I collected EDA data during sessions of the Iowa Gambling Task (IGT). I decided to compute skin conductance responses (SCRs) by simply measuring the absolute difference between onset and the maximum point within a 4-second window for each IGT event. I decided against computing for local baseline since trials were close to each other in time and prior SCRs would confound with a baseline computation. Am I right to do it this way?
Of corse, peak-to-peak amplitude is a significant measure for EDA measurement
With the short ISI the previous responses can (and will) still have effect due to the additive nature of the skin conductance signal. Ie if there's a previous response that has yet to be peaked, it will cause inflation of the current response and if there's one that is decaying from the peak it will cause deflation. Of course if there's proper randomization, this should just increase the error variance.
For short ISI settings the best way would be probably to remove/lessen the SCR overlaps via deconvolution. There are couple of open source packages for this: http://www.ledalab.de/ http://scralyze.sourceforge.net/
- Badges
- Science topic
- Similar topics
- Biological Science
- Bioecology
- Systems Ecology
- Landscape Ecology
- Connectivity