[FieldTrip] Source localization statistics for specific ROI using atlas

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

Dear Adi,

As discussed outside the scope of the other readers: could you please provide some additional information with respect to possible diagnostics you performed yourself so far? Also, perhaps you could point to some documentation on the website, or earlier e-mail threads to provide some context?

One thing you could do, is to read through the code of ft_sourcestatistics, and check whether the cfg.roi/cfg.atlas/cfg.inputcoord arguments are used to begin with. If you use the MATLAB debugger and the ‘dbstep’ and ‘dbstep in’ buttons you can recurse into lower level functions, such as ft_checkdata and ft_selectdata, which could be the place where the magic might be supposed to take place.

Best wishes,
Jan-Mathijs

On 14 May 2020, at 14:51, Adi Kilim <adikilim1 at gmail.com<mailto:adikilim1 at gmail.com>> wrote:

Dear community,

I am doing a source analysis on MEG data; I used the LCMV beamformer
to localize on time-locked data, and want to contrast two conditions: faces and vehicles in tempo-occipital regions. I used AAL atlas to do that. The code is running however when I'm looking at the stat struct all the voxel are include inside the analysis and thus the clusters are across different brain regions. I've searched around the tutorial web pages and the mailing list archive,
and didn't find explanation for this or similar problem in fieldtrip similar versions (I'm using fieldtrip-20200409). This is my code: ------- aal = ft_read_atlas('ROI_MNI_V4.nii');

cfgStat = [];

%
cfgStat.atlas        =  aal;
cfgStat.roi          = {'Temporal_Mid_L'}; %Example ROI
cfgStat.avgoverroi   = 'yes'; cfgStat.hemisphere   = 'left';
cfgStat.inputcoord   = 'mni' ;

cfgStat.avgovertime     = 'yes'; %'no'
cfgStat.parameter        ='avg.pow';
cfgStat.method           = 'montecarlo';
cfgStat.statistic        = 'depsamplesT'; cfgStat.correctm         = 'cluster';
cfgStat.clusteralpha     = 0.05;
cfgStat.clusterstatistic = 'maxsum';
cfgStat.minnbchan        = 2;
cfgStat.tail             = 0;
cfgStat.clustertail      = 0;
cfgStat.alpha            = 0.05;
cfgStat.numrandomization = 1000;
cfgStat.neighbours       = neighbours;
cfgStat.design           = designMat;


[stat_mainCond_spcTime] = ft_sourcestatistics(cfgStat, cond1_allSub{:},cond2_allSub{:});
-------
Any ideas on how to solve this atlas source statistics problem (performing statistics on specific ROI and identify the cluster) would be greatly appreciated!
Thanks, Adi

_______________________________________________
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/20200514/5ad4ede6/attachment.htm>