> for my thesis i am comparing two subjects:
> one is ill (Alzheimer) and one is healthy;
> my objective is to distingush in the ill subject the pathologic regions.
> I have three scans for subject.
> My analyse is :
>
> pre-processing:
>
> Realignement : #subject 1
> sinc interpolation
> select scans : six scans
> creat all+mean
>
> Normalization : determine e write normalisation
> #subject 1
> image to determine: Mean
> images to write : six scans
> Method interpolation bilinear
>
> Smoothing FWHM=0.75 mm
>
> Design type:
> singlesubject: condition e covariates
> select images: six scans
> [6] condition 111000
> [6] covariates 1,1,1,1,1,1
>
> Covariate interection : <none>
> covariate centre : around overall mean
> # nuisance variables : 0
> Global normalisation: proportional scaling
>
> Contrast (1,-1)
> p value corrected 0.05
>
> I allege my result
>
> Excuse-me and thank-you for all helps very important
I'm afraid that this analysis does not really tell you all that much. It
just shows you where one subject differs significantly from the other.
This difference need not be related to pathology.
In order to detect pathology, you probably need to define what is normal
using several subjects, and use a mixed (random) effects model. This
would involve creating a mean image for each of your subjects, which you
then spatially normalise and smooth by about 12mm. You could then
compare the control group with the patient.
I noticed that you treated all your scans as if they had
come from the same subject. For realignment and normalization, you
should say that you have two subjects, and enter the three scans for
each of them. This would mean that registration within subject is
rigid-body, whereas inter-subject registration is affine and nonlinear.
Something that I noticed from the figures you sent are that the orientations
of your spatially normalised images may have the anterior-posterior
directions reversed. Before doing anything to your images, you may want to
rotate then to the correct orientation using the Display button.
Best regards,
-John
--
Dr John Ashburner.
Functional Imaging Lab., 12 Queen Square, London WC1N 3BG, UK.
tel: +44 (0)20 78337491 or +44 (0)20 78373611 x4381
fax: +44 (0)20 78131420 http://www.fil.ion.ucl.ac.uk/~john
|