A comparative test of the links between sleep state, brain size, and incubation investment in birds

There are two electrophysiological states of sleep in birds (rapid-eye-movement sleep [REM] and slow-wave sleep [SWS]), which have different functions and costs. REM improves memory consolidation, while SWS is neuro-restorative but also exposes the animal to more risk during this deep-sleep phase. Birds who sleep in more exposed microsites are known to invest proportionally less in SWS (presumably to ensure proper vigilance), but otherwise little else is known about the ecological or behavioral predictors of how much time birds devote to REM v. SWS sleep. In this comparative analysis, we examine how proportional time spent in SWS v. REM is related to brain mass and duration of the incubation period in adults. Brain mass and incubation period were chosen as predictors of sleep state investment because brain mass is positively correlated with body size (and may show a relationship between physical development and sleep) and incubation period can be a link used to show similarities and differences between birds and mammals (using mammalian gestation period). We hypothesized that (1) species with larger brains (relative to body size and also while controlling for phylogeny) would have higher demands for information processing, and possibly proportionally outweigh neuro-repair, and thus devote more time to REM and that (2) species with longer incubation periods would have proportionally more REM due to the extended time required for overnight predator vigilance (and not falling into deep sleep) while on the nest. We found, using neurophysiological data from literature on 27 bird species, that adults from species with longer incubation periods spent proportionally more time in REM sleep, but that relative brain size was not significantly associated with relative time spent in REM or SWS. We therefore provide evidence that mammalian and avian REM in response to incubation/gestation period have convergently evolved. Our results suggest that overnight environmental conditions (e.g. sleep site exposure) might have a greater effect on sleep parameters than gross morphological attributes.