Gang Wrote:
-------------------------------------------------------
> Matthew,
>
> Add two more tests:
>
> -gltLabel 4 grp_task_1vs2 -gltCode 4 "group : -1*
> healthy 1* diseased strategy : 1*task_1 -1*task_2
> " \
> -gltLabel 5 grp_task_2vs3 -gltCode 5 "group : -1*
> healthy 1* diseased strategy : 1*task_2 -1*task_3
> " \
>
> Then the conjunction of the above two tests in the
> positive (or negative) domain would be what you're
> looking for.

Hello,

I had a similar question regarding 3dMVM and linear trends.

We have 3 levels for our condition as well.

Levels:

one two there.

Is there a difference modeling the trend as such with the following gltCode:

Case1:

stim: 1*one 0*two -1*three

vs

Case2:

the conjunction map of the following two contrasts:

stim: 1*one -1*two
stim: 1*two -1*three


Also, how does one interpret the result of Case1 and Case2.

We interested in the scenario where only one > two > three.

Thanks.

Michael