Hi Boris,
when you have R1+R2=R3+R4 in your matrix, ß can't be uniquely estimated.
if B={ß1,ß2,ß3,ß4} is ok, that is Y=ß1R1+ß2R2+ß3R3+ß4R4+E,
then with any constant, B={ß1+c,ß2+c,ß3-c,ß4-c} is also ok, because
Y=(ß1+c)R1+(ß2+c)R2+(ß3-c)R3+(ß4-c)R4+E is also true.
so maybe just try to define your matrix in another way.
Hope this helps,
Qing :)
boris suchan a écrit :
> Hi all,
>
> I am trying to specify a DCM model derived based on a 2 x 2 factorial
> design.
>
> First, I recombined the onsets of my 4 cells (A1 + A2; B1 + B2; A1 +
> B1; A2 + B2) to get the 4 conditions representing all 2 main effects.
>
> When specifying the spm model, I get the message under my design
> matrix: gray -à ß not uniquely defined…
>
> Indeed there are some overlap of onsets as these are the combinations,
> but this is the same as in the example data from the SPM homepage.
>
> Additionally it is also not possible to specify the DCM model…
>
> Any help? Many thanks
>
> boris
>
--
************************************************************************
Qing CAI
Institut des sciences cognitives
67, boulevard Pinel
69675 Bron Cedex, France
Telephone: +33 (0)4 37 91 12 53
E-mail : [log in to unmask]
************************************************************************
|