Abstract
Objectives: The human retina contains five photoreceptors:
The three spectral classes of cones — the S cones (peak spectral
sensitivity λmax ~420 nm), the M cones (λmax ~530 nm), the L cones
(λmax ~558 nm) — , rods ( λmax ~495 nm), and melanopsin ( λmax
~480 nm). Short-wavelength light is known have important effects
on what is conventionally called “non-image-forming” visual
function, including circadian photoentrainment, phase shifting and
the pupillary light reflex.
Methods: By presenting stimuli which are matched in the
activation of a given set of photoreceptors (so-called metamers)
but differ in activation of the target photoreceptors, the contribu-
tions of the different photoreceptors can be mapped out. We have
previously used this method to examine the properties of the hu-
man pupillary light reflex, and encoding of melanopsin-targeted
stimuli in human visual cortex (using fMRI). Here, we are examin-
ing metameric illuminants at five chromaticities along the daylight
locus (correlated colour temperatures 25,000K, 10,000K, 6,500K,
5,000K and 4,000K) differing in melanopsin activation by 200%
and measuring the pupil size in human observers under naturalistic
viewing conditions.
Results: While the colour rendering properties of these meta-
meric illuminants are generally poor, they elicit physiologically
differentiable pupil sizes by up to 20%. This is not a surprising
result, as it is known that melanopsin contributes to the pupil under
conditions of silent substitution; but it provides evidence that large
melanopsin contrast in isolation seen in naturalistic viewing condi-
tions has a tangible physiological effect.
Conclusions: This research represents a step towards pro-
ducing carefully controlled metameric melanopsin-targeted stim-
uli under naturalistic (e.g. sleep lab, home setting, office) viewing
conditions with a demonstrable physiological effect.
