The dolphin sleeps with one half of its brain at a time, while the other half stays awake. This was shown by the Russian scientist Lev Mukhametov, who recorded electrical brain waves from these maritime mammals. He found that the large slow waves that indicate deep sleep were confined to one brain hemisphere, but after a while switched to the opposite hemisphere.

Wouldn’t it be marvelous if we could assume a dolphin type of sleep with one half of our brain permanently awake? It would be a bonanza for all those who consider sleep to be an unnecessary loss of active lifetime. Although we humans lack this capacity, the dolphin phenomenon inspired us to ask whether our two brain hemispheres participate equally in the sleep process.

As measured by brain wave activity in the electroencephalogram (EEG), two distinct states of human sleep alternate in a cyclic fashion: the state of REM (characterized by rapid eye movements), and the non-REM, deep sleep state. A striking asymmetry in certain frequency ranges was found to alternate during the two sub-states of sleep — dominating in one hemisphere in REM sleep, and in the other hemisphere in non-REM sleep.

Of course, beyond the distinction between left and right brain, we can also distinguish between front (anterior) and rear (posterior). How might these areas show contrasting functions when we sleep? This question is interesting because different parts of the cortex clearly serve different functions while we are awake. For example, the frontal cortex (which sits under the forehead) is involved in high-level cognitive functions such as reasoning and problem solving. By contrast, the occipital cortex in the back of the head — the destination of a long pathway from the eyes — is engaged in routine processing of visual information. Indeed, we also find anterior-posterior regional distinctions during sleep.

Slow wave activity characterizes deep, intense non-REM sleep, which may reflect the need for recovery from wakefulness. This EEG pattern dominates during the first part of sleep, and gradually declines later in the night. While the decline is observable in both anterior and posterior areas of the cortex, it is steepest at the front — the same area that serves intense cognitive processing during the day. By contrast, the rear, posterior area deals with more mundane daytime activity like visual processing, and shows a less dramatic decline in slow wave activity during sleep.

It would be much more convincing to demonstrate that recovery during sleep is linked to the waking activity of a given brain area. Indeed, brain imaging had previously showed that sensory stimulation of the hand activates a specific sensory region of the cortex opposite to the hand used. In a sleep experiment to test this recovery hypothesis, right-handed university students served as research subjects. A vibrating stimulus was applied to the right hand for six hours before bedtime. EEG slow waves during sleep were more prominent in the left hemisphere. The effect was restricted to the area of the cortex that responds to touch, and it was limited to the first hour of sleep. Thus, we have proof of a highly specific sleep response to localized stimulation of the skin while awake. We can call this a local, use-dependent facet of sleep, with right hand / left cortex asymmetry.

Most often research progresses from laboratory animal models to attempts at replication in humans. This time it was different. With knowledge of the use-dependent effect in the student subjects, rats were tested with whisker stimulation on one side of the snout before they were allowed to go to sleep. Again, slow wave sleep measured at the sensory cortex on the opposite side showed heightened slow-wave EEG activity. It was very encouraging to find that use-dependent sleep is present also in animals.

Research on local, regional sleep has continued to gain traction since these discoveries, now with thousands of citations in the scientific literature. It looks like we have come closer to resolving the age-old conundrum, “Why do we sleep?”


Alexander Borbély MD, emeritus Professor at the University of Zurich, Switzerland, developed the two-process model of sleep regulation assuming an interaction of a homeostatic and circadian process. The model has provided a widely used conceptual framework for sleep research.