Hi - if you are taking out the subject means (which makes sense) then your
concern is probably correct, namely that you need to re-state the dummy
variables such that each is zero-mean. Obviously this will change the
meaning of the group-level EVs and you will need different contrasts as
well, but I'm sure you'll be able to ask the questions you want. A minor
query on your contrasts - I'm not sure why you included subject mean
values in the "dual task" contrast.
Cheers.
On Tue, 27 Jan 2004, Paige Scalf wrote:
> Hi FSL Guru's
>
> I've painted myself into a really complicated repeated measures,
> nested design, and I'd appreciate some feedback about whether the
> EV's and contrasts I've set up for my second level analysis are
> correct.
>
> This is a dual-task task visual attention study, so we have each
> of the dual task conditions in a single task form. Because we
> have some repeated measures within the dual task condition,
> however, we end up with an unbalanced (nested design).
> The conditions are
>
> A (single task)
> 1 center task
> 2 peripheral vision task, proximal to fixation
> 3 peripheral vision task, distal to fixation
>
> B (dual task)
> 1 center task + proximal peripheral task
> 2 center task + distal peripheral task
>
>
> I've set up my matrix using dummy variables, such that
>
> X1 = 1 single, 0 else
> X2 = 1 center single, 0 else
> X3 = 1 proximal peripheral single, 0 else
> X4 = 1 proximal dual single, 0 else
> X5 = subject 1
> X6 = subject 2
> X7 = subject 3
>
>
> The EV Matrix ends up being
>
> A1 1 1 0 0 1 0 0
> A2 1 0 1 0 1 0 0
> A3 1 0 0 0 1 0 0
> A4 0 0 0 1 1 0 0
> A5 0 0 0 0 1 0 0
>
> And my contrasts are structured
>
> single task 1 0 0 0 0 0 0
> dual task 0 0 0 .5 .165 .165 .165
> center single 0 1 0 0 0 0 0
> etc.
>
>
> My biggest concern is that my EV's are set up using a very simple
> "on/off" matrix, but if I set it up in the more traditional manner
> (i.e. 1 for single task, -1 for dual task) I find I can't make the
> maths work out for the contrasts I need. (I'll be happy to accept
> feedback about any errors, however).
>
> Thanks in advance for any help you can give.
>
> Paige
>
Stephen M. Smith DPhil
Associate Director, FMRIB and Analysis Research Coordinator
Oxford University Centre for Functional MRI of the Brain
John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
|