Hi Mayuresh,
> Does having an interval affect the experiment in terms of effective sampling
> of the hemodynamic response or efficiency of the design?
Offhand, this design sounds pretty good to me. Two considerations for
relatively rapid event related designs (that conflict with one another) are
that you would want as many stimuli as possible to achieve as much cognitive
activity as possible, and also that the stimuli are spaced well enough, and
jittered enough in time to be able to extract the separate event-related
responses. Allowing some of the stimuli to immediately follow just seems to
tilt the design slightly more in favor of the first goal.
One good but slightly older reference for jittering can be found in an Ollinger
paper around 2001- 2002.
> what would you suggest for such an experiment? Also I am interested in the
> comparison of different blocks rather than differences between stimuli types
> within each block.
If you are not interested AT ALL in the event related responses, then you can
have the stimuli follow each other within the blocks as rapidly as possible,
and not model the event-related part. This would make your experiment a pure
block design, and should make it a relatively powerful one. If you want to get
an event-related response to the individual stimuli, then such a mixed design
would probably be best accomplished with slightly longer blocks, and jittering
of individual stimuli within the block.
A few good papers on mixed designs come from K. Visscher, and Donaldson.
Ken
|