Dear SPMers,
Two questions on design issues for event-related
experiments:
(1)
We intend to run a 2x2 factorial event-related design,
employing a scanning sequence with TR=3.126 secs and a
random jitter to introduce a temporal shift of stimulus
presentation across the whole TR. We intend to model the
hdr with a hdf with time and dispersion derivatives. I was
wondering how many repetitions of each of the four events
per subject one would need as the *minimum* in order to be
able to measure the hdr with acceptable reliability (rfx
analyses are planned). As a rule of thumb, I came across
(TR in secs) * 10 = 30 repetitions, but we may want to even
use a lower number for time constraints. Does anyone have
experience knowledge or, even better, know references that
discuss that issue?
(2)
What is the minimum SOA/ISI for event-related studies
employing modelling with a hdf? I know some papers and
discussions on the list that state that one can go as low
as 2 seconds for more elementary sensory and attentional
processes. However, we are examining more complex reasoning
and decision processes that sometimes lead to long response
times (> 2 seconds). Does anyone have experience with the
minimum SOAs/ISIs of such designs, especially with the
required mean temporal distance between end of the response
and beginnig of the new stimulus presentation?
Any help would be greatly appreciated,
Christian
--
Christian Ruff
Center for Cognitive Science, IIG
University of Freiburg
Friedrichstrasse 50
D-79098 Freiburg
Germany
Phone: +49 761 203 4966
Fax: +49 761 203 4938
|