Behaviour 2019
Using a Homeograph to Detect Sleep
Matt Gaidica1, Emily Studd2, William Gonzalez1, Jeffrey Lane3, Andrew McAdam4, Stan Boutin2, Ben Dantzer1. 1University of Michigan, Ann Arbor, Michigan, United States; 2University of Alberta, Edmonton, Alberta, Canada; 3University of Saskatchewan, Saskatoon, Saskatchewan, Canada; 4University of Colorado, Boulder, Colorado, United States

Sleep is appreciated as a behavior critical to homeostasis, performance, and fitness. However, most of what we understand about sleep comes from humans or controlled laboratory experiments. Assessing sleep in the wild is challenging, as sleep is often hidden from view, and electrophysiological recordings are difficult to obtain. Accelerometers have offered great insight regarding gross movement, although ambiguous quiescent states like sleep have been largely ignored, limiting our understanding of gestalt behavior expressed by an organism. We developed a broadly applicable sleep detection algorithm that can be applied to accelerometer data collected from wild animals. We applied our methodology to detect sleep in free-ranging North American red squirrels (Tamiasciurus hudsonicus) in an area where it experiences drastic seasonal shifts in light, temperature, and behavioral demands. Our method expanded prior observations and reports to provide evidence that red squirrels apply unique sleep strategies to cope with changing environments. Applying a similar analytical strategy to accelerometer data from other species may open new possibilities for researchers studying wild animals.