Dear Marion,
> I have some PET data I'm trying to interpret, but
> I'm not sure about what is plotted in the
> 'contrast of parameter estimates' when they
> are plotted for each condition for an effect of
> interest in SPM99.
This bar-plot shows the mean-corrected parameter estimates of all
effects of interest. The red lines are the standard errors of the
parameter estimates.
> This plot seems to be representing
> the 'size' of the effect on interest at a given
> maxima - my question is how a negative value in this
> plot should be interpreted. Is it a 'deactivation'?
Because of the mean correction, the bar-plot shows the deviations of the
parameter of interest estimates from their mean. Therefore a negative
value does not necessarily mean that the parameter estimate is negative,
it is just lower than the mean of the parameter of interest estimates.
Note that the (non mean-corrected) parameter estimates of a given voxel
are stored in the workspace variable 'beta', when you plot them. By
typing beta in your matlab window, you can display them.
> Or, asked another way, what does the 0 effect size
> mean in these plots?
It means that this parameter estimate is equal to the mean of all
parameter of interest estimates. As a special case of only one parameter
of interest, it would mean that this parameter is zero.
I guess that the typical use of this plot is to easily assess the
relative sizes of the parameter estimates for a given voxel. You could
also use this plot to extract the vector of parameter estimates (and
other variables like the standard errors of the parameter estimates, the
fitted and the adjusted data stored in 'SE', 'Y' and 'y') from SPM99.
Stefan
--
Stefan Kiebel
Functional Imaging Laboratory
Wellcome Dept. of Cognitive Neurology
12 Queen Square
WC1N 3BG London, UK
Tel.: +44-(0)20-7833-7478
FAX : -7813-1420
email: [log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|