Hello, I would like to ask for advice regarding an unexpected change in EyeLink 1000 calibration behavior.
I am using an EyeLink 1000 (remote / arm mount setup) with pylink and SR Research SDK installed, running experiments via Python in VS Code (not PsychoPy Builder).
Recently, the system was updated from Windows 10 to Windows 11. Since this update, I have been experiencing a serious issue with calibration and validation behavior.
What worked before
Before the Windows update:
-
Calibration and validation were manually controlled
-
Each calibration point advanced only when I pressed SPACE/ENTER
-
The system waited until the participant clearly fixated the target
-
This behavior was stable across sessions using the same code
What is happening now
After updating to Windows 11:
-
Calibration and validation have become fully automatic
-
The tracker advances through calibration points without any key press
-
Fixations are accepted even when the participant is not actually looking at the target
-
In some cases, calibration/validation completes extremely quickly, producing inaccurate mappings
-
This happens even though:
-
The same experimental code is used
-
No autoAccept=True or similar option is explicitly set in the script
-
Manual acceptance options are enabled in the Host settings
-
In other words, the system now auto-accepts fixation and completes calibration on its own, making true manual calibration impossible.
Current setup details
-
Eye tracker: EyeLink 1000
-
Mount: Arm mount (remote mode)
-
OS: Windows 11 (previously Windows 10)
-
EyeLink Host version: v5.50
-
Calibration grid: 5-point (HV5) (cannot be changed due to an international collaborative study)
-
Tracking mode: Pupil-CR
-
Pupil tracking: Ellipse
-
Calibration options tested:
-
Pacing interval: 500 / 1000 / 1500 ms
-
Force manual accept: ON
-
Repeat first point: ON
-
Randomized order: OFF
-
Despite these settings, the system still accepts fixations automatically without waiting for user input.