A New Quantitative Technique to Determine the Number and Extent of Preferred Retinal Loci in Macular Disease

Abstract

Abstract: : Purpose: It has been shown that patients with macular disease and central loss of vision have poor fixation stability. Further, it has been reported that in some patients more than one retinal area is used for fixation. The number and stability of loci used for fixation will affect the rehabilitation of these individuals. We have applied a new quantitative technique to determine the number of preferred retinal loci (PRLs) used during an episode of fixation and the extent of each locus. Methods: Six adults with macular disease and unstable fixation were selected from patients seen in the laboratory. Patients were asked to fixate a round target of 3° diameter with an 18' central detail on a computer monitor for 10 seconds. Eye position was measured using an infra-red eyetracker (SensoMotoric Instruments Gazetracker, Germany) at a sampling rate of 250Hz. A "global" bivariate contour ellipse area (BCEA) was calculated for all the fixation points together, as has previously been performed. Data were further analysed using a non-parametric modelling technique (the kernel density estimator) to calculate the number of PRLs present. In the event of multiple PRLs, the EM (expectation, maximisation) algorithm was applied to calculate the parameters of each PRL, including the "individual" BCEA. Results: In 4 patients (67%) the kernel density estimator indicated there was only one PRL (BCEA range 1160 to 2160 minarc₂). However in two patients multiple PRLs were found and "individual" BCEAs (range 586 to 2000 minarc₂) provided a better fit of the data set than one "global" BCEA did (range 2760 to 6690 minarc₂). There was a marked difference between the sum of the "individual" BCEAs and the "global" BCEA in these patients. Both of the patients with multiple PRLs had recent (< 1 month) vision loss in the tested eye. Conclusion: Some patients with unstable fixation secondary to central scotomas display multiple PRLs when data are analysed with the kernel density estimator technique. In these patients fixation patterns are best described by assigning a BCEA to each PRL individually using the EM algorithm rather than using the traditional method of calculating a "global" BCEA. Patients with a recent change in visual acuity seem most likely to have multiple PRLs. These quantitative techniques will be of use to those involved in the investigation and rehabilitation of patients with central scotomas.

Keywords: 460 macula/fovea • 459 low vision • 408 eye movements: recording techniques