Hi Phlipp,
> Option (i) refers to strong differences in the spatial
> characteristics: do
> I understand it right if that basically refers to analyzing the
> subjects'
> modes (e. g. across conditions) to see if e. g. under condition A
> this
> component is more strongly contributing than under condition B, in an
> offline fashion?
Yes, that's right
>
>
> As within one tPICA component the spatial distribution should be quite
> similar, is there any use of voxelwise testing of these maps?
>
That's possible as a post-hoc processing step - though strictly
speaking does not fall into the model. You're basicaly assuming that
within the same areas there is a part that discriminates while other
parts within the component don't. if that's the case then this could
show up as two different components which (eve though they might have
very similar time course) have different subject-dependent vectors....
hth
Christian
|