Dear all,
I am analysing a drug study with fMRI. Each subject was scanned on
two occasions, once on drug once on placebo. For each occasion, there
was also a pre and post infusion session. Thus we have a total of 4
sessions: 1. pre-drug, 2. post-drug, 3. pre-placebo, 4. post-placebo,
each 96 scans long. We also have a 96 item regression vector
calculated based on the physiological/subjective responses to the
drug. I want to use this as a user-specified regressor to model
haemodynamic response to the drug.
In my analysis I have said I have 4 identical sessions. I am forced
to put in 1 condition (spm will not play with zero
conditions) which I am therefore entering as the steady state
response, ie one block 96 scans long. I then get to put in my
regressor. SPM whirrs away cheerfully and gives me an 8 column
design matrix with "steady state" response session 1 , followed by
"regressor" session 1, "steady state" response session 2 , followed
by "regressor" session 2 etc.
I am not interested in the " steady state" responses so I have used
0s in those columns when it comes to the design matrix and
concentrated on the "regressor" columns. I want to know where that
regressor is a good model for the haemodynamic response in the
post-drug session but not pre-drug or post-placebo sessions. Thus,
for example my key contrast (effectively the interaction term that
gives drug-placebo post-pre) is 0 -1 0 1 0 -1 0 1.
If anyone has followed thus far, I would be grateful for any comments
on whether this is a) an acceptable way to model the data and b) the
best way to model the data. I am a little uncomfortable with using
the regressor to model all 4 sessions, when it only has any meaning
for one of them. But then comparing this drug regressor for the
session where subjects actually had drug with a flat response in the
other 3 seems dubious too. Perhaps the model should be as it stands
but with the key contrast as 0 0 0 1 0 0 0 0??
Any help would be much appreciated
Thanks a lot
Rebecca
Dr Rebecca Elliott
Neuroscience and Psychiatry Unit
Room G907, Stopford Building
University of Manchester
Oxford Road
Manchester
M13 9PT
Phone: 0161 275 7433
Fax: 0161 275 7429
e-mail: [log in to unmask]
|