Abstract
Background: Entrainment of the circadian pacemaker de-
pends on both the intensity and spectral distribution of light reach-
ing the retina, in addition to other contextual modulators. Daily
ocular exposure to light is mediated by materials in our environ-
ment, whose reflectance properties manipulate incident light spec-
tra in ways which are not easily predictable. We refer to the set of
seen spectra as the ‘spectral diet’ which is a subset of all possible
spectra. The spectral diet contains useful information about how
an individual’s circadian system will respond to new stimuli over
time.
Methods: Here we outline some approaches to decode spec-
tral information in a way which maximizes insights about the bio-
logical response to changes in spectral distribution, while reducing
the dimensionality of our dataset. For our preliminary analysis we
chose to generate a set of possible spectra from existing illuminant
and reflectance datasets, quantified their effect on the five human
photoreceptor classes (CIE S 026), and looked at both relative and
absolute quantities.
Results: Toward this end, we present a dimension-reduced
framework with five photoreceptors. The next step will be to as-
semble a database of actual spectra from subjects in the built en-
vironment using novel, small-footprint wearable spectrometers.
Conclusion: We hope to use our method to identify a set of
‘basis spectra’ which are observed with high frequency in real life
to illustrate how statistical regularities in real world material prop-
erties and illuminant spectra further reduce the dimensionality of
the problem.
