[FieldTrip] Help with ft_sourcestatistics input

[Matteo Maran]

Dear Tao and Xavier,

I was also struggling with the .dim field of the cfg structure, and indeed I was getting the same error as the one you reported with the MNE data. From previous Q&A on the mailing list I found the following info which might be useful:

https://mailman.science.ru.nl/pipermail/fieldtrip/2017-June/037494.html
https://mailman.science.ru.nl/pipermail/fieldtrip/2016-July/023582.html

As far as I read from the reply to the first question, you need cfg.dim or cfg.tri, according to the source model you used. Loading the templates from http://www.fieldtriptoolbox.org/template/sourcemodel/ I see that the templates defined on a 3D grid have a .dim field (e.g. standard_sourcemodel3d10mm), but no .tri field. On the other side, source model templates for MNE have a .tri but no .dim field (e.g. sourcemodel = ft_read_headshape('cortex_20484.surf.gii') ). This is then reflected in the field of the data you give as an input to ft_sourcestatistics.

load(standard_sourcemodel3d10mm.mat)
sourcemodel = 
  struct with fields:
       dim: [18 21 18]                    
       pos: [6804×3 double]
    inside: [6804×1 logical]
       cfg: [1×1 struct]
      unit: 'cm'
% Note that 18*21*18 = 6804

sourcemodel = ft_read_headshape('cortex_20484.surf.gii') )
sourcemodel = 
  struct with fields:
     pos: [20484×3 double]
     tri: [40960×3 double]
    unit: 'mm'


Indeed, if you have a look at the second question posted above, you can see that cfg.tri instead of cfg.dim was specified. Maybe this can be of help.


Best wishes,

Matteo

----- Original Message -----
From: "Xavier Vrijdag" <x.vrijdag at auckland.ac.nz>
To: "fieldtrip" <fieldtrip at science.ru.nl>
Sent: Thursday, July 23, 2020 10:27:42 PM
Subject: Re: [FieldTrip] Help with ft_sourcestatistics input

Hi Tao, 



I am not sure about the cfg.neighbours setting, but the error looks like a SPM version issue I had before as well. I solved it by adding: 

cfg.spmversion = 'spm12' ; 



regards, 



Xavier 




From: fieldtrip <fieldtrip-bounces at science.ru.nl> on behalf of Tao Wang <wangtao0331 at tju.edu.cn> 
Reply to: FieldTrip discussion list <fieldtrip at science.ru.nl> 
Date: Friday, 24 July 2020 at 3:26 AM 
To: "fieldtrip at science.ru.nl" <fieldtrip at science.ru.nl> 
Subject: [FieldTrip] Help with ft_sourcestatistics input 





Dear community, 



My name is Tao Wang and I study at Tianjin University. My research direction is to explore ways(tDCS\tACS) to improve working memory. I ’ m looking to perform a cluster-based permutation test on my source data(the result from ft_sourceanalysis). 



I carefully read the function help and the tutorial( http://www.fieldtriptoolbox.org/example/source_statistics/ ). However, I receive the following error message: 



Error using spm_bwlabel 

spm_bwlabel: CONN must be 6, 18 or 26 

Error in clusterstat (line 191) 

[posclusobs, posnum] = spm_bwlabel(tmp, 2*numdims); 

Error in ft_statistics_montecarlo (line 353) 

[stat, cfg] = clusterstat(cfg, statrand, statobs); 

Error in ft_sourcestatistics (line 202) 

[stat, cfg] = statmethod(cfg, dat, design); 



I am running the following: 

cfg=[]; 

cfg.parameter='pow'; 

cfg.method='montecarlo'; 

cfg.statistic = 'ft_statfun_depsamplesT'; 

cfg.correctm = 'cluster'; 

cfg.clusteralpha = 0.05; 

cfg.clusterstatistic = 'maxsum'; 

cfg.tail=0; 

cfg.clustertail=0; 

cfg.alpha=0.025; 

cfg.numrandomization = 1000; 

nsubj = 10; 

cfg.design(1,:) = [1:nsubj 1:nsubj]; 

cfg.design(2,:) = [ones(1,nsubj)*1 ones(1,nsubj)*2]; 

cfg.uvar = 1; % row of design matrix that contains unit variable (in this case: subjects) 

cfg.ivar = 2; % row of design matrix that contains independent variable (the conditions) 



stat = ft_sourcestatistics(cfg, grandavgA{:}, grandavgB{:}); 



My data: 

display (grandavgA{1}) 

time: [1 × 400 double] 

inside: [8196 × 1 logical] 

pos: [8196 × 3 double] 

tri: [16384 × 3 double] 

method: 'average' 

avg: [1 × 1 struct] 

cfg: [1 × 1 struct] 

display(grandavgA{1}.avg) 

mom: {1 × 8196 cell} 

pow: [8196 × 1 double] %I used ‘ 8196 × 1 double ’ instead of ‘ 8196 × 400 double ’ 

noisecov: {1 × 8196 cell} 



Do I need to set cfg.neighbours, like the settings in the tutorial "Cluster-based permutation tests on event related fields"? (I noticed cfg.dim, but it doesn ’ t appear in my source struct) 

Can someone tell me if there is something wrong with the cfg settings I use or if I am doing something wrong at any other place? Any tips anyone has would be very much appreciated! 

Hope that helps! 



Best, 

Tao 

_______________________________________________
fieldtrip mailing list
https://mailman.science.ru.nl/mailman/listinfo/fieldtrip
https://doi.org/10.1371/journal.pcbi.1002202