Hi,
We did this study using code of Jean-Baptiste's which has now found it's way
into the MM toolbox so you seem to have the right tools. It is not clear to
me, however, whether this method will be helpful for VBM, given that you
only have one observation/subject. Maybe Jean-Baptiste can comment. In any
case, once you have the eigenimage, calculating the expression of it simply
requires you to multiply it with the (vector corresponding to) the scan in
question (taking care of course that they have the same format, i.e. row
coordinates in the one vector refer to the same spatial location as column
coordinates in the other).
Regards,
Andreas Meyer-Lindenberg
-----Original Message-----
From: Yasuhiro Kawasaki [mailto:[log in to unmask]]
Sent: Wednesday, January 29, 2003 9:04 PM
To: [log in to unmask]
Subject: multivariate methods of VBM
Dear SPMs,
Having read the Meyer-Lindenberg et al Am J Psychiatry (158, 1809-1817,
2001) canonical variates analysis paper, I am intending to adopt these
multivariate methods to VBM study, especially to calculate expression of
eigenimage of second subject group that is independent from the original
subject group to extracted eigenimages. I am doing multivariate analysis
using MM Toolbox, and obtained some preliminary result. Can we adopt this
toolbox for f-MRI to VBM? If possible, how can we obtain a formula of an
eigenimage expression? Thank you in advance
============================================
Yasuhiro Kawasaki, M.D.
Department of Neuropsychiatry
Toyama Medical and Pharmaceutical University
2630 Sugitani, Toyama 930-0194, Japan
TEL: +81-76-434-7322
FAX: +81-76-434-5030
e-mail: [log in to unmask]
============================================
|