Dear SPM'ers,
I am currently performing an analysis using Tc-99m HMPAO SPECT image data.
We have been looking at the effects of MMSE score on rCBF in two groups
of dementia types namely AD and DLB. We have carried out this covariate
analysis
separately on the two conditions (AD & DLB) and have noticed there is an
significant effect of MMSE on rCBF in the AD group but not in the DLB group.
My question is, is it legitimate to carry out two separate or if you like
independent covariate analyses of this nature and simply report on their
significance? Or should I carry out a factor by covariate interaction as
there seems to be a condition specific effect of MMSE with rCBF? Or can
anybody suggest
a suitable statistical model to use to investigate this finding.
Please help and any comments would be greatly appreciated.
Cheers
Sean Colloby
Research Associate
Imaging Research Group
Institute for Ageing and Health
Wolfson Research Centre
Newcastle General Hospital
Newcastle-upon-Tyne
NE4 6BE
Tel: +44 (0) 191 256 3372
Fax: +44 (0) 191 219 5051
email: [log in to unmask]
|