Dear Gonzalo,
Your question is a bit too vague to be answered in detail - a full description with each step would probably fill five to ten pages!
I really don't know how far you got, and what kind of pathology you're looking at. Let's assume you're using SPM5. Then the _overall_ steps for what you want to do would be:
1. Make an FDG template (search the archives for this)
2. Normalize your individual FDG images to this template ("NORMALIZE" button) -> wFDG images
3. Smooth the wFDG images - the amount of smoothing will depend on a lot of things, amongst which the most important will be the resolution of your system and the spatial extent of the effects you're looking for ("SMOOTH" button) -> swFDG images.
4. Define your design ("PET" button). There is excellent context-sensitive help in the bottom of the Graphics window - whatever you click on will bring this up, and this (in conjunction with the list archives) should solve most of your problems. There are too many details to give a comprehensive answer on "how to do it", but maybe you can get started and get back to the list with specific questions?
An important issue with FDG will be the normalisation for globally differing values. It's probably best to measure this as it will depend on acquisition timing & length, whether dose injected was fixed or adjusted per kg body weight, whether subjects were fasted etc.etc. You're likely to get a coefficient of variation for the global values around 30%, and then it's best to ANCOVA that out per group.
5. Examine the results ("RESULTS" button).
Some details on how to compare a single PET scan against a group can be found in previous papers by us and others - try eg.
Hammers A, Koepp MJ, Richardson MP, Hurlemann R, Brooks DJ, Duncan JS. Gray and white matter flumazenil binding in neocortical epilepsy with normal MRI. A PET study in 44 patients. Brain 2003, 126: 1300-1318.
Hammers A, Asselin MC, Hinz R, Kitchen I, Brooks DJ, Duncan JS, Koepp MJ. Upregulation of opioid receptor binding following spontaneous epileptic seizures. Brain 2007, 130(4): 1009-1016.
- those aren't SPM5, and I now realise that the detail given is probably insufficient for a beginner. Still, worth a go!
All the best and hope this helps to get you started,
Alexander
-----Original Message-----
From: SPM (Statistical Parametric Mapping) [mailto:[log in to unmask]] On Behalf Of Gonzalo Rojas
Sent: 24 March 2009 04:16
To: [log in to unmask]
Subject: Re: [SPM] compare patient to mean of normal subject group (FDG-PET images)
Dear Brahim:
Thanks you very much for your answer, but I don't could get that
paper...
I tried to do a Two-sample t-test to compare the patient PET image
to the controls database FDG-PET images, but I couldn't get good result,
because I am new user of SPM... Could you please send me te steps
necessary to do such comparison ?...
Sincerely,
Gonzalo Rojas Costa
Department of Radiology
Santa Maria Clinic
Avda. Santa María 0500, Providencia, Santiago, Chile.
Tel: 56-2-4613074
Cel: 56-9-97771785
www.csm.cl
Brahim HAMADICHAREF escribió:
> A concept can be used from this paper ...
>
> Voxel-Based Morphometry in Individual Patients: A Pilot Study in Early
> Huntington Disease
> M. Mühlau, A.M. Wohlschläger, C. Gaserd, M. Valet, A. Weindl, S.
> Nunnemann, A. Peinemann, T. Etgena and R. Ilg
> American Journal of Neuroradiology 30:539-543, March 2009
> http://www.ajnr.org/cgi/content/abstract/30/3/539
>
> Regards
> Dr Brahim HAMADICHAREF
|