Purchase this article with an account.
R.A. Schuchard, D. Ross; Evaluating retinal, pupil, and EOG eye trackers for eye movement measurements in low vision . Invest. Ophthalmol. Vis. Sci. 2004;45(13):5450.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: To compare the efficacy for three basic types of eye movement measurement systems: 1) Retinal tracking (scanning laser ophthalmoscope, SLO); 2) Pupil/corneal reflection tracking (ISCAN); and 3) EOG tracking (Colbourne). Methods: Eye movement data was simultaneously collected with all three systems from 15 normally sighted adults (5 in 4 age groups: 1) 20–40 yo; 2) 41–60 yo; 3) 61 to 80; and 4) 81 or older) as well as 11 people with central scotomas in both eyes. Each person made successive fixations on a 3 x 5 grid of fixation targets that were five degrees apart (horizontally and vertically) from each other. Each subject viewed the fixation grid in the scanning laser ophthalmoscope while eye movement measurements were simultaneously recorded in each system. Eye position as well as eye movements (fixation and saccade) were analyzed from each system. Results: An automated retinal image movement measurement analysis using Matlab (MathWorks) was developed and validated (by comparison to manual registration). A comparison of the 2 dimensional eye movement found no correlation of eye position or movement (including saccade amplitude) between EOG and either retinal or pupil eye movement measurements. Fixation areas derived from the pupil system had good correlation (r=0.72) to fixation areas found by retinal tracking but absolute eye position and absolute saccade length was not closely correlated between the two systems (r=0.38 and r= 0.33, respectively). In terms of age range, there was no significant difference in the eye movement measurements (fixation and saccade movements) between the age groups except the oldest (81 or older) while the visually impaired group had significantly different results compared to all other groups (e.g., fixation area at 95% frequency ellipse of 4.5 compared to 1.8 degrees for normally sighted and a saccade gain of). Conclusions: While pupil and EOG calibration was verified for this study due to simulataneous retinal monitoring during calibration, a general method was developed and will be proposed for accurate calibration of subjects with central scotomas and multiple fixation areas (PRLs) without retinal monitoring. While EOG systems are ideal when the anterior or posterior surfaces of the eye cannot be viewed (e.g., using low vision devices), the EOG system cannot be recommended for monitoring eye movements. The pupil/corneal reflection systems provide reasonable results while the retinal tracking system generally provides the highest accuracy.
This PDF is available to Subscribers Only