Caribbean box jellyfish have no centralized brain, yet a 2023 study published in Current Biology found they can learn from experience just like humans, mice, and flies do—a discovery that upends longstanding assumptions about what kind of nervous system is required for genuine learning. The finding suggests that learning may be a far more fundamental property of nervous systems than scientists previously believed, present in some form since the very earliest stages of neural evolution.
A Tiny Animal With a Surprisingly Complex Visual System
Tripedalia cystophora, the species at the center of this research, is no bigger than a fingernail and lives in the murky, root-tangled waters of Caribbean mangrove swamps, where it hunts tiny water fleas. Despite its simplicity, this jellyfish has a genuinely sophisticated visual system, with 24 eyes embedded across its bell-shaped body, organized into four clusters called rhopalia. Each rhopalium houses six eyes and roughly 1,000 neurons, functioning as an individual visual processing and integration center. This distributed setup replaces the single centralized brain that most learning animals rely on—instead, the jellyfish’s nervous system is spread across these four independent hubs.
Training Jellyfish to Dodge Obstacles
To test whether these animals could actually learn from experience, researchers led by Jan Bielecki at Kiel University and Anders Garm at the University of Copenhagen placed the jellyfish in cylindrical tanks lined with gray stripes meant to mimic the mangrove roots they navigate in the wild. At first, when faced with low-contrast gray stripes that resembled distant roots, the jellyfish swam directly along the tank wall, bumping into it repeatedly. But over the course of just 7.5 minutes, they learned to change course—increasing their distance from the wall by 50% and cutting the number of collisions in half.
The researchers were struck by how quickly this learning happened. Just five brief training sessions were enough for the jellyfish to demonstrate clear behavioral change, a surprisingly fast pace of learning for an animal without any centralized neural processing center.
Combining Visual and Physical Cues
The key to how this learning worked came down to combining two different types of sensory input. The gray stripes served as a visual stimulus, while the physical sensation of bumping into the tank wall served as a mechanical stimulus, and the researchers found that both kinds of input were necessary for the learning to occur. When jellyfish faced high-contrast black stripes, which were easy to see and avoid from a distance, they experienced few collisions and showed no meaningful change in behavior. But when faced with a completely uniform gray wall offering no visual cue at all, they kept bumping into it and never learned to avoid it. Learning only happened when both the visual stripe pattern and the physical bump were present together, allowing the jellyfish to form a genuine association between the two.
Isolating the Learning Center
To pinpoint exactly where this learning was actually happening, researchers took the experiment a step further by removing the rhopalia—the eye-bearing nerve clusters—from the jellyfish’s body entirely and testing them in isolation. They placed these detached structures in front of a screen showing low-contrast gray bars while delivering a weak electrical pulse timed to mimic the sensation of a physical bump. At first, the isolated rhopalia ignored the gray bars, treating them as distant and irrelevant. But after receiving the paired electrical “bump” signal, the rhopalia began generating the same kind of nerve signals a jellyfish produces when it darts away from an obstacle—demonstrating that these individual sensory clusters, on their own, without input from the rest of the animal’s body, were capable of learning the association.
Why This Discovery Matters
This research represents the first convincing demonstration of associative learning—specifically a form called operant conditioning—in the entire phylum Cnidaria, which includes jellyfish, corals, sea anemones, and hydras. This form of learning had previously been assumed to require an advanced, centralized nervous system, and the box jellyfish’s success challenges that assumption directly, since its nervous system is dispersed across four separate rhopalia rather than unified in a single brain.
Because Cnidaria represents one of the earliest-branching animal lineages, sitting as a sister group to the vast category of animals with bilateral body symmetry, the discovery raises an intriguing evolutionary possibility. If a jellyfish with roughly 1,000 neurons per sensory cluster can learn this way, it suggests that learning may have been an integrated part of neurons from the very beginning of nervous system evolution, rather than a capability that only emerged later alongside more complex, centralized brains.
What’s Next for This Research
The research team plans to dig deeper into the cellular mechanisms underlying this learning, aiming to understand exactly how and where these simple neural clusters store the memories that guide behavioral change. Beyond satisfying basic scientific curiosity about animal cognition, a better understanding of how memory works at this most fundamental cellular level could eventually inform research into memory-related disorders like dementia in more complex organisms, including humans.
Join The Discussion
Does it surprise you that an animal without a brain can learn from experience, or does it change how you think about intelligence and cognition across the animal kingdom? What do you think this discovery suggests about how learning first evolved in early nervous systems? Share your thoughts, questions, or anything else you find fascinating about this research below.
