Dear everyone,
I've just tried the script provided by Tor Wager for variable-duration modulation (a parametric modulator working on the duration, not the amplitude), but ran into problems right from the beginning.
SCN_Core_Support has been downloaded and added to the Matlab path. The model was specified with the latest SPM8 version via the GUI:
- one condition, onsets in seconds, duration set to 0
- SPM.Sess.U(1).P(1) is the time modulator (1st order time modulation)
- SPM.Sess.U(1).P(2).P reflects the durations in seconds, entered as parametric modulator (1st order)
After Xavg = onsets2fmridesign(onsets_in_sec, SPM.xY.RT, SPM.nscan, 'hrf'); I already receive a worrying message though:
Warning! onsets2fmridesign found 284 onsets after the end of the session and is omitting them.
Based on the model specifications none of the onsets is after the end of the session, SPM.xBF.UNITS = secs, SPM.nscan = 666 (~ no. of vol.), SPM.xY.RT = 2 (~ the TR), last onset is at 1326 s. The 283rd onset occurs at 666.178 s, so it seems the onsets might be treated as volume units for whatever reason. Any ideas?
Leaving this aside, should the duration be set to the average duration in seconds? Should trial-specific durations be entered?
Best,
Helmut
|