Dear spm’ers,
I would like to know how the a priori probabilities for GM, WM and CSF
(apriori/gray.mnc, apriori/white.mnc & apriori/csf.mnc) were created, ie
where did the necessary information on tissue classes come from. Or was
there no need for such information?
The background of my question concerns contrast at different field strengths.
At the increasingly used field of 3T – and in the future possibly higher –
(slightly but systemetically) different contrast than at 1.5T is likely to
be observed, eg owing to different sequence parameters and relaxation
behavior. Quite similarly to a study of, say, elderly brains this would mean
that the 3T data do not match the spm default data too well. For something
like carrying out VBM we know it might be advisable to produce one’s own
priors, in the present case from the 3T data. Am I not right that this
process involves the default prior probability images for GM, WM and CSF
produced from T1 weighted images acquired at 1.5T?
In other words, don’t we carry the information contained in 1.5T images
through to deviating conditions, eventually to data acquired at 7T, with
increasing risk of tissue misclassification?
Any comment would be highly appreciated.
|