Dear SPMers,
I have a simple fMRI event-related design with Words, Pseudowords, and a
Fixation Cross (which is a baseline condition).
To monitor participants’ vigilance in the scanner, I complicated it
slightly by introducing a star ‘*’ stimulus which is displayed in black
colour during the ISIs, between words, pseudowords and fixation crosses
(randomly mixed). On 10% of trials the star changes colour and the
participant is instructed to press a button.
In my first analysis, I was advised not to model the baseline condition in
the DM because the error term is over-specified. So I had the following
regressors: Words HRF, Words Time Derivative, Words Dispersion Derivative,
Pseudowords HRF, Pseudowords Time Derivative, Pseudowords Dispersion
Derivative and six movement regressors (to account for participants’
movement in the scanner).
As I am interested in BOLD signal in cerebellum due to reading, I started
to worry about my button press and was advised to model it as a separate
condition. This made me think about my star stimulus displayed during the
ISIs. It seems to me that because I did not specified it as a regressor in
the DM, my baseline (which was also not specified) has been confounded.
Do you think that the DM for this experiment would be improved if I include
the star as a separate condition, and as before do not include the baseline
as a separate condition? (Of course, I would also have all regressors
listed above, plus the button press, as a separate condition).
The black star was displayed for 3000ms, Can I enter ‘0’ for duration, as
for all the other stimuli, i.e., Words (which had duration of 1000 ms each)
in this event related design, or do I need to specify exactly 3 sec?
Many thanks for your help. All the best,
Agnieszka
--------------------------------------------------------------------
mail2web - Check your email from the web at
http://link.mail2web.com/mail2web
|