Optimising the glaucoma signal/noise ratio by mapping changes in spatial summation with area-modulated perimetric stimuli

www.nature.com/scientificreports OPEN Received: 7 August 2017 Accepted: 18 January 2018 Published: xx xx xxxx Optimising the glaucoma signal/ noise ratio by mapping changes in spatial summation with areamodulated perimetric stimuli Lindsay Rountree1, Pádraig J. Mulholland2,3, Roger S. Anderson2,3, David F. Garway-Heath3, James E. Morgan1 & Tony Redmond 1 Identification of glaucomatous damage and progression by perimetry are limited by measurement and response variability. This study tested the hypothesis that the glaucoma damage signal/noise ratio is greater with stimuli varying in area, either solely, or simultaneously with contrast, than with conventional stimuli varying in contrast only (Goldmann III, GIII). Thirty glaucoma patients and 20 age-similar healthy controls were tested with the Method of Constant Stimuli (MOCS). One stimulus modulated in area (A), one modulated in contrast within Ricco’s area (CR), one modulated in both area and contrast simultaneously (AC), and the reference stimulus was a GIII, modulating in contrast. Stimuli were presented on a common platform with a common scale (energy). A three-stage protocol minimised artefactual MOCS slope bias that can occur due to differences in psychometric function sampling between conditions. Threshold difference from age-matched normal (total deviation), response variability, and signal/noise ratio were compared between stimuli. Total deviation was greater with, and response variability less dependent on defect depth with A, AC, and CR stimuli, compared with GIII. Both A and AC stimuli showed a significantly greater signal/noise ratio than the GIII, indicating that areamodulated stimuli offer benefits over the GIII for identifying early glaucoma and measuring progression. Standard Automated Perimetry (SAP) is regarded as the current clinical standard for identifying glaucomatous visual field damage and change over time1. However, it has three cardinal limitations. First, SAP has poor sensitivity to early disease2, and although test-retest variability is lowest in early disease and in healthy individuals, it is unacceptably high for the identification of subtle damage3,4. Second, the greater variability in visual field locations with moderate damage (which increases with depth of defect), greatly inhibits the timely identification of change in those with established glaucoma4–6. Third, the test has a limited useable dynamic range, with test-retest variability spanning almost its entire range in individuals with advanced damage4,7,8, such that the measurement of remaining vision is difficult. Several studies have attempted to address the limitations of SAP by investigating the utility of alternative stimuli and comparing it to that of the clinical standard (Goldmann III). It has been suggested that employment of some alternative stimuli (e.g. the larger Goldmann V stimulus, area: 2.3 deg2) could enable measurement of a larger range of damage, with an accompanying reduction in test-retest variability7,9. This addresses the ‘noise’ component of the signal-to-noise ratio, but ascertaining whether such stimuli allow the test to maintain the same sensitivity to early damage (‘signal’) is not straightforward. In the absence of a clear rationale for using alternative stimuli, in terms of physiology, beyond reports that they may offer lower measurement variability, it is premature to confirm their superior utility, or otherwise, in clinical testing. Furthermore, a comparison of the utility of different stimuli on existing clinical instruments is not straightforward, particularly if it is not possible to precisely control their parameters, and without a precise knowledge of the workings of the thresholding algorithm employed. A comparison of stimuli with a non-clinical technique such as the Method of Constant Stimuli (MOCS) is also difficult if one is restricted to using the stimulus step size and scale provided on the clinical instrument. This becomes even more challenging if 1 School of Optometry and Vision Sciences, Cardiff University, Cardiff, United Kingdom. 2Optometry and Vision Science Research Group, School of Biomedical Sciences, Ulster University, Coleraine, N. Ireland, United Kingdom. 3NIHR Biomedical Research Centre, Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of Ophthalmology, London, United Kingdom. Correspondence and requests for materials should be addressed to T.R. (email: ) SCIeNtIfIC REPOrTS | (2018) 8:2172 | DOI:10.1038/s41598-018-20480-4 1 www.nature.com/scientificreports/ Figure 1. Schematic illustrating the change in spatial summation found by Redmond et al.31 GIII: Goldmann III stimulus, currently used in SAP. The distance between normal and glaucoma spatial summation curves for a stimulus of fixed area, varying in contrast is small in this region. ‘A’, ‘AC’ and ‘CR’ indicate an area-modulated, area-contrast modulated, and contrast modulated (within RA) stimulus, where the distance between the spatial summation curves is greater. comparing stimuli between different instruments. Such a restriction could well affect the resolution and accuracy with which the psychometric functions can be sampled for different stimuli, thereby increasing the risk of slope bias10–12. A full understanding of the diagnostic benefits of using alternative stimuli, requires the removal or minimisation of confounding factors that are unrelated to the stimulus configuration, such as the thresholding algorithm or unequal psychometric function sampling between stimuli. The optimisation of stimulus parameters for use in SAP to maximise the signal to noise ratio (SNR), should be based on the underlying physiological mechanisms being measured. Spatial summation describes the way in which the visual system integrates light energy across the area of a stimulus. Ricco’s law states that, for a range of small stimuli, within a critical area (Ricco’s area, RA), the intensity of the stimulus at threshold is inversely proportional to its area (threshold x area = constant)13. This is referred to as ‘complete spatial summation’. Beyond RA, spatial summation is incomplete. RA is not a constant value, and has been found to vary with visual field eccentricity14–17, retinal illuminance18–20, and stimulus duration16,21. Traditionally, RA was thought to have a physiological basis at the retinal level18,22,23, however increasing evidence indicates that it is likely a perceptual result of spatial filtering at multiple hierarchies of visual processing, in the retina and visual cortex24–29; i.e. the ‘perceptive field’26,30. An enlarged RA has been found in patients with primary open angle glaucoma, and differential amounts of sensitivity loss to a range of stimulus sizes can be mapped to a lateral shift in the spatial summation function31. The finding has important implications, not only for a better understanding of the pathophysiological changes that occur in glaucoma, but also for the development of methods to identi (...truncated)