Dear SPMers,
Some papers I read used experimental designs in which the ISI was selected from a random uniform distribution as to jitter the fMRI timeseries. However, I also read that the best method is selecting values according to an exponential distribution, in which shorter ISIs are more often used. Is there any practical sgnificance in using one method over the other, for instance, should selection of ISIs obey an exponential distribution if SOAs are short/long?
Relatedly, what would the optimal SOA for an erfMRI study be?
Thanks in advance,
Angela
|