Hi,
To start with the documentation says:
"One difference is that FIRST does not model separate group variances
and so
ignores the "Group" column in the Glm gui."
Therefore it does not matter what you put into the Group column.
So there is no point in changing these values.
When doing vertex analysis with FIRST, it will add in a group-mean
regressor if one is not explicitly present. Therefore you do not need
to do option 3. Either option 1 or option 2 are fine, and they should
give you exactly the same areas of statistical significance. The
difference will be that the vectors will point in opposite directions in
the two cases. It says in the documentation that:
"These vectors point from the mean surface of whatever subjects
corresponded to a value of 0 or less in the design matrix (demeaned EV)
to the mean surface of others subjects." The demeaning is done
internally, not by you.
Does this explain the way in which option 1 and option 2 differ?
All the best,
Mark
On 16 Nov 2009, at 03:06, SUBSCRIBE FSL hankee wrote:
> Hi, FSL experts,
>
> I am a newcomer to FSL. When using FIRST to make group analysis,
> for example, 5 AD patients vs 6 controls, I need to make a design
> matrix using
> Glm gui. I launched the Glm gui in the terminal. I chose "Higher
> level/nontime
> series design" and specified the input number to be 11. I tried to
> make the
> design matrix in the following different ways, but I am not quite
> sure which
> was correct for group analysis between AD and con;
>
> Option 1,
>
> Group EV1
> 1 0
> 1 0
> 1 0
> 1 0
> 1 0
> 0 1
> 0 1
> 0 1
> 0 1
> 0 1
> 0 1
>
> Option 2,
> Group EV1
> 0 1
> 0 1
> 0 1
> 0 1
> 0 1
> 1 0
> 1 0
> 1 0
> 1 0
> 1 0
> 1 0
>
> Option 3,
>
> Group EV1 EV2
> 1 1 0
> 1 1 0
> 1 1 0
> 1 1 0
> 1 1 0
> 2 0 1
> 2 0 1
> 2 0 1
> 2 0 1
> 2 0 1
> 2 0 1
>
> Which of the above options is the correct one for the difference
> between AD
> and con?
>
> What is the difference between option 1 and option 2?
> Do they mean that AD-con and con-AD respectively?
>
> I have tried option 1 and option 2 respectively. The results
> obtained from
> them differ a lot! I do not quite understand that.
>
> Thanks in advance!
>
> Hankee
>
|