Dear FSL users,
I have some data of a event related experiment with 12 runs per participant. We're interested on plotting the individual (subject level) time-series of the whole experiment. Working at the single-subject level, so far I've extracted a time-series for each run, but:
1) If I look at the time-series after putting all runs together in a fixed effects analysis, the X-axis time points are 12 (that is, one per run).
2) If I try to concatenate every preprocessed run and then run a first-level analysis, I get a time-series for the whole experiment. However, I've read some other threads here which discourage this approach. Indeed, the time-series pattern looks strange, with nasty jumps in the Y-axis between runs.
My questions then are:
- For single-subject data, is it possible to obtain a time-series plot with all volumes in the X-axis, instead of just one time point per run?
- Why do I get those jumps when using concatenated data? If I understand well, high-pass filtering should account for intensity differences between runs...
- Additionally, I read on the Feat manual that registration to the template space is already done in the first-level analysis, but FSL will plot data on the native brain for the sake of saving disk space. I wonder if there is any way of actually plotting this first-level stats (that is, single run data) onto a template.
Thanks in advance!
Carlos
|