Person:

Dunn, Timothy

In the absence of salient sensory cues to guide behavior, animals must still execute sequences of motor actions in order to forage and explore. How such successive motor actions are coordinated to form global locomotion trajectories is unknown. We mapped the structure of larval zebrafish swim trajectories in homogeneous environments and found that trajectories were characterized by alternating sequences of repeated turns to the left and to the right. Using whole-brain light-sheet imaging, we identified activity relating to the behavior in specific neural populations that we termed the anterior rhombencephalic turning region (ARTR). ARTR perturbations biased swim direction and reduced the dependence of turn direction on turn history, indicating that the ARTR is part of a network generating the temporal correlations in turn direction. We also find suggestive evidence for ARTR mutual inhibition and ARTR projections to premotor neurons. Finally, simulations suggest the observed turn sequences may underlie efficient exploration of local environments. DOI: http://dx.doi.org/10.7554/eLife.12741.001

Comprehensive descriptions of animal behavior require precise measurements of 3D whole-body movements. Although 2D approaches can track visible landmarks in restrictive environments, performance drops significantly in freely moving animals, where occlusions and appearance changes are ubiquitous. To enable robust 3D tracking, we designed DANNCE, a method using projective geometry to construct inputs to a convolutional neural network that leverages learned 3D geometric reasoning to track anatomical landmarks across species and behaviors. We trained and benchmarked DANNCE using a new 7-million frame dataset relating color videos and rodent 3D poses. In rats and mice, DANNCE robustly tracked dozens of landmarks on the head, trunk, and limbs of freely moving animals in naturalistic settings, achieving over an order of magnitude better accuracy than prior techniques. We extend DANNCE to rat pups, marmosets, and chickadees, and demonstrate a novel ability to quantitatively profile behavioral lineage over development. DANNCE offers unprecedented analytical access to animal behavior across species and environments.

Behavior is generated via brain-wide coordination of neural circuits. But until recently, it was difficult to analyze neural dynamics at cellular resolution throughout the brain during behavior. With the genetic and optical accessibility of the larval zebrafish, however, we are now beginning to dissect neural circuits on larger scales. Here, I describe the neural origins of two prominent innate behaviors of the larval zebrafish: (1) looming-evoked escape behavior and (2) a self-generated exploratory behavior. In zebrafish, punctuated mechanosensory stimuli, signaling proximal threats, elicit escape behaviors that rely on a compact neural circuit. Visual identification of threats, however, is more complex: instead of detecting an impulse-like stimulus, danger must be recognized by computations on the spatiotemporal properties of visual scenes. Here, I characterize behavioral responses to visual stimuli simulating predator approach using a high-speed, closed-loop system that enables precise control over the visual environment of free-swimming fish. I report that the visual system alone recruits lateralized, rapid escape maneuvers in response to looming but not static stimuli. Brain-wide calcium imaging isolated the optic tectum as an important visual center processing looming stimuli, with ensemble activity encoding escape latency. Finally, ablations of hindbrain circuitry confirmed that visual and mechanosensory modalities share a premotor output network. In the absence of specific stimuli, however, animals continue to exhibit rich self-generated behavior. In featureless environments, fish exhibit stereotypical behavioral sequences, which consist of repeated turns in one direction followed by stochastic switches to repeated turns in the other direction. Using whole-brain imaging in behaving animals, we found antisymmetric activity in distinct hindbrain populations that was ipsilaterally correlated with turning and exhibited slow time courses well-matched to behavioral sequences. These populations correspond to the “hindbrain oscillator” (HBO), and cell ablations demonstrated a causal role for the HBO in determining the statistics of spontaneous swimming. We revealed that the HBO comprises separate glutamatergic and GABAergic clusters interacting across the midline, suggesting a mutual-inhibitory circuit motif shaping HBO dynamics and downstream behavior. These findings establish a circuit underlying spatiotemporally structured spontaneous behavior that, in simulations, supports efficient exploration of environments in the absence of explicit sensory cues.