[pymvpa] Searchlight statistical inference

[Yaroslav Halchenko]

On Wed, 12 Aug 2015, Roni Maimon wrote:

>    So the full design is I have 4 conditions in 8 runs. 5 blocks of each
>    condition in each run.
>    All runs have all the conditions but I'm interested only in two
>    classifications and the differences between these classifications.
>    The order of trials is different across runs.
>    Some recommend I only permute the labels within runs, is this what you're
>    referring to? 

it depends on the meaning of "permute the labels":

- completely randomly shuffle all the labels, e.g.

  ababab  could become anything of aaabbb, ababab, aabbab, ..

 that is the default behavior of  AttributePermutator --
 strategy='simple'

- or "reassign" the labels i.e. 
  ababab  can become only bababa or ababab in that run (since we have
  only two labels)

  that is the strategy='uattrs'

in both of above cases, permutations should happen strictly within each
chunk/run (that is a must) -- use    limit='chunks' .  But as you might
have noted, complete permutation (strategy='simple') might not pin
point deficient designs (i.e. ababab).

And then the strategy='chunks'  is when you take complete sequence of
trials from one run and assign to another.  If your trials lost original
order when you e.g. extracted betas -- this strategy is not applicable.

> Is there a quick way to do that in pyMVPA?

see above ;)  AttributePermutator is the one doing it 'quickly' ;)

-- 
Yaroslav O. Halchenko, Ph.D.
http://neuro.debian.net http://www.pymvpa.org http://www.fail2ban.org
Research Scientist,            Psychological and Brain Sciences Dept.
Dartmouth College, 419 Moore Hall, Hinman Box 6207, Hanover, NH 03755
Phone: +1 (603) 646-9834                       Fax: +1 (603) 646-1419
WWW:   http://www.linkedin.com/in/yarik