Hi,
On 9 Mar 2007, at 19:34, Deane Aikins wrote:
> Hello everyone,
>
> I am interested in learning how to model the within-subjects effect
> of time on a two-stimulus
> presentation.
>
> Each participant viewed two stimuli (CS+, CS-) for a total of ten
> trials (5 per stimuli).
>
> In my first-level FEAT, I have created 2 original EVs and added the
> temporal derivates to
> investigate the main effect of stimulus type. Thus, the contrasts:
>
> 1 -1 (CS+ > CS-)
> -1 1 (CS+ < CS-)
> 1 0 (CS+ > rest)
> -1 0 (CS+ < rest)
> 0 1 (rest > CS-)
> 0 -1 (rest < CS-)
>
> We also belive the signal strength in our ROI may be greatly
> affected by repeated stimulus
> exposure. For some participants, the signal may habituate, in
> others it may sensitize.
>
> Do I need to create 5 original EVs to represent the 5 trials of
> stimulus exposure, and then create
> contrasts that model each potential "shape" of the within-subjects'
> time effect?
That's the most flexible way of doing it, yes. Or you could assume
linearity and produce, in addition to the constant height original
EV, an orthogonal one containing the linear trend within it, but yes
it's probably easiest to start as you suggest.
> I belive I would model a linear signal habituation as:
>
> 2 1 0 -1 -2
That's right.
> (and a sensitization model would be the opposite).
>
> For a cubic effect, would I model:
>
> 2 1 -6 1 2
?? I presume you mean quadratic (as this contrast is symmetric about
the centre), and that would therefore want to be something like
4 1 0 1 4 before demeaning, resulting after demeaning in:
2 -1 -2 -1 2
> The interaction of the main effects of stimulus and time are also
> of interest, as we predict CS- to
> remain constant over time and CS+ to change.
>
> Are all interaction effects analyzed if I select the "interaction"
> basic shape of an addition EV?
Not quite - it will ask you which existing EVs you want to interact,
so you just need to specify the appropriate ones.
Cheers, Steve.
>
> Thank you so much for your help,
>
> Deane
------------------------------------------------------------------------
---
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director, Oxford University FMRIB Centre
FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
------------------------------------------------------------------------
---
|