[FieldTrip] indepsamplesF with sourcestatistics

[Frederic Roux]

Dear all,

I noticed that the result of sourcestatistics can be quite different when using cfg.statistic = 'indepsamplesF'
or also cfg.statistic ='indepsamplesregrT', depending on whether you specify a unit of observations (UO's)
variable in the design matrix or not.

For example when running the code without the UOs-variable as following

cfg.design(1,:) = [ones(1,length(indx1)) 2*ones(1,length(indx1)) 3*ones(1,length(indx1)) 4*ones(1,length(indx1))];
cfg.ivar = 1;

I get a different result than if I run:

cfg.design(1,:) = [ones(1,length(indx1)) 2*ones(1,length(indx2)) 3*ones(1,length(indx3)) 4*ones(1,length(indx4))];
cfg.design(2,:) = [indx1' indx2' indx3' indx4'];
cfg.ivar = 1;

cfg.uvar = 2;

Note that in the second case the indexes in the UOs variable are not sorted.

I guess that you are not supposed to run the code as in the second case,
but I would still like to know why I get different results. Shouldn't it be the same anyways?

It would be great to get an explanation for the discrepancy between both codes.

Best,
Fred

-- 
Frédéric Roux, PhD student
Department of Neurophysiology
Max Planck Institute for Brain Research
D-60529 Frankfurt am Main
Frederic.Roux at brain.mpg.de
+49(0)69630183225


 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.science.ru.nl/pipermail/fieldtrip/attachments/20120318/c391e704/attachment-0001.html>