Deciding under duress: motor decisions of larval zebrafish under attack


We perform many movements every day without much deliberation. However, moving can be seen as a form of decision-making since one of many possible movements must be selected and executed. The decision-making processes that underlie movements are influenced by various factors, including sensory perception, energetics, time, perceived rates of failure along with uncertainty in the accuracy of sensory information and the movements themselves. The computations needed to evaluate these factors and make motor decisions are known to occur within the nervous system but the processes of movement selection and execution are still not well understood. Simpler organisms, such as fish at early developmental stages, are important model systems to investigate the neurobiology of motor decisions. Specifically, the escape responses of larval fish can be mapped to a finite number of neurons that are identifiable under microscopy, allowing for an analysis of the neural substrate underlying the behavior. Moreover, selecting and executing the appropriate escape movement is critical for survival. Therefore, escape behaviors are simple models of decision-making where the biomechanics of locomotion and the associated neural circuitry are under severe evolutionary pressure. This thesis investigates both the kinematics and the neurobiology of the movements made by larval zebrafish during escape behaviors initiated in response to real and simulated predators. Through this work, I aim to answer four main questions about the evaluations and motor decisions relevant to the escape maneuvers of larval zebrafish: - Do larval zebrafish perform a graded assessment of the threat posed by a predatory attack? - How is the assessed threat used to make motor decisions about deploying specific escape movements or strategies as a response? - What are the utilities of specific escape movements and their associated neural activations in producing maneuvers that successfully evade predators? - Since escape behaviors must be optimally timed while being resistant to false negatives and false positives, how are the recruitment of escape circuits influenced by noise or uncertainty in sensory information? By showing larval zebrafish visual stimuli of virtual predators approaching at different speeds, I interrogated how larval fish evaluate threat. Through these experiments, I found that the fish performed a graded assessment of threat and responded by deploying escape maneuvers with characteristic patterns of neural activation. To understand the utility of specific escape maneuvers in producing successful evasive responses, I studied the escape responses of larval zebrafish from the attacks of a natural predator, the dragonfly nymph. By combining extensive analysis of the predator-prey interaction with computational methods, I found that the intersection of the reachable spaces of predator and prey plays a crucial role in determining escape success. Lastly, I studied how noise in the visual percept of an approaching predator influenced the recruitment of escape circuits and the resulting escape behavior. Through a computational framework, I demonstrate that the sources of sensory noise and the methods of compensating for them may interact to serve functional roles in producing the resulting behavior. Overall, these results draw a continuous thread through the sub-processes relevant to making escape decisions. The findings provide insight into sensory reception, sensory evaluation, neural recruitment, and movement generation in larval zebrafish under attack. While the four main questions presented above are an atomistic deconstruction of escape behavior, the investigation traverse scales of analysis---from the neural, to the organismal and the ecological.

Alternate Identifier
Date created
Resource type
Rights statement