dear Karl,
thanks for your advice on our problem.
To try to avoid mood induction being confounded by order effects, the
experimental design counterbalances the order of the three mood induction
tasks across all control subjects and Bipolar subjects.
while for each individual subject we would not be able to distinguish
effects of mood induction from those of session order, we wonder whether it
would be possible to examine the main effects of mood induction over all
subjects using a group analysis.
is it possible and valid using spm99 to identify the main effects of mood
induction for each group (control and Bipolar).
Again many thanks
alan mcbride & David Barbenel
> The aim of the experiment is to examine the effects of experimental
> mood induction on verbal fluency activations, in bipolar patients and
> normal controls.
>
> Each subject undergoes 3 mood induction tasks each lasting approx 6.5
> mins. Each mood induction task is followed by verbal fluency tasks
> during which an fmri time series is obtained. Each verbal fluency task
> consists of 30 sec epochs of each of three conditions Repition (A),
> Generation (B) and Rest(C), ordered as follows:
>
> C A B C C B A A B C C B A
>
> The first rest epoch is discarded prior to analysis.
>
> In analysing the data so far, to examine verbal fluency effects, and
> modulation of these by mood, we have used two cut off filters as
> follows:
>
> haemodynamic response function ( HRF)
>
> high frequency cut-off: 180 180 180
>
> the high frequency we presume is optimised depending on previous input
data
>
> we would be grateful for your help in the determination of the
> appropriate filter and contrasts to allow us to best examine the main
> effects of mood.
I am afraid you cannot assess the main effects of mood per se because
they are completely confounded with session. You can assess the mood x
condition interaction which is the aim of your study and would involve
modelling all the sessions in a session-separable design matrix
(enforced by SPM99) and using appropriate contrasts. These contrasts
can be assessed directly in terms of SPM{T} at the first level for each
subject (fixed-effect analsis) or entered into a one-sample T test at
the seond-level (random effects analyis). I would use a two sample T
test at the second level to assess group differences in the mood x
condition interaction contrasts.
e.g. for a deisgn matrix with the following efffects
subject 1 subject 2 ...
mood- mood+ mood- mood+ ...
A B C A B C A B C A B C ...
0 1 -1 0 1 -1 0 0 0 0 0 0 Main effect of fluency -
subject 1
0 -1 1 0 1 -1 0 0 0 0 0 0 Mood x fluency (vs. rest) interaction -
subject 1
0 0 0 0 0 0 0 -1 1 0 ........
I hope this helps - Karl
|