afni.nimh.nih.gov

> Would my statistical power be higher when the trials are more clumped
> together that when they are more spaced out?

Assuming that you are going to model each event separately, I feel the situation would be the other around: you would have higher statistical power with events less clumped together simply because they tend to be less correlated with each other.

See more discussion here:

[imaging.mrc-cbu.cam.ac.uk]

HTH,
Gang