Dear Jack,
> We designed an fMRI study in the following way :
>
> First fMRI session we present in an encoding situation :
> 0 1 0 1 0 1 0 1 0 2 0 2 0 2 0 2
>
> Then a few min later the retrieval session :
> 0 3 0 3 0 3 0 3 0 4 0 4 0 4 0 4
>
> 0 -> rest
> 1 -> lecture 1
> 2 -> encoding
>
> 3 -> retrieval
> 4 -> lecture 2
>
> This study was designed as categorical for comparison of 2 vs 1 (say
> comparison A) and 3 vs 4 (comp B) in a subject by subject manner. But
> we tried also 3 vs 1 (comp C) and 1 vs 4 (comp D = the 2 lecture).
>
> As we are finding preservative results in all those conditions, and
> particularly in cond C and D, we would like to know whether such an
> interpretation is correct for MRI time series when comparing to streams
> of data acquired in 2 different fMRI session. Patient stay in the MRI
> between the 2 sessions => very few movements, however scale factor
> between the series is different. Can it be a problem ?
We commonly pool sessions in our analyses when using fixed effects
models in this way. We do however ensure that each session is modelled
separately with its own constant term and other confounds. One problem
you may face is that comparisons C and D could be confounded with
non-specifc time by condition interactions (i.e. adaptation of
hemodynamic responses). This would not be a problem if you used a
random effects analysis and the order of the two sessions was balanced
over subjects.
With best wishes - Karl
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|