Temporally resolved analyses of aperiodic features track neural dynamics during sleep

Sleep has long been classified into stages by looking at well-known brain rhythms, such as slow waves and sleep spindles. Although these rhythms provide important information, recent research shows that they do not capture everything happening in the sleeping brain. A growing area of study focuses on the aperiodic or non-rhythmic part of the brain’s electrical activity.
This part does not contain clear rhythms, but it still changes in meaningful ways that reflect the brain’s state.

One way to describe this background activity is by measuring the slope of the brain signal across different frequencies. This slope is thought to reflect the balance between excitatory and inhibitory neural processes and tends to change from one sleep stage to another. Another feature, called the “knee,” marks a point where the slope shifts. The frequency at this knee may
reveal how quickly groups of brain cells process information, “processing timescales”. Earlier studies hinted that both the slope and the knee might help describe sleep more precisely.

To explore this further, the authors analysed a wide range of brain-signal frequencies so they could reliably detect the knee. They also used methods that allowed them to track changes in the slope and the knee continuously throughout the night, rather than averaging brain activity over longer periods of entire sleep stages. These methods were applied to two types of recordings: extracranial and intracranial EEG. This made it possible to see how aperiodic features changed during sleep-stage transitions and how they reacted to sounds presented during sleep.

The results suggest that the knee frequency provides a sensitive marker of state-dependent shifts in neural processing timescales across sleep stages. Moreover, the time-resolved nature of the analysis reveals that sleep is not a static sequence of stages but a continuously evolving landscape where aperiodic metrics track rapid transitions and even transient responses to external stimuli.

Together, these findings open new avenues for understanding how the brain reorganises itself during sleep. They also highlight the potential of aperiodic features to complement classical oscillation-based sleep metrics, offering a richer and more mechanistic view of neural activity across the sleep-wake cycle.