Hello
I am a "new SPM user" and currently work on a set of 99mTc-HMPAO SPECT
scans acquired with a triple head gamma camera (fan beam collimation and
simultaneous acquisition of 153Gd rodsource for attenuation correction).
The data were acquired for 52 normal healthy volunteers using a split-dose
protocol (reflecting a baseline and an activation condition, 1 scan by
condition). This protocol was assessed twice, in a two-year interval. This
means that we have a final sample set of 208 scans, 52 Baseline Time 1, 52
Activation Time 1, 52 Baseline Time 2 and 52 Activation T2.
However, because the split-dose protocol lead to different counts in the
baseline and in the activation condition, we performed on every scan a
normalization using the individual cerebellar uptake value. This was done
using Toshiba scripts, directly after image reconstruction.
Subsequently, every scan was transformed into ANALYZE format, transformed to
fit in the Talairach space using MedX, realigned, normalized and smoothed
with SPM99.
We now want to test the Time x Condition effect. However, as "new SPM
users", we face a problem regarding normalization parameters. Since our
scans have been already normalized on cerebellar values, should we
nevertheless use a Proportional Scaling procedure (as recommended for SPECT
data) or would it be wiser to adopt another way of normalizing our data ? Or
eventually not normalize at all in SPM (to avoid double normalization) ?
This question is important because we soon will add to the over mentioned
sample, a sample of Alzheimer patients and a sample of young adults .
Regarding this latter project, will we be able to test the Group (AD, Young,
Elderly) x Condition (Baseline, Activation) x Evaluation (T1, T2)
interaction using SPM (assuming that we'll have 40 subject per group) ? If
this was possible, what would be the most appropriate design in SPM99,
considering that we have only 1 scan per subject in each condition ?
Thanks in advance for providing me with answers that will help to
go further into the analyses.
Catherine
|