We are planning an eye tracking experiment and would like to measure pupil size. Our stimuli are dynamic videos entailing human interaction, and they are about 10 seconds in length. To get reliable measures of pupil size, how should the baseline be designed? How closely do we need to match the baseline to the brightness/luminosity of the videos? How long do we need to show it, and do we have to show it before every trial? Thanks!
Hi Kathrin,
I'll send you our article on baseline correction and luminance
Jamie Reilly can you please also link the paper here? or share with me privately?
Thanks
Hi Kathrin,
if you do the experiment with the current generation of eye-trackers, I would not try to analyze the absolute pupil size, as this will always be distorted (foreshortening bias etc). You may rather go for a baseline-corrected relative value as dependent variable.
Concerning baseline assessment. One measurement over less than 5 seconds before the experiment should be absolutely sufficient. The baseline stimulus should be a grey screen that is matched in luminance to the mean of the video stimuli's luminance. You may also consider correcting measurements by the luminance of the presented videos. More information can be found here:
Kret, M. E., & Sjak-Shie, E. E. (2018). Preprocessing pupil size data: Guidelines and code. Behavior research methods. doi:10.3758/s13428-018-1075-y
DiCriscio, A. S., & Troiani, V. (2017). Pupil adaptation corresponds to quantitative measures of autism traits in children. Sci Rep, 7(1), 6476. doi:10.1038/s41598-017-06829-1
- Badges
- Science topic
- Similar topics
- Psychology