Hi - you can just set up your model like you would for any group
comparison.
I.e.
1 0
1 0
1 0
1 0
1 0
0 1
contrast
1 -1
-1 1
The problem is that you cannot use randomise, as you don't have
enough potential permutations (only 5!), so you will have to assume
normality and just use the t-maps from a single iteration of randomise.
Multiple comparison correction is obviously a bit of a problem if you
can't use permutation testing though, so you will need pretty high t-
values to convince a reviewer.
T
On 30 Jan 2008, at 07:59, [log in to unmask] wrote:
> Hi all,
>
> I have one patient (n=1) and would like to compare his FA with
> a group of controls (n=5).
> I have been through all tbss steps as described in
> http://www.fmrib.ox.ac.uk/fsl/tbss/index.html (using the
> template FMRIB58 for corregistration).
> I am now ready to run the statistics but I am not sure about
> which procedure is the correct one. My problem is that I am
> trying to compare 1 patient with a whole group.
> Does anyone know which test can give me a map of FA differences
> between the patient and controls?
>
> I thank you in advance,
>
> Peace,
> Joyce!
>
> --
> joyce
>
|