[pymvpa] group-level analysis differences vs. pattern classification

[Nick Oosterhof]

On Sep 23, 2013, at 3:37 PM, Vadim Axel wrote:

> BTW, if at first step I run conjunction analysis with contrasts A > C and B >C; then, for ROIs resultant from conjunction analysis analysis, I  run MVPA using the same data. Can I be accused by  N.Kriegeskorte in double-dipping crime?

Not if your contrasts are independent of the classes that you want to classify. 
We did such a conjunction analysis (with 'C' being baseline) followed by MVPA:

Oosterhof, N. N., tipper, S. P. & Downing, P. E. Viewpoint (in)dependence of action representations: an MVPA study. J Cogn Neurosci 24, 975–989 (2012).