Hi Cherif! I think you're going in the right direction, yes. Your final
example below sounds like the right approach - this should work fine.
Cheers, Steve.
On Sat, 3 Dec 2005, Cherif Sahyoun wrote:
> I'm not too sure how to analyze the following situation in Feat. I am trying
> to combine runs in a subject where each run is not an exact repeat in terms
> of EVs:
>
> A subject does 6 runs of a study that has 3 conditions (A, B, C),
> but only 2 per run i.e.
> run 1: A vs B
> run 2: A vs C
> run 3: B vs C
> runs 4, 5, 6 repeat 1, 2, and 3
>
> In addition, the EVs for a given condition are not the same across
> runs (they are self-paced).
>
> What would be the best way to deal with this? possibilities i thought
> of was to concatenate data from all runs and fix EVs appropriately, or
> analyzing each run separately, but then how do i combine the data? i'm
> thinking by weighing particular contrasts in FLAME, but not sure how
> exactly. Note that because of the pair-wise presentation of conditions in
> the runs, the COPE numbers will never match, which
>
> really is the source of the problem and led me to use lower level copes as
> inputs as describes below.
>
> Here are two ways I tried, but i am not sure what is more robust/elegant:
>
> so for the first level, I basically analyzed each run separately so that for
> each run I have 2 EVs and the contrasts are
> 1 0
> 0 1
> 1 -1
> -1 1
>
> which gives copes 1 through 4 for each run.
>
> now for higher level analysis I entered all 24 copes in that order and all
> were set to group1, then 9EVs:
> 1 0 0 0 0 0 0 0 0
> 0 1 0 0 0 0 0 0 0
> 0 0 0 1 0 0 0 0 0
> 0 0 0 0 1 0 0 0 0
> 1 0 0 0 0 0 0 0 0
> 0 0 1 0 0 0 0 0 0
> 0 0 0 0 0 1 0 0 0
> 0 0 0 0 0 0 1 0 0
> 0 0 1 0 0 0 0 0 0
> 0 1 0 0 0 0 0 0 0
> 0 0 0 0 0 0 0 1 0
> 0 0 0 0 0 0 0 0 1
> 1 0 0 0 0 0 0 0 0
> 0 1 0 0 0 0 0 0 0
> 0 0 0 1 0 0 0 0 0
> 0 0 0 0 1 0 0 0 0
> 1 0 0 0 0 0 0 0 0
> 0 0 1 0 0 0 0 0 0
> 0 0 0 0 0 1 0 0 0
> 0 0 0 0 0 0 1 0 0
> 0 0 1 0 0 0 0 0 0
> 0 1 0 0 0 0 0 0 0
> 0 0 0 0 0 0 0 1 0
> 0 0 0 0 0 0 0 0 1
>
> and the contrasts simply:
> 1 0 0 0 0 0 0 0 0
> 0 1 0 0 0 0 0 0 0
> 0 0 1 0 0 0 0 0 0
> 0 0 0 1 0 0 0 0 0
> 0 0 0 0 1 0 0 0 0
> 0 0 0 0 0 1 0 0 0
> 0 0 0 0 0 0 1 0 0
> 0 0 0 0 0 0 0 1 0
> 0 0 0 0 0 0 0 0 1
>
> The other option was to just have in the second order analysis, EVs for just
> the main conditions vs rest (i.e. cope1 and
>
> cope2 from each first level analysis). Again, all in the same group:
> 1 0 0
> 0 1 0
> 1 0 0
> 0 0 1
> 0 0 1
> 0 1 0
> 1 0 0
> 0 1 0
> 1 0 0
> 0 0 1
> 0 0 1
> 0 1 0
>
> then the contrasts would be:
> 1 0 0
> 0 1 0
> 0 0 1
> 1 -1 0
> -1 1 0
> 1 0 -1
> -1 0 1
> 0 1 -1
> 0 -1 1
>
> I feel this second way is better just intuitively, but it seems I'm going at
> it the wrong way altogether.
>
> Any tips appreciated!
> thx in advance!
> cherif
>
--
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
|