Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Supported by

Re-calibration of eye-tracker during the experiment

Hi there,

I am programming an eye-tracking change-detection experiment using OpenSesame 3.0.7 on an old Tobii120 machine.

I am using the wonderful Python code by Hiroyuki Sogu to interface with the Tobii SDK. I got a basic experiment running, and it seems to be collecting decent data. However, there are couple of things that I have still not yet fully understood: (a) say I want to re-calibrate my participant, every so often, for example, by pressing a key on the keyboard (and call the eye-tracker), is there a way I could do it without interrupting the experiment, and risk to loose it all? (b) is it possible to perform drift correction for every trial? TobiiController does not look like it allows that ... (c) is there a way I could somehow monitor whether the gaze of my participant drifted completely off during the course of the experiment? (I have two screens connected to my Tobii machine).

Thanks a lot for your time!

Comments

  • Hi,

    First of all, I am not familiar with Tobii eye trackers. Assuming that the works in a similar way as Eyelink eyetrackers, the following advice might be useful to you:

    (a) say I want to re-calibrate my participant, every so often, for example, by pressing a key on the keyboard (and call the eye-tracker), is there a way I could do it without interrupting the experiment, and risk to loose it all?

    Yes, you can recalibrate a subject, every time your experiment is executing the drift_correct

    (b) is it possible to perform drift correction for every trial?

    Yes, this is also possible (see my response before). The only thing you need to keep in mind is that the drift correct cannot be executed while the eyetracker is recording. So, only if you start and stop the eye tracker on a trial-by-trial-level you can run the drift_correct. In this case you should keep in mind that it takes a while to start and stop the eyetracker (~300ms), so if the timing of your intertrial interval is important, you need to be careful.

    (c) is there a way I could somehow monitor whether the gaze of my participant drifted completely off during the course of the experiment?

    ehm, probably. Basically you can test the calibration at the very end of your experiment, and see how far off the fixations are. So, define a couple of points on the screen, and instruct your subjects to fixate them one by one (maybe make them also press a key once they think they are fixating the points) and save their fixations. Then you can simply take the distance between the points to evaluate your calibration.

    Does this make sense? Again, not sure whether this is useful for Tobii eyetrackers. If not, then sorry for wasting your time :)

    Eduard

  • edited October 2016

    yeah, that makes a lot of sense. cheers Eduard :) I will try to implement few of the things you suggest, and post here back whether it worked or not. Not sure if Tobii allows for drift correction ... as far I can see in the code is no-where mentioned, but I will keep looking. Your response to (c) is truly a good suggestion, and it is easy to implement. God bless SR and Eyelink anyway, they are far better than anyone else in this business.

Sign In or Register to comment.