Evaluation of the infrared reflection method for saccadic eye movement velocity analysis

Abstract

Saccades are rapid eye movements that change the orientation of the eyes onto objects of interest. In an every day situation saccadic eye movements are used for e.g. reading, scanning objects and enviromnents.

The neuronal control of the saccadic eye movement is complex and involves many different parts of the brain. The saccade is the fastest of eye movements and its velocity is not subject to volitional control. Velocity measurements of saccades are therefore interesting for the detection and characterization of diseases that affect the oculomotor system of the brain and the extra ocular muscles.

Different techniques have over the years been developed for tracking eye movements. Among the methods, the magnetic scleral search coil (MSC) system has stood out as a "gold standard" for recording fast eye movements, as the saccade. The device is sensitive and recording noise is low. It also records at a high frequency level, which is particularly important when recording fast eye movements. However, one of the drawbacks of the method is that it involves wearing, on the surface of the eyes, rather uncomfortable silicone lenses in which a copper wire is embedded. There have been suggestions that the coil may slide on the eye surface and hence introduce an unwanted filtering effect. Moreover, the weight of the coil might affect the oculomotor signal elicited in the brainstem.

For conducting saccadic eye motility research and high velocity eye-tracking in the clinics, a non invasive recording system would be of great value. The infrared reflection (M) method is able to record the eye's position and velocity at a high frequency level, but without any contact with the eyes. The system works through emitting infrared light from illuminators built into goggles against the three-dimensional surface of the eyes. The reflected light is detected and processed in a computerized system into eye movement signals. The recent developments of the IR system have not been evaluated previously for detailed saccadic eye movement velocity recordings. In the presented studies healthy subjects performed saccadic tasks with both the MSC and IR methods. The data were analysed regarding the amplitude-velocity relationship, also called the main sequence. The results showed that the IR method generated saccadic velocities that were significantly higher compared to the MSC method. Intra individual as well as inter individual variability of the main sequence was shown with both methods, but there was generally larger variability with the IR recordings.

The reasons for the IR method generating higher eye velocity data compared to the MSC method may be explained by drawbacks that are related to both methods. First, there is an inherent risk of filtering in the MCS recordings if the silicon contact lens slides on the eye surface and the coil changes the oculomotor signal. Second, artefacts that are associated to the IR system may induce inaccurately high peak velocity measures. A risk that can not be left out of account is that the IR system does not merely detect eye movements, but also concurrent changes in the three-dimensional reflecting surface of the eyes and eyelids. It is also sensitive to changes in lighting conditions, which was shown in our experiments. Moreover, the inferior resolution and linearity of the IR system may introduce falsely high velocity data as well as a generally larger variability.

The IR system may be used for rapid eye velocity analysis and has some advantages compared to the MSC system. However, the velocity data differ from those measured with the coil system. Intra individual and inter individual variability should be considered, especially in follow-up studies. Improving the calibration procedure, paradigms and recording conditions might increase the scope for using the method for reliable saccadic eye movement recordings, both in scientific and clinical settings.