09-04-2020, 05:02 PM
Automating the EyeLink Drift Correction procedure itself is not currently supported - the Host PC software requires manual input to confirm that the drift is acceptable and the trial can proceed. A manual decision is required because automatically accepting a Drift Correction has a high probability of accepting a false positive (e.g. when any "drift" is actually the participant failing to accurately fixate the target). This would decrease the spatial accuracy of the recorded data.
If you have a task with a large number of short trials, you may wish to implement a Drift Check every Nth trials. It is also possible to skip the Drift Check procedure entirely, and replace it with a "Pseudo Drift Check". By monitoring the participant's gaze position at the start of the actual recording sequence, it is possible to determine if the participant has stably fixated a target (within a specific boundary) for a given period of time. If these conditions are met, then the trial will proceed. If the participant cannot maintain a reasonable fixation on the target within the time limit (for example if the tracker is reporting gaze inaccurately due to drift), then the trial will be halted, and the experiment will instruct the Host PC to perform a genuine drift check, or go to the camera setup screen - thus allowing for a manual recalibration / validation to remove the drift. Once the participant has been recalibrated, recording can resume and the experiment will continue from where it left off.
The attached template illustrates how this can be implemented in Experiment Builder.
Simple_PseudoDriftCheck.ebz (Size: 190.82 KB / Downloads: 53)
If you have a task with a large number of short trials, you may wish to implement a Drift Check every Nth trials. It is also possible to skip the Drift Check procedure entirely, and replace it with a "Pseudo Drift Check". By monitoring the participant's gaze position at the start of the actual recording sequence, it is possible to determine if the participant has stably fixated a target (within a specific boundary) for a given period of time. If these conditions are met, then the trial will proceed. If the participant cannot maintain a reasonable fixation on the target within the time limit (for example if the tracker is reporting gaze inaccurately due to drift), then the trial will be halted, and the experiment will instruct the Host PC to perform a genuine drift check, or go to the camera setup screen - thus allowing for a manual recalibration / validation to remove the drift. Once the participant has been recalibrated, recording can resume and the experiment will continue from where it left off.
The attached template illustrates how this can be implemented in Experiment Builder.
Simple_PseudoDriftCheck.ebz (Size: 190.82 KB / Downloads: 53)