[FieldTrip] ft_sourcestatistics cannot find field 'pos'

Fri Jan 24 12:40:21 CET 2014

Hi Joram,

Probably this is my bad. 
ft_sourceinterpolate intentionally removes the pos field, which has to do with the representation of the data. FieldTrip either represents source reconstructed data that can be defined on a regular 3D grid as a so-called 'source-structure' (with a pos field), or as a so-called volume-structure (without a pos field). After the sourceinterpolate step your data is represented as the latter, lacking a pos field (intentionally), but unintentionally causing a crash in ft_sourcestatistics.
A workaround for now would be for you to change line 228 in statistics_wrapper into if isfield(varargin{1}, 'transform') || (isfield(varargin{1}, 'dim') && prod(varargin{1}.dim)==size(varargin{1}.pos,1)).
Could you try this out and let me know if that works? Then I can incorporate it in FT.

Best and sorry for the inconvenience,
JM

On Jan 24, 2014, at 12:21 PM, Joram van Driel wrote:

> Sorry, this should have been:
> This 'pos' field is removed in further steps (ft_sourceinterpolate). When calling ft_sourcestatistics in version fieldtrip-20140109, the function statistics_wrapper searches for this field (line 228) and can't find it.
> 
> 
> On Fri, Jan 24, 2014 at 12:16 PM, Joram van Driel <joramvandriel at gmail.com> wrote:
> Hi Nietzsche,
> 
> Thanks for the suggestion, but unfortunately that's not what's going wrong. My input data is a subject-by-condition array, so if I fill in sourcedat{:,1} and sourcedat{:,2}, that would be equivalent to having two separate variables and do source_condition1{:},source_condition2{:}. I tried that but I get the same error "??? Reference to non-existent field 'pos'."
> 
> In fact, the error is I think a bug of the newest fieldtrip version, because when I tried an older version (fieldtrip-20131031), it works (although it later crashes on a design array issue, but that's something I have to figure out myself ;)).
> 
> The 'pos' field is a field that is present in the output of ft_sourceanalysis; according to the help it's a N-by-3 matrix of the x-y-z position of all the sources, where N is the sum of the length of the 'inside' and 'outside' fields.
> This 'pos' field is removed in further steps (ft_sourceinterpolate). When calling ft_sourceanalysis in version fieldtrip-20140109, the function statistics_wrapper searches for this field (line 228) and can't find it.
> 
> Chrs,
> 
> - Joram
> 
> 
> 
> On Fri, Jan 24, 2014 at 9:55 AM, Lam, Nietzsche <n.lam at fcdonders.ru.nl> wrote:
> Hi Joram,
> 
> I'm not entirely sure if this is the solution, but when you call ft_sourcestatistics, you can try this:
> 
> <stat = ft_sourcestatistics(cfg,sourceDiffNorm{:}, sourceDiffNorm{:})>
> 
> FieldTrip statistics functions understands that you want to use the data from all subjects when you use {:}, so there's no need to call individual columns with {:,X}.
> 
> Best,
> Nietzsche
> 
> 
> 
> ----- Original Message -----
> > From: "Joram van Driel" <joramvandriel at gmail.com>
> > To: "FieldTrip discussion list" <fieldtrip at science.ru.nl>
> > Sent: Friday, 24 January, 2014 8:54:24 AM
> > Subject: [FieldTrip] ft_sourcestatistics cannot find field 'pos'
> > Hi all,
> >
> >
> > I'm getting stuck with ft_sourcestatistics.
> > I want to do a simple two-condition contrast on neuromag MEG data,
> > where I did frequency beamforming on a pre vs. post tf-window.
> >
> >
> > I followed the instructions of the tutorial, so for each subject and
> > condition:
> >
> >
> > 1) ft_sourceanalysis with subject-specific vol and grid structures,
> > where I did the pre vs post contrast as follows:
> >
> >
> >
> > sourceAll = ft_sourceanalysis(cfg, freqAll(condi));
> >
> > cfg.grid.filter = sourceAll.avg.filter;
> > sourcePre_con = ft_sourceanalysis(cfg, freqPre(condi) );
> > sourcePost_con = ft_sourceanalysis(cfg, freqPost(condi));
> >
> >
> >
> > sourceDiff(condi) = sourcePost_con;
> > sourceDiff(condi).avg.pow = (sourcePost_con.avg.pow -
> > sourcePre_con.avg.pow) ./ sourcePre_con.avg.pow;
> >
> >
> > 2) ft_sourceinterpolate with the subject-specific mri
> > 3) ft_volumenormalize to MNI with coordsys 'neuromag'.
> > 4) The output is stored in a subject-by-condition cell array, which I
> > put into ft_sourcestatistics with the following cfg:
> >
> >
> >
> > cfg = [];
> > cfg.parameter = 'avg.pow';
> > cfg.method = 'analytic';
> > cfg.statistic = 'depsamplesT';
> > cfg.correctm = 'no';
> > cfg.alpha = 0.05;
> >
> >
> > Nsub = 10;
> > cfg.design(1,1:2*Nsub) = [ones(1,Nsub) 2*ones(1,Nsub)];
> > cfg.design(2,1:2*Nsub) = [1:Nsub 1:Nsub];
> > cfg.tail = 0; % number, -1, 1 or 0 (default = 0)
> > cfg.ivar = 1; % number or list with indices, independent variable(s)
> > cfg.uvar = 2; % number or list with indices, unit variable(s)
> >
> >
> > stat = ft_sourcestatistics(cfg,sourceDiffNorm{:,1},
> > sourceDiffNorm{:,2});
> >
> >
> >
> >
> > This results in the error that it cannot find the field 'pos'; however
> > this field is only present in the result from ft_sourceanalysis (and
> > differs for each subject), but disappears as soon as
> > ft_sourceinterpolate is applied. I tried to put the result from
> > ft_sourceanalysis straight into ft_sourcestatistics (which according
> > to the help should be possible), but this doesn't recognize the input
> > as volume data (and apart from that, the subjects aren't spatially
> > aligned this way).
> >
> >
> > I hope someone can help me with this; any help is much appreciated!
> >
> >
> > Thanks,
> > Joram
> >
> >
> > --
> >
> > Joram van Driel, MSc.
> > PhD student @ University of Amsterdam
> > Brain & Cognition @ Department of Psychology
> > _______________________________________________
> > fieldtrip mailing list
> > fieldtrip at donders.ru.nl
> > http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
> 
> --
> Nietzsche H.L. Lam, MSc
> PhD Candidate
> 
> Max Planck Institute for Psycholinguistics
> Wundtlaan 1, 6525 XD Nijmegen, The Netherlands
> 
> Donders Institute for Brain, Cognition and Behaviour,
> Centre for Cognitive Neuroimaging,
> Kapittelweg 29, 6525EN Nijmegen, The Netherlands
> 
> n.lam at fcdonders.ru.nl
> +31-24-3668219
> 
> 
> neurobiologyoflanguage.com
> _______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip
> 
> 
> 
> -- 
> Joram van Driel, MSc.
> PhD student @ University of Amsterdam
> Brain & Cognition @ Department of Psychology
> 
> 
> 
> -- 
> Joram van Driel, MSc.
> PhD student @ University of Amsterdam
> Brain & Cognition @ Department of Psychology
> _______________________________________________
> fieldtrip mailing list
> fieldtrip at donders.ru.nl
> http://mailman.science.ru.nl/mailman/listinfo/fieldtrip

Jan-Mathijs Schoffelen, MD PhD 

Donders Institute for Brain, Cognition and Behaviour, 
Centre for Cognitive Neuroimaging,
Radboud University Nijmegen, The Netherlands

Max Planck Institute for Psycholinguistics,
Nijmegen, The Netherlands

J.Schoffelen at donders.ru.nl
Telephone: +31-24-3614793

http://www.hettaligebrein.nl

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20140124/4429034f/attachment.html>