The stillness of sleep – Science Magazine

When animals fall asleep, skeletal muscle movement largely ceases. The lack of movement during sleep is an actively controlled process, just like sleep itself. There are specialized sleep-inducing neurons that mostly reside in the brainstem and hypothalamus (1). Until now, active repression of movement during sleep was thought to mainly apply to rapid eye movement (REM) sleep, which is when the neocortex exhibits a wake-like activity and dreaming is vivid. Conversely, for the first stage of sleep, non-REM (NREM) sleep, when activity of neurons in the neocortex synchronize at 0.5 to 4 Hz (called delta waves), it was unknown whether movement was actively repressed. On page 440 of this issue, Liu et al. (2) find that entering NREM sleep and stopping movement are wired together in mice. This is controlled by a brain region called the substantia nigra pars reticulata (SNr), which was thought to control motor actions only when mice are awake.

Liu et al. studied an inhibitory neuronal subtype in the SNr of mice, marked by the expression of the gene glutamic acid decarboxylase 2 (Gad2), which encodes a protein that synthesizes the inhibitory neurotransmitter molecule -aminobutyric acid (GABA). They discovered that these neurons send their axons to areas of the brain that simultaneously induce NREM sleep and inhibit movement (see the figure). For example, to inhibit movement, the Gad2+ SNr neurons connect to the motor thalamus and other motor areas of the brain. But to induce sleep, they also inhibit arousal-inducing centers such as the locus ceruleus and dorsal raphe. Because of these connections, a specific circuitry now explains how movement is repressed during NREM sleep, as well as during REM sleep.

The neural circuitry that suppresses movement during REM sleep seems entirely different from the circuit that suppresses movement in NREM sleep. During REM sleep, brain-stem circuits actively suppress motor neurons in the spinal cord, which control skeletal muscle contraction. This means that skeletal muscles have no tone (atonia) during REM sleep (3). The two different circuits that suppress movement in NREM and REM sleep further emphasize the differences between these types of sleep; indeed, sleep researchers are still puzzled as to why two states of sleep exist.

In mice, inhibitory Gad2+ neurons in the SNr are a central hub in the brain for triggering both immobility and NREM sleep. A parallel circuit, located in the brainstem, imposes muscle atonia during REM sleep.

However, falling asleep is not like throwing a switch. Animals are not just awake, and then suddenly asleep (unless suffering from narcolepsy). Indeed, preparing to sleep is a complex and slow behavior. Humans perform specific sleep-preparatory behaviors, such as putting on bed clothes, getting into bed, and adopting particular postures. After feeling drowsy, humans slip into NREM sleep. There seem to be intriguing hints that, at least in mice, specific circuitry controls sleep-preparatory behavior: Nesting behavior in mice prior to sleep is initiated by inhibiting dopamine neurons in the ventral tegmental area (VTA), which is a brain region closely related to the SNr that controls goal-directed behaviors (4). Once in the nest, sleep-preparatory postures promote skin warming which, via the preoptic hypothalamus, induces sleep (5). Liu et al. discovered that the Gad2+ SNr neurons might be contributing to the drowsiness factor of sleep-preparatory behavior. The authors used machine learning to make nuanced assessments of motor behavior and sleep entry based on electroencephalogram (EEG) readings from mice, which detect brain activity wave patterns. They correlated EEG activity of the Gad2+ SNr neurons and found that they function to bias mice to enter NREM sleep by both promoting delta waves (during drowsiness) and inhibiting movement.

The study of Liu et al. adds to recently identified long-range neuronal connections that induce NREM sleep from brain areas, collectively known as the basal ganglia, that are classically considered to promote aspects of wakefulness such as motivation, movement, and motor planning. Nearly all of these basal ganglia and closely associated regions also have neurons that strongly promote NREM sleep, including the nucleus accumbens and caudate putamen (6), globus pallidus externa (7), and now the SNr (2), as well as the VTA (8). On the basis of these collective findings, the distributed nuclei of the basal ganglia seem to play a major role in regulating NREM sleep (6). It seems to be no coincidence that deep brain stimulation of the basal ganglia nuclei to reduce tremor in Parkinson's disease patients often reduces the sleep disturbances these patients suffer and improves their sleep (7). The added and distinctive feature of the new findings is that SNr neurons actively innervate and inhibit motor-generating centers as well as promote sleep, whereas presumably other NREM sleeppromoting neurons inhibit movement indirectly.

Muscle activity is not completely suppressed during healthy sleep. Breathing continues, the heart beats, and in REM sleep, eye muscles are highly active, causing the eyes to flicker behind closed lids. But overall, one of the big mysteries of sleep is why it is necessary to go offline and be unconscious (9). Why has such a circuit evolved so that sleep and immobility are linked? It has been suggested that immobility serves the restorative function(s) of sleep (10). If the animal has to be offline and unconscious to carry out whatever restorative process is being served by sleep, then this may be the reason it needs to stay immobile. For example, it is widely supposed that paralysis during REM sleep stops dreams from being acted out (a process that fails during REM sleep behavioral disorder in humans). But there are also dreams during NREM sleep, so the Gad2+ SNr-imposed immobility discovered by Liu et al. could serve a parallel function.

What happens to the Gad2+ SNr neurons in animals (e.g., dolphins, and some migrating birds) that keep moving while they are partially asleep (11)? In these cases it might be expected that a modified circuitry exists, such that the sleep-promoting neurons in the SNr do not inhibit movement. Gad2+ SNr neurons could control two desirable features of anesthesia: immobility and unconsciousness (12). It would be interesting to know whether stimulating Gad2+ SNr neurons would aid entry into anesthesia, and if sedatives could selectively target this type of neuron (12).

Acknowledgments: The authors are supported by the Wellcome Trust and the UK Dementia Research Institute.

Follow this link:
The stillness of sleep - Science Magazine