Hello Chris-
If you do decide to normalize your patient's scans, you may want to examine
the brain masking function described by Brett, Leff and Ashburner.
Essentially, you create a lesion mask so that the normalization routines
are not disrupted by the unusual features of the lesion:
http://www.apnet.com/www/journal/hbm2000/6786.html
I have written a step-by-step tutorial for this procedure:
http://www.psychology.nottingham.ac.uk/staff/cr1/mritut.html#Normalize
My web page also discusses the pitfalls of patient normalization (e.g.
problems with the nonlinear transforms).
-chris
_________________________________________________________
At 14:55 18/11/2000 +0100, [log in to unmask] wrote:
>hi SPM
>
>we would like to scan a patient with a large posterior brain lesion using
>fMRI
>time-series. Looking at old correspondence from the list about this issue
>I have
>found a number of relevant mails, but I have some more specific
>questions relating to the normalization process and the template to
>be used.
>
>James Patterson's mail of 19/4/2000
>
>http://www.mailbase.ac.uk/lists/spm/2000-04/0131.html
>
>suggests that you could use the mean image of all the patients
>scanned as a template. My question is this: if you have a single subject,
>could
>you use the mean image from all the realigned scans of that subject?
>What would you gain by this? Clearly this
>would not help you put the brain in a Talairach space. On a more
>general note, in a situation such as this what would be lost by NOT
>normalizing at all - is it likely that your activations would be
>spatially displaced? Is there another way to avoid this?
>
>thanks -
>
>Chris Summerfield
>departament psiquiatria i psicobiologia clinica
>universitat de barcelona
__________________________________________
Dr. Chris Rorden
Department of Psychology
University of Nottingham
Nottingham NG7 2RD, UK
Tel: +44 [0]115-951-5294
Fax: +44 [0]115-951-5324
http://www.psychology.nottingham.ac.uk/staff/cr1/
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|