External Email - Use Caution
Hello FreeSurfer Developers,
I am building a brain age model from the stats extracted with FreeSurfer. I
tested my model normalizing volumes by eTIV, which I think is the
recommended, and surfaces by WM total surface. For surfaces, I haven't
found any recommendations. I am getting better results by using the
features without normalization.
From what I have read, volume normalization clearly should be performed,
but I am in doubt about the surface normalization. Although, the surface is
different depending on head size and along with lifespan, I am not sure if
all cortex surfaces evolve the same way, and therefore that normalization
might be adding noise to my results. Then, is it correct to normalize my
surface features by the total WM surface? Does anybody know any reference
that talks about the topic or have done similar work?
Thank you for your support
-----------------------------------
|| Rafael Navarro González
|| Ph.D. Student
|| ETSI Telecomunicación
|| Universidad de Valladolid
|| Campus Miguel Delibes
|| 47011 Valladolid, Spain
|| e-mail: rafael.navarro(a)uva.es
|| Phone: +34 672 394 294
||_________________________________
-----------------------------------