Feasibility of size-modulation as a strategy to improve smartphone perimetry

Abstract

Purpose : Automated visual field-testing using bowl perimeters such as the Humphrey Field Analyzer use a fixed-size stimuli with varied luminance along a wide dynamic range that commercial 8-bit depth digital screens cannot accommodate. This study aimed to investigate the viability of using size modulation to exploit the spatial resolution of digital devices to increase testing capabilities of a smartphone-based visual field device.

Methods : Frequency-of-seeing (FOS) curves were generated at perimetric locations with normal and abnormal sensitivities in glaucoma and normal subjects using varied stimulus size and fixed contrast (MOCS procedure). Next, a group of normal and glaucoma patients with known visual field sensitivity underwent threshold testing using a Zippy Estimation by Sequential Testing (ZEST) procedure at four levels of contrast (stimulus luminance 11.5,13,15,20cd/m² against a background of 10cd/m²).

Results : MOCS testing successfully yielded FOS curves for 86% of locations (Fig. 1). Variability (FOS curve slope) was not correlated with sensitivity [R²=0.04, p=0.86]. FOS-derived size-thresholds showed decreasing sensitivity with eccentricity, correlation with luminance thresholds gave R²=0.45, p<0.0001 (Fig. 2). Control subjects using ZEST showed a ceiling effect at a contrast of 10cd/m², while glaucoma patients showed a floor effect at 1.5 and 3cd/m² in areas of moderate to severe damage. Neither group exhibited a ceiling or floor effect at 5cd/m².

Conclusions : MOCS testing showed that size-modulation is a viable strategy for smartphone perimetry, helping devices with limited contrast resolution but high spatial resolution conduct automated perimetry. Luminance needs to be set between 3 and 10 cd/m² above a background of 10cd/m² to avoid ceiling and floor effects in normal and damaged locations.

This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.

 

Figure 1: (A) FOS curves from a healthy subject at three locations show the probability that a stimulus of a given size was seen (Threshold=50% seen). Slope is related to threshold variability and remained steep in healthy (B) and glaucoma (C) cases. No correlation between threshold and slope.

 

Figure 2: (A) Size thresholds showed an expected decrease with eccentricity among normal patients. Glaucoma patient thresholds varied with damaged locations falling well below normal range (dotted line). (B) Size thresholds showed a highly significant correlation when compared to luminance thresholds.