[FieldTrip] databrowser and merged data

Tue Aug 9 14:11:34 CEST 2011

Hi Eelke,

I have read your mail and would like to emphasize that I disagree. If a 
sampleinfo field is constructed that cannot be used to infer anything 
from, it should not be done. Thus, the fix should not be in 
fixsampleinfo, because it will result in all kind of other issues when 
it comes to reconstructing sampleinfo, although it might help to solve 
this particular problem.  Let's discuss this at the next FT meeting 
rather than here :)

Best,
Jörn

On 8/9/2011 12:30 PM, Eelke Spaak wrote:
> PS: In other words, fixsampleinfo now generates an ad-hoc sampleinfo
> with no relation to the original dataset, if requested to do so.
> Marco's preprocessing steps are not relevant for this.
>
> 2011/8/9 Eelke Spaak<eelke.spaak at donders.ru.nl>:
>> Hi Jörn,
>>
>> As you can read in my email, I actually changed fixsampleinfo so that
>> it *does* reconstruct the sample info now :)
>>
>> I agree that sampleinfo does not make sense in the case of data that
>> is (1) timelock-averaged and (2) combined from multiple datasets, but
>> the issue here was to get the databrowser working on such data. Since
>> the data used internally by the databrowser (after the call to
>> ft_checkdata) is never exposed to the user, this seems like a good
>> solution to me.
>>
>> Best,
>> Eelke
>>
>> 2011/8/9 "Jörn M. Horschig"<jm.horschig at donders.ru.nl>:
>>> Hi Marco and Eelke,
>>>
>>> sampleinfo cannot be reconstructed after resampling the data, because there
>>> is no way to precisely relate your resampled samples to the original
>>> samples. My guess would be that you, Marco, did that and therefore encounter
>>> the problem you mentioned. If that was so, I am not sure whether
>>> fixsampleinfo should actually recompute the sampleinfo field, because it
>>> makes absolutely no sense to have a sampleinfo field in your resampled data
>>> that relates back samples in the original data.
>>> Anyway, if you did not resample your data, then fixsampleinfo should work.
>>> Could you let us know how you preprocessed your data?
>>>
>>> Best,
>>> Jörn
>>>
>>> On 8/9/2011 11:00 AM, Eelke Spaak wrote:
>>>> Hi Marco,
>>>>
>>>> The subfunction fixsampleinfo (used by ft_checkdata) attempts a number
>>>> of things in order to create a sampleinfo field. My guess is that it
>>>> detected a trial definition somewhere in your cfg "tree" (the data.cfg
>>>> and data.cfg.previous fields) that was inconsistent with your data,
>>>> and therefore refused to continue rebuilding a sampleinfo from
>>>> scratch. I have now changed this function so that it will always
>>>> reconstruct your sampleinfo, even if it detects an inconsistent trial
>>>> definition in a previous configuration structure.
>>>>
>>>> So, it should now work :) At least, with tomorrow's update on the FTP
>>>> server.
>>>>
>>>> Best,
>>>> Eelke
>>>>
>>>> 2011/8/8 Marco Dahmane<marco.dahmane at gmail.com>:
>>>>> Hi Eelke,
>>>>> So in order :)
>>>>> 1) Fair enough. And I'm not using cfg.keeptrials='yes' anyway
>>>>> 2) I'm actually using databrowser because I find it really convenient to
>>>>> just remove and add channels, or see to which channel each line
>>>>> corresponds.
>>>>> I know I could do this using Matlab's plot, but I was just wondering why
>>>>> it
>>>>> wouldn't work with databrowser.
>>>>> 3) Ok this is weird then. Something that will probably help you, whenever
>>>>> I
>>>>> use databrowser on my timelocked data, I get this :
>>>>> Warning: the trial definition in the configuration is inconsistent with
>>>>> the
>>>>> actual data
>>>>>> In utilities/private/warning_once at 75
>>>>>    In utilities/private/fixsampleinfo at 51
>>>>>    In ft_checkdata at 579
>>>>>    In ft_databrowser at 155
>>>>> Warning: failed to create sampleinfo field
>>>>>> In utilities/private/warning_once at 75
>>>>>    In utilities/private/fixsampleinfo at 83
>>>>>    In ft_checkdata at 579
>>>>>    In ft_databrowser at 155
>>>>>   So there must be something wrong right?
>>>>> And yes I'm using the latest FT...
>>>>> Many thanks,
>>>>> --Marco
>>>>> _______________________________________________
>>>>> fieldtrip mailing list
>>>>> fieldtrip at donders.ru.nl
>>>>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>>>>>
>>>> _______________________________________________
>>>> fieldtrip mailing list
>>>> fieldtrip at donders.ru.nl
>>>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>>>
>>> --
>>> Jörn M. Horschig
>>> PhD Student
>>> Donders Institute for Brain, Cognition and Behaviour
>>> Centre for Cognitive Neuroimaging
>>> Radboud University Nijmegen
>>> Neuronal Oscillations Group
>>>
>>> P.O. Box 9101
>>> NL-6500 HB Nijmegen
>>> The Netherlands
>>>
>>> Contact:
>>> E-Mail: jm.horschig at donders.ru.nl
>>> Tel:    +31-(0)24-36-68493
>>> Web: http://www.ru.nl/donders
>>>
>>> Visiting address:
>>> Trigon, room 2.30
>>> Kapittelweg 29
>>> NL-6525 EN Nijmegen
>>> The Netherlands
>>>
>>> _______________________________________________
>>> fieldtrip mailing list
>>> fieldtrip at donders.ru.nl
>>> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
>>>
> _______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip

-- 
Jörn M. Horschig
PhD Student
Donders Institute for Brain, Cognition and Behaviour
Centre for Cognitive Neuroimaging
Radboud University Nijmegen
Neuronal Oscillations Group

P.O. Box 9101
NL-6500 HB Nijmegen
The Netherlands

Contact:
E-Mail: jm.horschig at donders.ru.nl
Tel:    +31-(0)24-36-68493
Web: http://www.ru.nl/donders

Visiting address:
Trigon, room 2.30
Kapittelweg 29
NL-6525 EN Nijmegen
The Netherlands