For a mixed blocked/event-related study, I would like to model blocks and
events together. If I understand Visscher et al. [NeuroImage 19 (2003)
1694–1708] correctly, I would need to model/deconvolve blocks and events
with different functions (blocks: gamma/boxcar; events: delta regressors at
time zero and additional regressors for each of the following time points,
covering 14-20 seconds total), but I am unsure how to do this with SPM5.
My design has 3 block-types (ABC) and two event-types (1,2), resulting in a
factorial design: 1A, 1B, 1C, 2A, 2B, 2C. (=6 conditions). The 60s blocks
are separated by 30s baseline blocks and I have additional null events
within task blocks.
Questions:
1) Do I need to select different functions and if so how?
2) How would I model the additional time points explicitly? And what is the
difference between including the events as 'conditions' or 'regressors'?
3) Do I need to model baseline blocks and/or null trials?
Thanks for any advice!
Esther
|