Purchase this article with an account.
Daria Ivanchenko, Ziad Hafed, Frank Schaeffel; How correlated are drifts in both eyes during fixational eye movements?. Invest. Ophthalmol. Vis. Sci. 2018;59(9):5792.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
While microsaccades during fixation are highly correlated in both eyes, the literature is less coherent regarding drifts (Otero-Millan et al, 2014). If drifts show uncorrelated components, the positions of the images of a fixated target on the retina are no longer (mirror)-symmetrical. Symmetrical corrective microsaccades should then no longer be optimal for both eyes. To better understand this question, we studied interocular correlations in ocular drifts.
We built a high resolution binocular eye tracker with two USB3 infrared monochrome cameras having 640x480 pixel resolution and 320 Hz sampling rate (TheImagingSource, Model DMK33UX174). Both inputs were merged into one video buffer to synchronize binocular tracking, cutting the sampling rate down to 160 Hz. We developed software to precisely track both pupil centers and first Purkinje image centers generated by a field of IR LEDs (angular noise level <2 arcmin). After an automated self-calibration procedure, fixational eye movements were recorded while 6 subjects fixated a small target (13x13 pixels) on a display. A chin rest was used. Directions and magnitudes of ocular drifts were analyzed by a sliding regression algorithm (in steps of 20 ms) applied to either horizontal or vertical eye position traces in both eyes and correlating the slopes. Regression windows containing microsaccades were excluded.
While microsaccades were highly correlated both in time of occurrence and amplitudes (R>0.9, p<0.01, in all cases), as expected, their frequency was highly variable among subjects. One subject displayed no microsaccades over a period of 10 seconds. In the others, they occurred every 2-3 seconds. In 5 subjects, drift directions were significantly correlated in both eyes in the vertical direction (R=0.5, n=70, p<0.01) but uncorrelated or only weakly so in the horizontal direction. Drift directions were always uncorrelated in the subject with no microsaccades (R=0.17, n=59, p=0.21 for horizontal and R=0.02, n=59, p=0.85 for vertical).
Vertical ocular drifts are highly binocularly correlated but horizontal ones are less so. High variability of fixational eye movement patterns across subjects suggests different strategies for analysing drifting retinal images. Moreover, a memory for drifts may be needed to keep the eyes on target for extended time. Finally, the key question remains as to why drifts remain perceptually unnoticed.
This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.
This PDF is available to Subscribers Only