[FieldTrip] Help on ft_statfun_correlationT

[Ronny Ibrahim]

Dear Fellow fieldtrip users,
I hope that you don't mind me emailing you directly enquiring about the procedures of using ft_statfun_correlationT. I have been reading the fieldtrip forum but seems to still don't quite understand fully on how to get the function to run properly. I currently have power spectrum values (2047 frequency points) from 15 electrode channels (which will give a powspctrm size of 15x2047) and I would like to see whether any of the frequency bands (alpha, delta, beta, gamma values) do correlate to behavioral measures such as distress, etc. As it had suggested in the forum
http://mailman.science.ru.nl/pipermail/fieldtrip/2015-February/008950.html
I should create a dummy variable with equal size to powspctrm and fill them with my behavioural value (eg 'distress' lets call it D which is essentially D*ones(15,2047)).
I have created a design matrix as follow :
cfg.design           = [1*ones(1,36) 2*ones(1,36); 1:36 1:36];

as I have 36 replications

Here is the script that I have put together.
%% Conditions
%C1 - actual powerspectrum values of 15 electrodes
%C2 - distress values
C_1 = {TIN1{2} TIN1{4} TIN1{5} TIN1{6} TIN1{7} TIN1{8} TIN1{9} TIN1{10} TIN1{11} TIN2{2} TIN2{4} TIN2{5} TIN2{6} TIN2{7} TIN2{8} TIN2{9} TIN2{10} TIN2{11} TIN3{2} TIN3{4} TIN3{5} TIN3{6} TIN3{7} TIN3{8} TIN3{9} TIN3{10} TIN3{11} TIN4{2} TIN4{4} TIN4{5} TIN4{6} TIN4{7} TIN4{8} TIN4{9} TIN4{10} TIN4{11}};
C_2 = {TFI1{1} TFI1{2} TFI1{3} TFI1{4} TFI1{5} TFI1{6} TFI1{7} TFI1{8} TFI1{9}  TFI2{1} TFI2{2} TFI2{3} TFI2{4} TFI2{5} TFI2{6} TFI2{7} TFI2{8} TFI2{9} TFI3{1} TFI3{2} TFI1{3} TFI3{4} TFI3{5} TFI3{6} TFI3{7} TFI3{8} TFI3{9} TFI4{1} TFI4{2} TFI4{3} TFI4{4} TFI4{5} TFI4{6} TFI4{7} TFI4{8} TFI4{9}};

cfg = [];
cfg.layout = 'MEG.lay';
lay = ft_prepare_layout(cfg);


%% Performing cluster statistics
cfg                  = [];
cfg.frequency        = [8 12];
cfg.channel          = 'all';
cfg.avgovertime      = 'no';
cfg.avgoverfreq      = 'yes';
cfg.avgoverchan      = 'no';
cfg.statistic        = 'ft_statfun_correlationT';
cfg.numrandomization = 1000;
cfg.correctm         = 'cluster';
cfg.method           = 'montecarlo';
cfg.clusteralpha     = 0.05;
cfg.clusterstatistic = 'maxsum';
cfg.minnbchan        = 1;
cfg.tail             = 0;
cfg.clustertail      = 0;
cfg.alpha            = 0.05;
cfg_neighb.layout    = lay;
cfg_neighb.method    = 'triangulation';
cfg.neighbours       = ft_prepare_neighbours(cfg_neighb);
cfg.uvar             = 1;
cfg.ivar             = 1;
cfg.computecritval = 'yes';
cfg.design           = [1*ones(1,36) 2*ones(1,36); 1:36 1:36];
stat                 = ft_freqstatistics(cfg, C_1{:}, C_2{:});
%% Plotting significant clusters
cfg                  = [];
cfg.alpha            = 0.05;
cfg.parameter        = 'stat';
cfg.layout           =lay;
ft_clusterplot(cfg, stat);
I would like to thank you guys for your kind help and assistance.

Kind Regards,

Ronny

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20160215/61d5ffa1/attachment.html>