Disentangling cephalopod chromatophores motor units with computer vision

Cephalopod chromatophores are skin pigment organs enabling rapid, neurally controlled camouflage, yet the organization of their motor control remains poorly understood. Previously, we developed CHROMAS, a computer-vision pipeline for high-resolution analysis of chromatophore dynamics (Ukrow et al., 2025). Here, we apply it to investigate motor control and innervation in Euprymna berryi and Sepia officinalis. By segmenting chromatophores into radial slices and analyzing anisotropic deformations, we used dimensionality reduction and source separation to estimate the number and spatial influence of motor neurons controlling individual chromatophores and groups thereof. On average, four independent components were detected per chromatophore, each forming contiguous petal-shaped domains. Clustering thousands of components revealed motor units spanning multiple chromatophores, most involving fewer than 14, with diverse geometries ranging from compact local groups to elongated or fragmented structures; chromatophore pairs were co-innervated more often than expected by chance. Expansion was consistently faster and more stereotyped than relaxation, consistent with active contraction and passive recoil. These results show that chromatophores are not uniform pixels but contrast elements fractionable into sub-territories coordinated across neighbors. This geometry of neural control enables the generation of 'virtual chromatophores', that is, functional groupings of adjacent chromatophore territories that act as single units, as well as that of noise in the distribution of pixel shapes.