Published in Computational and Systems Neuroscience (CoSyNe), Mar 2026.
Cuttlefish are masters of camouflage. To camouflage the cuttlefish brain needs to process the visual environment, compute an appropriate pattern, and then orchestrate the expansion and contraction of half a million pigment filled sacs, known as chromatophores, to generate it. We hypothesize that the animals operate by computing a set of statistics of their visual environment, and then adjusting their chromatophores to create skin patterns that match these statistics. Here, we develop a framework to test what statistics they are matching. First, we develop a behavioral paradigm that elicits camouflage behavior to a variety of natural substrates. Next, we build a model of the skin, a "renderer", which simulates the reflectances of over 125,000 chromatophores (~1/4 of the animal's skin). We test if a measured set of statistics are sufficient to mimic the animal's patterns by optimizing the radii of the simulated chromatophores to create an image whose statistics match that of a background image. We demonstrate that the chromatophore renderer can match statistics from a standard texture model based on the primate visual system (Portilla et. al. 2000), which produces simulated skin patterns that are well camouflaged. However, analyzing the behavioral data suggests that the animals are not matching the full set of these texture statistics. A simple center-surround model designed to mimic cephalopod early visual processing produces images that are somewhat camouflaged, but to a lesser extent than the observed skin patterns. However, images generated by directly optimizing pixel intensities to match the statistics produce worse camouflage than the renderer. Suggesting the natural constraints imposed by the physical chromatophore process biases the patterns towards effective camouflage. Overall, our framework allows testing candidate sets of statistics, and future work will combine this with additional behavioral data to determine a set of statistics sufficient to explain cuttlefish camouflage behavior.