Hi
EVs 1-3 need to be mean 0, i.e. the values in there should be -10/19
and +9/19 respectively. However, it seem that in your design each
subject was scanned repeatedly? If so, the design would have to be
different altogether, the 2x2 example in the feat manual describes a
scenario where in each factor&levels group (cell of the ANOVA table)
there exist 2 measurements from separate subjects and each subject
only ever appears once in the ANOVA table. How many measurements exist
in your ANOVA table per cell?
cheers
Christian
On 5 Jul 2008, at 23:38, Anne Schneider wrote:
> Hello,
>
> I am a student at the University of York and would be very grateful
> if you
> could perhaps give me some advice with regards to the following fsl-
> related issue.
>
> I am planning to run higher-level analysis for 19 subjects using an
> ANOVA
> 2x2 repeated measures - I have 2 factors, each at two levels: Valence
> (positive and negative) & Arousal (high and low). (I.e. each subject
> listened to the same stimuli, which were grouped according to these
> categories)I've had a look at the feat manual but can't quite
> understand
> the illustrated example. Would my design matrix look this:
>
> Group EV1 EV2 EV3 EV4
> Input1 1 -1 -1 1 1
> Input2 1 -1 -1 1 1
> Input3 1 -1 1 -1 1
> Input4 1 -1 1 -1 1
> Input5 1 -1 -1 1 1
> Input6 1 -1 -1 1 1
> Input7 1 -1 1 -1 1
> Input8 1 -1 1 -1 1
> Input9 1 -1 -1 1 1
> Input10 1 -1 -1 1 1
> Input11 1 1 1 1 1
> Input12 1 1 1 1 1
> Input13 1 1 -1 -1 1
> Input14 1 1 -1 -1 1
> Input15 1 1 1 1 1
> Input16 1 1 1 1 1
> Input17 1 1 -1 -1 1
> Input18 1 1 -1 -1 1
> Input19 1 1 1 1 1
>
> The way I understand it - EV1 models the effect of A, EV2 of B, EV3
> the
> interaction between both & EV4 the grand mean (?)
>
> With three F-tests being performed:
>
> A 1 0 0
> B 0 1 0
> AB 0 0 1
>
> I am new to fsl (& neuroimaging in general) so please forgive the
> naivety
> of the following question as well - would I have to use the 4 lower-
> level
> copes for (i) positive-valence vs baseline (ii) negative-valence vs
> baseline (iv) high-arousal vs baseline (v) low-arousal vs baseline?
> OR 2
> lower-level copes for (i) negative > positive valence (ii) high
> arousal >
> low arousal??
>
> I would be extremely grateful for any sort of advice or help you could
> provide on this issue.
>
> Best wishes,
> Anne
|