[pymvpa] null classification performance in the presence of strong univariate signal??

[J.A. Etzel]

I agree with Nick that something might have went wrong with the t-test. 
I've never tried one in fsl, either, but usually use R. Here's a little 
bit of R code to do a voxelwise t-test: 
http://mvpa.blogspot.com/2014/09/demo-r-code-to-perform-voxelwise-t-test.html

I assume you looked at the 19 subjects' accuracy maps before trying the 
t-test ...

Jo



On 9/18/2014 8:58 AM, Nick Oosterhof wrote:
>
> On Sep 18, 2014, at 3:36 PM, David Soto <d.soto.b at gmail.com> wrote:
>
>> I reiterate I did this separately for each of the 19 subjects.
>>
>> I then aimed to carry out a group analyses using the individual
>> accuracy maps. to do this I merged the 19 nifti accuracy maps into
>> a 4D file and run a one-sample t-test in FSL using randomise -i
>> searchpred -o searchpredOneSampT -1 -v 5 -T.
>>
>> Weirdly the output gives a group map with all brain voxels over
>> p<0.001 !? which cannot be right...
>
> I am not familiar with FSL's randomise tool (and a quick google
> search only gave me limited information).
>
> One possibility: did you test against the null hypothesis of a mean
> of zero? In your case chance level is .5 (=1/2), not 0. It could
> explain why you are getting all highly significant voxels. If that is
> the case, you should subtract .5 from the classification accuracy
> maps (of individual subjects) before testing against a mean of zero.
> _______________________________________________ Pkg-ExpPsy-PyMVPA
> mailing list Pkg-ExpPsy-PyMVPA at lists.alioth.debian.org
> http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa
>
>
-- 
Joset A. Etzel, Ph.D.
Research Analyst
Cognitive Control & Psychopathology Lab
Washington University in St. Louis
http://mvpa.blogspot.com/