Dear Carol
> Having read Karl's response to my email about interscan intervals on
> 11/10/98, I have a further query. Must the interscan interval remain
> constant over the whole time-series, or is it possible to have one
> interval between individual scans within an epoch, and another interval
> between epochs?? Also, how important is the interval between epochs??
> Must this interval remain constant over the time-series??
The interscan interval is generally assumed to be constant for a
time-series. This is important if you use temporal smoothing or
convolve stimulus functions with a hemodynamic response function. It
is not important if you assume no temporal correlations and do not
smooth and have a regressor that accounts for differences in
acquisition rates.
The interals between epochs and the lengths of epochs can be as
variable as you want.
I hope this helps - Karl
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|