California two-spot octopus in front of a mirror in the Octopus Lab at Dartmouth. (Credit: Mary Kieseler)
In A Nutshell
- Three California two-spot octopuses learned to use a mirror’s reflection to locate hidden food, a first for any invertebrate.
- All three chose the correct side about 73% of the time, well above random chance, across dozens of trials.
- In more than half of successful trials, octopuses climbed over a chamber wall to reach food they couldn’t yet see, suggesting they were using more than just a simple visual reflex.
- Researchers say the finding points to convergent evolution, the idea that octopuses and vertebrates independently arrived at similar cognitive abilities despite diverging over 520 million years ago.
For the first time, an animal without a backbone has used a mirror to find something it couldn’t directly see. The animal is an octopus, and what happened in a lab tank at Dartmouth College is making researchers rethink what kinds of brains can solve a tricky spatial problem.
Scientists trained three California two-spot octopuses to locate a hidden food reward using only a mirror’s reflection as a guide. Each octopus learned to read the reflection, figure out where the real target was located behind it and out of view, and navigate there deliberately. Using a mirror to locate hidden objects had previously been documented in vertebrates, including mammals and birds, but not in invertebrates. Octopuses don’t share that vertebrate blueprint. Their evolutionary lineage split from ours more than 520 million years ago.
Chimpanzees, elephants, pigs, dogs, and some birds have been tested on whether they can use mirrors to solve problems or locate hidden food. Because this ability had been shown mainly in vertebrates, researchers didn’t know whether an invertebrate brain could pull off the same kind of mirror-guided navigation. These three octopuses answered that question.
Octopuses Had to Ignore What They Could See to Find What They Couldn’t
Published in Current Biology, the experiment was designed to confirm the octopuses were using the mirror and not picking up on other cues. Each octopus was placed inside a small open-topped start chamber along the back wall of a large tank. A mirror spanning the tank’s width sat at the midline, facing the back wall, and the chamber walls blocked any direct view of what lay behind the octopus.
A projected image of a crab, one that octopuses respond to with real predatory behavior, was displayed on either the left or right side of the back wall. From inside the chamber, the only way to see it was to look in the mirror. To earn a live crab as a reward, the octopus had to exit the chamber and move to the correct side, matching where the virtual crab appeared in the reflection. No humans were present during trials; experimenters watched remotely through cameras in an adjacent room.
Octopuses Learned to Override Their Own Hunting Instinct
When an octopus spots something that looks like prey, its instinct is to go straight for it. During the first mirror-use learning trial, when a live crab in a jar was visible only in the mirror, all three animals moved toward the reflection as though the crab were actually there. But over roughly 10 to 12 learning trials per animal, each learned to override that impulse and move toward the actual location of the reward on the opposite wall.
Across all trials, the octopuses chose the correct side about 73% of the time, well above what random chance would produce. Multiple statistical tests confirmed the result was not a fluke. One octopus showed a slight preference for moving right regardless of where the crab appeared, but even accounting for that, all three animals performed better than chance. As trials progressed, they also got faster, a sign they were becoming more confident and efficient.
Climbing Over Walls to Reach Food They Couldn’t Even See Yet
Most telling was what the octopuses did with their bodies. In 59% of correct trials, instead of walking around inside the start chamber toward the back wall, they climbed up and over the side wall to reach the target.
That path took them away from the mirror reflection before they had any direct view of where the crab actually was. If the octopuses were simply reacting to what they saw, it makes no sense. But if they were using some kind of internal sense of the tank’s layout, as the authors suggest, climbing over the wall becomes a reasonable shortcut. The study cannot prove that explanation on its own, but the behavior points in that direction. Researchers describe this as evidence of “spatial mapping.” “These findings extend mirror-use capabilities to invertebrates,” the authors write, “demonstrating that cephalopods can employ mirror reflections for spatial navigation.”
What Octopus Intelligence Tells Us About How Smarts Evolve
Octopuses developed sophisticated brains and behaviors entirely independently from the line that produced mammals, birds, and humans. Yet in a tank in New Hampshire, three of them demonstrated a cognitive ability that looks remarkably similar to what researchers have documented in those other animals.
Scientists call this convergent evolution, when two unrelated groups independently arrive at the same solution to a shared problem. Here, that problem is navigating a world where not everything is visible at once. The ability to use mirrors for navigation, appearing in both octopuses and vertebrates despite separate evolutionary histories, points to something worth considering: this kind of spatial navigation may be a basic requirement for any animal that needs to hunt, hide, and survive in a world full of obstacles.
None of this means octopuses recognize themselves in mirrors. But the study raises the question of whether they may have other mirror-related abilities worth testing.
Paper Notes
Limitations
The authors are straightforward about several important limitations. The sample size was small, with only three octopuses completing the study. Others were excluded because they began showing signs of natural aging-related decline before trials could be completed, and resource constraints limited how many animals could be tested. The authors also acknowledge that the study was not designed to definitively distinguish between two competing explanations for the octopuses’ success: one being simple learned association (the reflection at position X means food at position Y), the other being genuine spatial reasoning using an internal map of the tank. Future studies using larger sample sizes, new mirror configurations, and tests of whether learning transfers to novel situations would be needed to resolve that question. Individual variability, including mood and hunger level, may have influenced results, and one octopus showed a significant rightward bias in its choices, though statistical analysis confirmed this bias alone did not explain its performance.
Funding and Disclosures
Funding for this research was provided by the National Science Foundation under grant numbers 2122962, 1632738, and 1844589. The authors report no competing interests. During the preparation of the manuscript, the authors used the AI tool Claude to improve readability and clarity of some text, and all content was reviewed and edited by the authors, who take full responsibility for the publication.
Publication Details
Paper Title: Octopus bimaculoides can learn to utilize a mirror to localize a reward outside the line of sight | Authors: Mary Kieseler, Marvin R. Maechler, Kelly R. Finn, Carl Harris, Jay Michael Vincelli, Zachary Hoffman, Navneet Dhanoa, Jean Fang, Scott Gies, James McHugh III, Julia Valenti, Mira Ram, John O. Fitzgerald, Madison Augusto, David Edelman, and Peter U. Tse, all affiliated with the Department of Psychological and Brain Science, Dartmouth College. Mary Kieseler is additionally affiliated with the Department of Psychology, University of Fribourg. David Edelman is additionally affiliated with the Association for Cephalopod Research (CephRes). | Journal: Current Biology, Volume 36, June 22, 2026 | DOI: https://doi.org/10.1016/j.cub.2026.05.012







