1) What benefits would we achieve in our fMRI study by selecting
either "time" or "other" when asked about Parametric Modulation, rather
than "none". We are doing a study that involves three runs per subject,
with four identical trials in each run, but in a randomized order, so the
runs do not occur in the same order in each trial. We had previously
used "none" when asked about Parametric modulation, but wanted to know if
there were added statistical benefits inherent in utilizing this function.
2) What are the benefits and draw-backs of using a high-pass filter? When
using the high-pass filter, our data appears to be much more neat and
constrained. Is this a good thing? We are concerned that using this
option will remove valuable voxels from our results, causing us to miss
some important information when reviewing these results.
3) When we perform a high-pass filter, the session cutoff period (secs)
automatically displays "136 136 136". We are concerned about this because
when we calculated two times the experimental cycle, we got "128 128 128".
Is SPM known to be wrong in this calculation, or did we miscalculate. Our
design consists of 4 trials, each consisting of 8 images, with a TR of 4
seconds. We have a total of 72 images, leaving 40 images for 'baseline.'
There are 5 baseline conditions of 8 images each in each run, and baseline
precedes the first trial.
If anybody has any insight into any or all of these questions, we would
greatly appreciate answers to them.
Thank you,
Eric Long
|