Dear PET SPM experts,
I have the FDG brain PET images of the following groups
Lesion_24,
Lesion_7D,
Lesion_1M
Lesion_2M,
Lesion_6M,
Lesion_12M
Sham_1M,
Sham_2M,
Sham_6M,
Sham_12M
Healthyrats_BL
and I would like to see if and where there is a metabolic change in
cerebral metabolism
Should I simply use a 2-sample T-test?
I would say Yes for independence as they are not always the same rats
I would say Unequal to variance when comparing the lesioned to the
sham but equal when comparing the same type of surgery (correct?)
then the rest is not quite clear to me
grand mean scaling...
ancova
covariates...
masking..
any suggestions are more than welcome,
thank you in advance
Laura Ravasi
On Fri, Oct 3, 2014 at 10:05 PM, Katie Surrence <[log in to unmask]> wrote:
> Dear SPM experts,
>
> If you have a parametric modulator but are very occasionally missing data
> (you have an onset, but the RT and the answer did not get recorded, perhaps
> because there was no button press that trial), what is the best way to
> handle that? I've tried to look for discussion of this issue in the
> archives, but to my surprise I can't find it.
>
> I tried setting the missing values to NaN, but I got this error:
>
> Failed 'fMRI model specification'
> Error using svd
> Input to SVD must not contain NaN or Inf.
>
> Do I need to choose a method of interpolation myself and construct a
> parametric modulator myself that I enter as a separate regressor?
>
> Thanks very much for your help, as always.
>
> Best,
> Katie
>
>
>
>
--
Cordialement,
Regards,
Laura Ravasi
|