Hi Nasser,
On 27 Mar 2006, at 18:32, nasser kashou wrote:
> Hello again,
> I did some more reading and this is what I came up with. But I'm
> not sure if this is entirely correct. Can someone please give
> feedback.
> EV1 EV2
> skip 0 0
> off 40 40
> on 20 20
> phase 20 40
> OC1 1 0
> OC2 0 1
> OC3 1 -1
This looks right, yes, if EV1 is showing B, EV2 is showing A and rest
(where both EVs are "down") is C. You don't need a third EV.
> OC3 gives me the expected shape of my paradigm, where A(off), B
> (on), C(3rd condition bw off and on) in which all are presented for
> 20s each.
OC1 = B-C
OC2 = A-C
OC3 = B-A
> I also, read that to ignore a time point, then can add EV and make
> it 1's for ! that time and 0's everywhere else. So does that mean
> if I want to look at the activation going from A to B ignoring C, I
> would add another EV with 1's at the times C was presented? If so,
> then in order to get the wanted contrast, would I just do this?
Nope - the comments you're thinking about are a different thing -
they are how to model parts of your data (e.g. dropouts) that
otherwise wouldn't get modelled at all and which would otherwise show
up as noise. Here you are already modelling eveything that you need
to so all is fine already.
When you add the temporal derivatives the design is still nowhere
near rank deficient - it doesn't need to be _completely_ orthogonal -
so I think everything's fine.
Cheers, Steve.
>
> EV1 EV2 EV3
> skip 0 0 20
> off 40 40 40
> on 20 20 20
> phase 20 40 20
> OC1 1 -1 0 activation going from A
> to B?
> OC2 1 0 -1 activation going from C
> to B?
> OC3 0 1 -1 activation going from A
> to C?
>
> EV1: when present B
> EV2: when present A
> EV3: when present C
>
> The problem with this model is it becomes non-orthoganal when I add
> the temporal derivatives. Please give me some feedback on if my
> understanding is correct or if I'm way off. Thanks
>
> nasser kashou <[log in to unmask]> wrote:
> Hello,
>
> Recently, I changed my design paradigm from a simple box ABABA to
> ABCABCABC. With this I was hoping to analyze
> A (OFF1) vs. B (ON)
> C (OFF2) vs. B (ON)
> and even
> A (OFF1) vs. C (OFF2)
>
> so I was initially setting up the EVs as
> EV1 EV2
> skip 0 20
> off 20 20
> on &! nbsp; 20 20
> phase 0 0
>
> OC1 1 0
> OC2 0 1
>
> with 1 f-test.
> But after doing some reading, I began to think that this may not be
> the correct way to model this. Any suggestions is appreciated.
>
> Thanks
> nasser
>
> New Yahoo! Messenger with Voice. Call regular phones from your PC
> for low, low rates.
------------------------------------------------------------------------
---
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
------------------------------------------------------------------------
---
|