Related Experiment Video
Updated: Feb 23, 2026
Using Eye-tracking to Assess the Relative Importance of Visual and Vestibular Input to Subcortical Motion Processing in the Roll Plane
Published on: August 22, 2025
Multi-centre comparison of five eye movement detection algorithms
Alpo Värri1, Bob Kemp1, Agostinho C Rosa1
1Signal Processing Laboratory, Tampere University of Technology, Tampere, Finland,Department of Clinical Neurophysiology, University Hospital Leiden, Leiden, The Netherlands,CAPS, IST, Technical University of Lisbon, Lisboa, Portugal,Department of Medical Informatics and Image Analysis (MIBA), Aalborg University, Aalborg, Denmark,Medical Policlinic of Philipps University, Marburg, Germany,Department of Clinical Neurophysiology, Tampere University Hospital, Tampere, Finland.
Comparing automatic sleep eye movement detection algorithms is challenging. This study objectively compared five algorithms using identical data, finding electrode montage and EEG delta activity crosstalk impact accuracy.
Area of Science:
- Sleep Medicine
- Neuroscience
- Biomedical Engineering
Background:
- Objective comparison of automatic sleep eye movement detection algorithms is difficult due to varied recording setups.
- Standardized evaluation is crucial for advancing automated sleep analysis.
Purpose of the Study:
- To objectively compare the performance of five different automatic eye movement detection algorithms.
- To assess the impact of electrode montage and electroencephalogram (EEG) delta activity on algorithm accuracy.
Main Methods:
- Five distinct eye movement detection algorithms were applied to a single, standardized dataset.
- Algorithm outputs were compared against visually scored data.
- Performance was evaluated using true and false detection rates across different threshold levels.
Main Results:
- Algorithm performance varied, with best results achieved when using the electrode montage for which the algorithm was originally designed.
- Using non-native electrode montages significantly decreased detection accuracy.
- Subtracting EEG delta activity crosstalk improved the correctness of eye movement detections.
Conclusions:
- Standardized data and comparison are essential for evaluating sleep eye movement detection algorithms.
- Algorithm performance is highly dependent on the electrode montage used.
- Mitigating EEG delta crosstalk enhances the reliability of automated eye movement detection during sleep.
