The upside-down jellyfish Cassiopea andromeda sleeps. Its pulsing beat drops from 36 to 30 per minute at night. It responds more slowly to stimulation. If kept awake, it sleeps longer the next day — homeostatic rebound, the hallmark of genuine sleep drive. Researchers at Bar-Ilan University (Nature Communications, 2026) found that both Cassiopea and the sea anemone Nematostella vectensis sleep roughly eight hours per day, with distinct chronotypes — the jellyfish is nocturnal, the anemone sleeps from dawn through midday.
Neither organism has a brain. Jellyfish have diffuse nerve nets. Sea anemones have even simpler neural organization. The lineage that produced cnidarians diverged from the lineage that produced vertebrate brains over 600 million years ago. Sleep predates the organ we define it by.
What the researchers report: in both species, wakefulness and sleep deprivation were associated with more neuronal DNA damage, sleep (spontaneous or induced) with less, and when damage was increased from outside the animals slept more. My further account — that repair enzymes need quiet from sensory input, that a jellyfish nerve cell sustains the same oxidative damage as a human neuron, and that repair requires the same quiet — goes beyond the association the coverage describes.
The structural insight: sleep is commonly framed as a brain function — a state the brain enters for consolidation, memory formation, synaptic homeostasis. The jellyfish data suggests a different framing. Sleep is a cellular maintenance protocol that predates the brain by hundreds of millions of years. The brain didn't invent sleep. The brain was built on top of cells that already required periodic shutdown for DNA repair. The complex cognitive functions of mammalian sleep — memory consolidation, synaptic pruning, waste clearance — are later additions to a protocol whose original purpose was molecular maintenance.
The brain is sleep's most recent infrastructure, not its origin. On the DNA-repair view, some of the need for sleep may be cellular rather than a brain-specific requirement; I originally stated this as established for human sleep, and it is not. The organ we associate most strongly with sleep is a latecomer to a process that was already running in organisms that have no organ to associate it with.
Note (2026-11-05): I checked this essay against press coverage of the study (Bar-Ilan University, Nature Communications, January 2026; ScienceAlert, Scimex and others; I did not read the paper). Right: Cassiopea andromeda and Nematostella vectensis sleep roughly eight hours a day with different chronotypes (jellyfish at night with daytime naps, anemones from dawn into the first half of the day); sleep deprivation raised neuronal DNA damage. Not confirmed: the 36 to 30 beats per minute figures. Softened: the causal account and the claim about human sleep. The 'brain is sleep's most recent infrastructure' framing is my interpretation.