Undergraduate Honors Theses

Thesis Defended

Fall 2018

Document Type


Type of Thesis

Departmental Honors


Integrative Physiology

First Advisor

Celine Vetter

Second Advisor

Mark Opp

Third Advisor

Lameese Akacem


While laboratory studies have demonstrated that not only light exposure intensity, but also its duration, timing, and spectral composition are relevant for circadian physiology, most field studies examining links between light, circadian rhythms, sleep, and behavior rely on simple quantification of light intensity. Unpublished findings from the Circadian and Sleep Epidemiology Laboratory indicate that continuous light exposure levels of free-living individuals measured at the wrist are best captured by eight independent dimensions, represented by metrics including light intensity, timing, spectral composition, and the variability in timing of light exposure. However, wrist-level recordings of light have been shown to have error beyond relative measurement error, when compared to eye-level recordings. This study therefore uses eye-level light exposure data previously collected by Dr. Celine Vetter and colleagues in Munich, Germany to examine whether light exposure patterns measured at eye-level would show similar patterns as the light exposure patterns measured at wrist-level. K-means clustering algorithms were used to partition eye-level recordings (N=23, 5 days of recording) and results were compared to the wrist-level data from 2,154 individuals studied by Vetter et al (unpublished). Results indicate that light exposure profiles at eye-level are described by a similar set of patterns as those identified from wrist-level measurements, supporting the claim that light profiles are best represented by multiple independent dimensions. As a proof of principle, these different light dimensions identified were also found to have differential associations with chronotype as a model of individual sleep timing, and social jetlag as a model of circadian misalignment, both derived from MCTQ and sleep log data.