Comprehensive analysis of behavioral dynamics in the protochordate Ciona intestinalis

Locomotion is broadly conserved in the animal kingdom, yet our understanding of how complex locomotor behaviors are generated and have evolved is relatively limited by the lack of an accurate description of their structural organization. Here we take a neuroethological approach to break down the motor behavioral repertoire of one of our nearest invertebrate relative, the protochordate Ciona intestinalis , into basic building blocks. Using machine vision, we track thousands of swimming larvae to obtain a feature-rich description of larval swimming and show that most of the postural variance can be captured by six basic shapes, which we term “Eigencionas”. Using multiple complementary approaches, we built representations of the larval behavioral dynamics and systematically reveal the global structure of behavior. By employing matrix profiling and subsequence time-series clustering, we reveal that Ciona swimming is rich in stereotyped behavioral motifs. Combining pharmacological inhibition of bioamine signaling with Hidden Markov Model we discover underlying behavioral states including multiple modes of roaming and dwelling. Finally, performing a spatio-temporal embedding of the postural features onto a behavioral space provides insight into the behavioral repertoire by project it to a low-dimensional space and highlights subtle light stimulus evoked behavioral differences. Taken together, Ciona larvae generate their spontaneous swimming and visuomotor behavioral repertoire by altering both their motor modules and transitions between, which are amenable to pharmacological perturbations, facilitating future functional and mechanistic investigations. ### Competing Interest Statement The authors have declared no competing interest.