Dear SPM experts,
I sent the message below earlier today and wanted to add the following:
Could I model the categorical variable by using dummy variables and enter
them as parameters convolved with the trial-specific HRF:
trial 1 with e.g. 6 events of this trial type:
adjective 0 1 0 1 0 1
noun 1 0 0 0 1 0
verb 0 0 1 0 0 0 ????
Does that make any sense ?
Thanks for any feedback ....
My question might be trivial to the community, but I just wanted to make
sure not to get something wrong here. It concerns whether I can enter a
categorical variable (nominal scale) as a "parametric modulation" of the
trial-specific HRF, although it is not a continuous variable.
I am using SPM5 to analyze a study using 3 different conditions of word
pairs and null events.
I have several psycholinguistic parameters, which I want to use as
parametric regressors. Some as regressors of interest, some of no interest.
Under fMRI model specification -> Data & Design -> Conditions
There is the option "parametric modulation" to enter a vector of values
for each trial of a given condition (say word pair type 1)
I figured this was the best way to enter my parameters, since I want to
test whether they interact with the HRF to a specific type of word pair.
Most of my parametric variables are continuous and most of them normally
distributed.
However I have one non-parametric factor (word class, e.g. verb, noun,
adjective, overall there are 6 different levels of this factor) which I
would like to include as a confound. Each condition (word pair type)
contains different types of words.
Would it be correct to enter this categorical factor as a "parametric
modulation" ?
Thanks for any help !
Best,
Roland
Roland Zahn, Dr. med.
NIH / NINDS
Cognitive Neuroscience Section
Building 10, 5C206
10 Center Drive, MSC 1440
Bethesda, MD 20892-1440
Tel.: (+1)-301-402-6392
Fax.: (+1)-301-480-2909
|