Sleep deprivation is one of the most common forms of mental impairment, as it reduces performance in a variety of cognitive and motor tasks. A new paper has identified what might be the cause of this poor
performance: small parts of a mammal’s brain appear to go into a sleep-like state while they’re otherwise apparently awake. And, in rats, the degree to which the brain is experiencing these tiny episodes of sleep correlates with their decline in performance on a simple task.
The study relies on differences in the resolution at which we measure neural activity. On a crude level, rats and humans show a similar pattern of neural activity when non-REM sleep is recorded with an EEG: rhythmic peaks of activity, in between which the brain goes largely silent, with the neurons effectively “offline,” as the authors call it.
To get down to the level of individual cells, you have to implant electrodes in the brain. When this is done in rats, it’s possible to see that the “offline” periods are the product of nearly every neuron in a given brain region going silent, while the peaks of activity involve the frequent firing of most neurons. So, sleep seems to be the product of the coordinated action of most of the brain’s neurons, which drop offline and return to activity on a regular cycle.
For the new work, the researchers had both an external EEG reading and electrodes implanted in the frontal motor cortex. It turns out it’s remarkably easy to cause rats to deprive themselves of sleep: simply keep giving them new toys for several hours. Although the rats were, by all appearances, wide awake, their performance on a simple motor task (reaching for a sugar pellet) declined with the sleep deprivation.
The readings of the brains, however, often looked like the rats were experiencing what the authors termed “local sleep.” After four hours of deprivation, the EEGs would often show the spikes associated with the rhythmic activity of sleep, and most of the neurons near the implanted electrodes would go silent. This wasn’t a global phenomenon, however, as other areas of the brain didn’t show the same pattern of activity. And, in some cases, the electrodes recorded standard waking activity from many of the neurons they were near. Thus, instead of being part of a global, coordinated sleep pattern, small clusters of neurons were simply taking themselves offline while the animal itself was awake. As the sleep deprivation continued, these local sleep events became more frequent and coordinated.
The authors also obtained an indication that these local sleep events might be related to the poor mental performance seen during sleep deprivation. They looked at the pattern of brain activity when the animals were performing the motor task, and found that they were more likely to fail if a bit of local sleep occurred just before the rats attempted the task.
The paper concludes that local sleep is a precursor to a more general sleep, and resembles the converse of the waking process. Its authors also make comparisons between this local sleep and the needs of birds and marine mammals, which can apparently take large regions of their brains offline instead of sleeping to meet the needs of their lifestyle. That adaptation, they suggest, may be an elaboration of the local sleep that rats appear to use.
An accompanying perspective makes a separate suggestion: the fundamental unit of sleep may be a single cell. When an animal is sleep deprived, individual cells are more likely to take themselves offline. By chance, that will eventually start producing the clusters that produce local sleep events. Over time, the frequency of local events goes up, and the neurons begin to coordinate their activity, ultimately producing the large-scale rhythms seen in sleep.