[pymvpa] pyMVPA with betas

[Nick Oosterhof]

> On 3 Mar 2017, at 13:46, jasi3k <jasi3k at gmail.com> wrote:
> 
> Dear all,
> 
> I am starting my adventure with pymvpa and I would like to use beta maps for my analysis. I have collected data in 10 sessions with 6 different events repeated twice in each session, which gave me 120 beta maps.
> 
> Would this be an appropriate setup for attributes file?
> 
> Ev1 0
> Ev2 0
> Ev3 0
> Ev4 0
> Ev5 0
> Ev6 0
> Ev1 0
> Ev2 0
> Ev3 0
> Ev4 0
> Ev5 0
> Ev6 0
> Ev1 1
> Ev2 1
> Ev3 1
> Ev4 1
> Ev5 1
> Ev6 1
> Ev1 1
> Ev2 1
> Ev3 1
> Ev4 1
> Ev5 1
> Ev6 1
> 
> etc.

That seems fine.

> 
> No matter how i split my data for training and prediction my error is always equal 0.
> 
> This is my code https://paste.ofcode.org/7jw4Ecu5LZHJaggUHKv4Za

Line 48: fds = fds[fds.sa.targets == 'Ev1']

This suggests you are only keeping one condition in the dataset. It would explain a zero prediction error.