psychopy.org | Reference | Downloads | Github

Task crashing following eye tracker calibration

OS (e.g. Win10): Win10
PsychoPy version (e.g. 1.84.x): Latest

**What are you trying to achieve?: I have created a motor task where a cursor is moved to targets with a joystick. The targets disappear once the cursor reaches the target and the next one appears. I plan to collect data using eyelink 1000plus. I have used the components in builder to add calibration and validation of eyelink. The task starts and completes calibration and validation connecting with host PC running eyelink. When I press escape to get out of eye link setup the screen goes black and does not move on to the joystick task. If I remove the USB connecting the joystick and start again this does allow the screen to move on from validation to the task but it does not show the cursor. This suggests to me that the computer is struggling to process all the requirements at once. There is no specific error message. I noticed another topic where psychopy was struggling with speed due to eyelink so wondered whether the problem is the same. Any thoughts on this would be gratefully recieved. Can share details of the task if that would help. Dave

There are a couple known issues with the Builder eye tracking components when used with eyelink.

Until they are resolved you probably want to revert to using custom code to start / stop eye tracker recording etc. instead of using the ETRecord and ROI components.

Not sure what to suggest about the interaction with the joystick, if you are still having issues please send the simpliest task you can make that reproduces the problem.

Hi Sol

Thank you for the reply. I disabled the ETRecord and ROI components because of the reported problems but the issue seemed to be with ending of the validation and returning to the psychopy flow. I am considering collecting the eye gaze manually and therefore not control it from psychopy. The challenges with this is the data from the task and eyelink will be collected separately. My question is can I send a trigger from psychopy to the eyelink, to time stamp, via the ethenet connection. Dave

Do you mean the eyelink validation or the Builder ET Validation component?

Are you able to run the Experiments->visualSearch or Feature Demos->eyetracking Builder demos after switching it to use the EyeLink 1000?

Could you send me your project so I can try to reproduce the issue?

Thank you

Update: @dxm986 and I have figured out that the issue has to do with using the joystick in an experiment that also uses an ET Calibration component. It seems that the joystick is breaking the ability of psychopy to minimize and then maximize the full screen experiment window; only a black screen is displayed. This is the case regardless of which eye tracker is being used, including the MouseGaze simulator.

The current suggested work around is to have one experiment that just does the ET Calibration and then a second experiment that includes the actual task and starts / stops recording etc.