Olga Prilipko wrote:
>> Dear SPMers,
>> After unsuccessfully searching the archives and the net I would very
>> much appreciate your help on the following:
>> I want to identify which regions of the brain show an activation
>> pattern that follows 1) linear and 2) a quadratic trend with the
>> increasing task load.
>> I have 4 levels of load and was going to define the matrix with 1
>> condition (all onsets of all 4 loads) with a categorical parametric
>> My questions are:
>> 1) how should I enter the parametric modulation vector? Would 1,2,3,4
>> for the 4 loads be correct or should I orthogonize this and enter -1.5
>> -0.5 0.5 1.5(for linear) and 1 -1 -1 1 (for quadratic) instead?
Just use [-1.5:1.5] for the linear trend and then choose a polynomial
expansion of 2. SPM will the automatically generate the regressors for the
quadratic trend and orthogonalizes it w.r.t. to the linear trend.
You will end up with 3 regressors:
2. linear trend
3. quadratic trend (orthogonal to 2.)
>> 2) should I specify 2 parametric modulations (with identical vectors)
>> one being described as linear and the other as quadratic in order to
>> tease apart regions that may follow different trends ?
>> 3) finally, when defining the contrast, should I specify 0 1 0 if I
>> want to visualize the regions with the linear trend and 0 0 1 with
>> quadratic trend (the 1st regressor beeing the condition)?
Correct. 0 1 0 test the linear trend (increase), 0 -1 0 tests a linear
decrease across load conditions, and 0 0 1 test for the quadratic trend.
>> Thank you very much, any help will be very much appreciated.
>> Olga Prilipko
Jan Gläscher, Ph.D. Div. Humanities & Social Sciences
+1 (626) 395-3898 (office) Caltech, Broad Center, M/C 114-96
+1 (626) 395-2000 (fax) 1200 California Blvd
[log in to unmask] Pasadena, CA 91125