What Is a Spindle? The Burst of Brain Activity That Protects Sleep

A sleep spindle is a burst of activity lasting roughly half a second to two seconds, in which oscillations in a band of around eleven to sixteen cycles per second grow and then fade, producing a shape on the trace that gives them their name. They are one of the defining features used to identify the second stage of non-rapid-eye-movement sleep, appearing alongside another characteristic waveform, and their presence is how a scorer distinguishes that stage from lighter drowsiness. They occur many times per hour, are generated by an interaction between a relay structure deep in the brain and the cortex, and vary between individuals in density and frequency to a degree that is stable across nights and substantially heritable, which makes them something like a signature.

Several roles are supported by evidence of different strengths:

• •Gating sensory input, with the mechanism generating them also blocking signals from reaching the cortex
• •Protecting sleep, since people with more of them are harder to wake with noise
• •Consolidating memory, with spindle activity after learning predicting what is retained
• •Coordinating with slow waves and with hippocampal activity in a timed sequence
• •Differing systematically in several clinical conditions, which makes them a candidate marker
• •Correlating with measures of reasoning ability, which is reported repeatedly and is not well understood

One line of evidence is unusually direct. Participants sleeping in a laboratory were exposed to recorded sounds at controlled volumes while their activity was monitored, and those producing more spindles tolerated more noise before waking. The finding supports the gating account, in which the burst reflects the brain actively blocking incoming signals at the relay stage rather than merely happening to occur while the sleeper is undisturbed. Further work found that the sounds themselves could trigger spindles, suggesting a response rather than a coincidence. The practical implication is that individual differences in tolerating a noisy sleeping environment have a measurable physiological basis rather than being a matter of habit or complaint, and attempts to increase spindle activity deliberately, whether by sound, by stimulation or by drugs, follow directly from that.

Spindles are one of several characteristic patterns and the others are worth placing alongside them. Slow oscillations of about one cycle per second dominate the deepest stage of non-rapid-eye-movement sleep, involve most of the cortex switching between active and silent states together, and are what makes that stage difficult to wake from. Sharp bursts from the hippocampus carry compressed replays of recent experience and are the clearest physiological evidence that the brain revisits the day during sleep. Another characteristic waveform appears in response to sudden noise and is thought to suppress arousal. Rapid-eye-movement sleep has a quite different profile resembling waking, with vivid dreaming and with the muscles paralysed. The proposed memory mechanism involves these rhythms occurring in a nested timed relationship rather than any one of them acting alone.

The link to memory rests on several kinds of study that support each other. Spindle density increases on nights following intensive learning compared with control nights. The amount of spindle activity after learning predicts how much is retained when tested later, across several types of material. Spindles cluster in time with slow oscillations and with bursts of hippocampal activity in a repeating sequence, which fits the proposal that information recorded rapidly during the day is transferred to longer-term cortical storage during sleep. Drugs that increase spindle activity have been reported to improve retention in some studies. The evidence is correlational in most designs and the causal claim rests on the smaller number of studies that manipulated spindles directly, which is the honest summary of where the field stands.

Short bursts of rhythmic activity around eleven to sixteen cycles per second define a stage of light sleep and recur many times an hour. People producing more of them tolerate more noise before waking, which supports the idea that the burst reflects active blocking of incoming signals. Spindle activity after learning predicts retention, though most of that evidence is correlational.