Dear experts,
I work with PET images, were the old-normalize routine is the best thing we have. Generally I want to use it, and confidently if possible, as for many studies we do not necessarily have MRI scans.
Suppose we have a number of PET scans for which indeed 3D-T1-MRI are available. How can I estimate "how good" the PET-based normalization is?
One hypothesis could be to measure the impact on a “known” test. I.e.: I take 40 healthy controls, 40 AD patients, and “take a look” at t-values from the comparison. If the MRI-based normalization is “much better” I should see “much larger” t-values. Or do I? What could I infer from such a test? If t-values remain identical I could say that improved spatial normalization does not impact PET data analysis. But what if they increase and, say, my highest peak goes from t=15 to t=16.5? (note how all the words in quotes aren’t well defined to begin with).
This, however, isn’t really looking at the spatial aspect of it, only whether it affects the following analysis. I guess I could also take a look at the normalization itself. As an example, I could take the segmented gray matter class from the MRI normalization and apply the MRI-based transformation or the PET-based one and see “how far” they are. How could I do that? I could compute the sum of squares. But suppose I find something like 2*10^2. What would that mean? What would be a cut-off to say “this is good”?
Other ideas in general to compare two different strategies for normalization?
Thank you very much,
Luca
Rispetta l’ambiente: non stampare questa mail se non è necessario.
Respect the environment: if it's not necessary, don't print this mail.
IL TUO 5XMILLE AL SAN RAFFAELE DI MILANO
Devolvi il tuo 5 per mille all’Ospedale San Raffaele: perché al centro della Ricerca ci sei TU. CODICE FISCALE: 07636600962, nel riquadro RICERCA SANITARIA. Non c’è cura, senza ricerca. Non c’è ricerca, senza il tuo 5xmille. Scopri come su http://www.5xmille.org
[http://www.5xmille.org/img/731x129.jpg]
|