A mouse curls into a ball and drifts off to sleep. Inside its brainstem — the stalk-like structure connecting the brain to the spinal cord — hundreds of neurons begin to flicker in unison, building slowly toward something remarkable: the dream state. Now, for the first time, researchers have caught this quiet electrical orchestra in the act, and what they've found could reshape our understanding of why we dream.
A team at the University of Pennsylvania and the Champalimaud Foundation discovered that coordinated slow waves of activity in brainstem neurons precede the transition from non-REM sleep to REM sleep in mice. Their findings, published in Nature Neuroscience, suggest the brain doesn't stumble into dreaming randomly — it prepares for it. Senior author Franz Weber has spent years wondering how the brain decides the time is right. "REM sleep is a very distinct brain state, but it does not occur completely randomly," he told Medical Xpress. "It is typically preceded by a period of non-REM sleep, and somehow the brain must determine when the time is right to transition into REM sleep."
To watch this decision unfold in real time, Weber and his colleagues used high-density Neuropixels probes — tiny devices developed collaboratively by researchers at the Howard Hughes Medical Institute, University College London, the Allen Institute for Brain Science, and other institutions. These probes allowed the team to simultaneously record the activity of roughly 185 neurons in the brainstem while mice slept. The recordings revealed two broad patterns of activity during sleep, but one stood out: during non-REM sleep, waves of population-wide activity gradually accumulated, like water rising behind a dam. When these slow fluctuations reached a certain threshold, the mice entered REM sleep.
"Notably, these dynamics could predict when the brain was more likely to enter REM sleep," Weber said. The researchers then combined their recordings with optogenetic experiments, stimulating specific REM-promoting neurons in the brainstem and cortex to test whether they could nudge the system. The results affirmed what the slow waves had suggested: the transition to dreaming is not a flip of a switch but a tide that builds and recedes.
The implications stretch beyond mice. Understanding how the brainstem orchestrates sleep transitions could eventually inform treatments for sleep disorders, which affect tens of millions of people worldwide. If researchers can learn to read and influence these preparatory slow waves, they may one day help those whose brains struggle to reach or maintain restorative REM sleep. For now, the discovery offers something more fundamental: a glimpse into the elegant choreography that carries us from wakefulness into the vivid worlds we inhabit each night.