Jerome Siegel: UCLA Dept. of Psychiatry

Sleep amounts vary by more than an order of magnitude across mammalian species. Either the amount of time spent sleeping has no relation to underlying function, which would distinguish sleep from many other homeostatically regulated processes, or sleep need varies considerably across species.

Prior data and new data on primitive mammals and cetaceans indicate a strong negative correlation between total sleep time and weight. Because metabolic rate is strongly and negatively correlated with body mass, this is also a positive correlation between metabolic rate and sleep time. Some evidence suggests that brain regions with high metabolic rate have higher levels of sleep deprivation induced damage. We hypothesized that non REM sleep serves to repair damage caused by oxidative stress (Eiland et al., 2002; Ramanathan et al., 2002).

REM sleep and nonREM sleep amounts are positively correlated. One explanation for this is that REM serves to stimulate the brain to prepare for waking after a period of nonREM (Ephron, Carrington, 1966; Snyder, 1966; Vertes, 1986).

However, much of the variation in REM amounts is independent of nonREM duration. Animals born in a relatively immature state, have more REM early in development (Jouvet-Mounier, 1970). One may hypothesize that REM facilitates development. A major mystery is why immaturity at birth is correlated with REM time in adulthood.

Cetaceans show unihemispheric sleep, with both hemispheres never being in deep sleep at the same time. Fur seals show both unihemispheric sleep and bihemispheric sleep and can switch between these two modes. Unihemispheric sleep appears to largely do away with sleep rebounds after deprivation of bilateral sleep. Unihemispheric sleep is linked to low or absent REM sleep. Understanding the mechanisms and functional relations underlying these unusual sleep adaptations of marine mammals can offer a major insight into the function and mechanisms of sleep.