[pymvpa] detecting changes between fMRI sessions: a new mapper?

Yaroslav Halchenko debian at onerussian.com
Wed Sep 17 18:14:49 UTC 2008

So I see 2 questions here actually:

1. technical -- how to combine features of two (or more datasets) into 1
dataset. For now we have some way to ease the life of a user:
MetaDataset, which can join features of multiple datasets. Imho it is
just a temporary solution since
 * it is not a full-featured dataset since is not really a
    Dataset-derived class, so many things could go wrong

 * imho  there should be some mapper (come up with nicer names than 
   MetaMapper, CombinedDatasetsMapper, etc), which would do similar
   thing -- take multiple data(set)s and combine into a single blob.
   But it raises another aspect which we haven't fully discussed yet:
   interaction between Mappers and Datasets. For now forward/reverse
   operate on data, although I believe I made them also digest
   datasets's .samples if dataset is given. .reverse is always spitting
   out .data. But now we have also .train method (optional) so some
   mappers have it and it takes dataset... .reverse for "MetaMapper"
   should spit out corresponding Dataset, thus we just need to agree
   upon something... or just let it go... what am I talking about
   anyways? :-) sorry for the moot

2. neuroscientific -- how to contrast two sessions... That is imho more
   complicated. I had a limited experience with 1 dataset where
   different conditions were collected in different runs... so the
   problem was to remove such 'run-specificity' so clf don't rely on
   simple misalignment or trend to classify samples. apparently detrend
   + zscoring helped -- at the end sensitivity maps were quite sensible.

   In Emanuele's example situation is more interesting, although could
   be done indeed just simple way (in addition to suggested): double
   number of labels -- ie. dog1 (for session 1) dog2(for session 2).
   Do regular analysis, and then look at the sensitivity maps between
   dog1-vs-dog2. If dog category wasn't trained for at home -- I hope
   that generalization and sensitivity would be at-chance. If it was
   trained at home -- should be sensible...

   noone would know which one is the best way to proceed unless tried
   them all I guess ;-)

Yaroslav Halchenko
Research Assistant, Psychology Department, Rutgers-Newark
Student  Ph.D. @ CS Dept. NJIT
Office: (973) 353-5440x263 | FWD: 82823 | Fax: (973) 353-1171
        101 Warren Str, Smith Hall, Rm 4-105, Newark NJ 07102
WWW:     http://www.linkedin.com/in/yarik        

More information about the Pkg-ExpPsy-PyMVPA mailing list