I had also got confused and thought that the new version allowed you
to get PPI data for all runs at once but obviosuly, like you say, this
is done independently still because of the use of independent VOI
files for each run... it all becomes clear!
I have now used the new function in SPM tasks>stats>PPI and then
looked at the PPI.ppi, PPI.P and PPI.Y for a single subject whos PPI
data I had already obtained using the old PPI button on the front of
the main GUI, and it seems that they both give the same results so I
guess I am ok to continue with my old PPI.mat files.
Thanks for your help
Sophie
Quoting Darren Gitelman <[log in to unmask]>:
> Sophie:
>
> The input should be an n x 3 matrix. n = number of conditions you are
> including in your PPI. For example there might be 3 conditions in your
> design, but if you only use 2 in your PPI, then n will be 2.
>
> The columns are #1: Index into SPM.Sess.U. Basically the U's are the list of
> your conditions for a particular session. So SPM.Sess.U(1) is the first
> condition and so on. You need to know the order of your conditions in your
> GLM design for the SPM.mat file that you choose.
>
> Column #2 is an index into the name of the condition
> (SPM.Sess.U(i).name(ii). ) This will always be a 1 unless you have
> parametric effects, then you will have to look at the SPM structure to get
> the proper index.
>
> Column #3 is the contrast weight for the condition you are including in the
> PPI estimation.
>
> Example from the attention to motion dataset: Let's say you have 3
> conditions: Stationary, NoAttention and Attention in that order, and you
> want to examine a PPI for Attention - NoAttention. Then the entry would be:
>
> [ 2 1 -1; 3 1 1]
>
> So [2 1 -1] means the NoAttention condition given a -1 contrast weight, and
> [3 1 1] means the Attention condition with a +1 contrast weight.
>
> You don't need to worry about the Session number because that is
> automatically specified by the VOI file you choose. Notice that the VOI
> files end in a number _1 is the first session, _2 the second session and so
> on.
>
> Hope this helps,
> Darren
>
> On Thu, Jul 1, 2010 at 11:29 AM, Sophie Green <
> [log in to unmask]> wrote:
>
>> Hi all,
>>
>> I'm trying to do PPI in SPM8 and have been using the PPI button on the main
>> GUI to create my PPI terms, the way I have done in SPM5. I have done this
>> for almost 60 people for 3 runs.
>>
>> I've just realized that PPI is now included separately in the job/batch
>> editor under task>stats> which will create the PPI terms for all three runs
>> together. I'm wondering whether there are any other differences between the
>> two methods? Basically, do I need to re-create the PPI files using the new
>> all-in-one way?
>>
>> I've tried to check this out myself by creating the PPI terms for one
>> participant using the new method and comparing them to the old method that
>> can still be used in SPM8 but I'm struggling with understanding the 'input
>> variables and contrast weights' part. If any one could shed any light onto
>> this for me that would be great. I've checked the manual and see that column
>> one is SPM.Sess.U(i) and second column is SPM.sess.Uiiname but I'm
>> struggling to understand what this means.
>>
>> Please forgive my ignorance!
>>
>> Any help at all would be much appreciated,
>>
>> Kindest regards,
>> Sophie
>>
>
--
Sophie Green
PhD Student
Neuroscience and Aphasia Research Unit (NARU)
Zochonis Building, 3rd Floor
Brunswick Street
University of Manchester
Manchester M13 9PL
Tel: +44 (0)161 306 0451
|