include here the 3 outputs, first
>>> two are correct, third one is the problematic one. Also included is the
>>> source. Matlab version used: R2019a. Could someone check whether it is a
>>> script error or perhaps a change in Matlab that generates the odd result?
>>>
>>> links to figures and source:
>>>
>>> https://drive.google.com/drive/folders/1XPZd5l3og1WAEt8dK6Bx7a9W8xgSzCMF?usp=sharing
>>>
>>> Thanks,
>>> Zoltan
>>> University of Pannonia, Hungary
>>> _______________________________________________
>>> fieldtrip mailing list
>>> https://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>>> https://doi.org/10.1371/journal.pcbi.1002202
>>>
>>>
>>> _______________________________________________
>>> fieldtrip mailing list
>>> https://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>>> https://doi.org/10.1371/journal.pcbi.1002202
>>>
>> _______________________________________________
>> fieldtrip mailing list
>> https://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>> https://doi.org/10.1371/journal.pcbi.1002202
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20200304/f6a26342/attachment.htm>