Spatial vision in the mesopic range: the effect of illumination spectra

Abstract

Purpose: The variation in the spatial resolving power of the cone and rod systems with the level of illumination is well established. In the mesopic range, when both cones and rods are active and operating at the limit of their range, the situation is more complex. Recently, the ability to fine-tune the spectral power distribution of LED lighting installations has led to the potential for optimisations of illumination spectra. These optimisations are limited by our current understanding of how cones and rods contribute to functional vision in the mesopic range. This study assessed visual acuity (VA) over some 4.7 log units change in illumination with emphasis on the mesopic range. A four-primary LED system was used to provide controlled rod and cone stimulation.

Methods: VA thresholds were estimated on an eink reflective display by a staircase procedure. The participants’ task was to indicate the orientation of a Landolt C by button press. The illumination was set to five different photopic luminances (0.001, 0.01, 0.1, 1 & 50 cd/m²). At each light level six different scotopic luminances were investigated; scotopic/photopic (S/P) luminance ratios varied from 0.5-5.5. Measurements were carried out at the fovea and 12 degrees in the periphery. Pupil diameter was recorded throughout and used to calculate retinal illuminance. To ensure adequate fixation eye position was measured and trials were repeated if fixation deviated more than ±1.5 degrees.

Results: The fovea was only affected significantly by variation in S/P luminance ratios at the lowest light level when VA thresholds were high and led to part of the stimuli lying outside the rod free zone of the retina. In the periphery S/P ratios significantly improved vision at light levels below 1 cd/m². In the mid-mesopic range (0.1 and 0.01 cd/m²) VA was improved with increasing S/P ratio, whereas at the lowest light level VA thresholds only improved only up to an S/P ratio of 2.5. The change in pupil diameter with variations in illuminants S/P ratio was small. Based on these findings we explore the possibility that rod-cone interactions play a role in the improvement of VA in the mid- to high-mesopic range.

Conclusions: In the mesopic range, VA in the periphery can be improved by increasing the S/P ratio of the illuminant. When vision is dominated by rod signals, VA improves significantly with increasing S/P ratio.