Hi,
Can I just ask here why you are using voxelwise SIENA rather than VBM?
All the best,
Mark
On 27 Jan 2014, at 13:36, Anderson M. Winkler <[log in to unmask]> wrote:
> Hi Elijah,
>
> Don't use Feat for this. Use the command Glm (or Glm_gui in the Mac). You already have the 4D file as I understood from the original post. So, all what you need is to use Glm to create the design matrix. In the window that appears, select higher-level design; the rest is self-explanatory I think. Once you save, use the design.mat, design.con e and design.fsf in randomise.
>
> All the best,
>
> Anderson
>
> Am 26.01.14 18:45, schrieb Elijah Mak:
>> Hi Anderson,
>>
>> Many thanks for your response to get me going. I have a few questions regarding the construction of the GLM. Could you take a look at my step by step procedure to see if I might have gotten any step wrong?
>>
>> First, I type Feat_gui.
>>
>> 1) I select Higher Level Analysis.
>>
>> 2) Since I have 65 subjects, I enter 65 for "Number of inputs"
>> At this point, am I supposed to select FEAT directories? What does this mean? I am in a folder containing my SIENA data in subject folders:
>>
>> 1A_to_3B_siena
>> 2A_to_3B_siena
>> 3A_to_3B_siena
>>
>> 3) Clicking on the Stats tab, I select the "Full mode setup"
>> Number of main EVs : 3 (3 groups)
>> Next, I paste the design, assigning the 65 subjects to their respective groups (X,Y,Z)
>>
>> 4) Contrasts and F-Tests
>>
>> X - Y
>> X - Z
>> Y - Z
>>
>> Contrasts: 3
>> F-tests : 1
>> I clicked the greyed out F1 box beside each contrast. However when I tried to view design, an error pops up, saying that "each included contrast cannot be a linear combination of the others". What does this mean ?
>>
>> Thank you very much. Appreciate your help!
>>
>> Best Wishes,
>> Elijah
>>
>>
>>
>
|