Dear SPMers,
we conducted a single case study of a rare patient comprising pre- & post-treatment fMRI (short intervention, both acquisitions on a single day).
During the task, picture ratings were assessed via response grips (right hand button press moving upwards on the visual analogue scale (VAS), left hand press moving downwards). Accidentally, the subject switched the right and left response grip in the second session, so that ratings on the VAS were given by opposed button presses as compared to the first session (left hand up, right hand down).
To control for confounding motor activity when comparing both sessions, we included the button presses’ as separate regressors of no interest (separate regressors for right & left button). The analyses revealed stronger activity (among other regions) in the motorcortex post-treatment. The finding would make sense in terms of the intervention – however, it could also be an artifact of the switched response buttons & thus between-session differences in the motor response.
Unfortunately these patients are very rare & our paradigm cannot be simple applied again – therefore we would like to control the potentially confounding motor effects in the current data-set.
My question: does modeling the button-presses allow to exclude confounding motor effects? Is there an alternative more conservative approach to control for confounding effects of the switched response grips?
Thanks in advance & best regards,
Ben
|