Recently, ethological studies of animal behavior uncovered its complexity while
neuroscientific work began unraveling the neural bases of behavior. Improvements
in algorithmic understanding of behavior and neural function contributed to re-
cent breakthroughs in robotics and artificial intelligence systems. Yet, animals’
decision-making and motor-control are unequalled by human engineered systems
and the continued investigation of the behavioral and neural bases of these abilities
is crucial for understanding brain function and inform further technological devel-
opments. In my PhD work, I first investigate escape path selection in mice presented
with threat, demonstrating how mice combined rapidly acquired spatial knowledge
with an innate choice heuristic to inform decision-making. This strategy minimizes
the requirement for trial-and-error learning and yields accurate decision-making by
combining knowledge acquired at an evolutionarily time-scale with that acquired
by the individual. Future work aimed at understanding how these sources of in-
formation are combined in the brain to inform decision-making may lead to more
efficient artificial learning agents. Next, I studied goal-directed locomotion behav-
ior in which mice move rapidly through an environment to reach a goal location.
Successful goal-directed locomotion behavior requires substantial navigation and
motor control skills and, additionally, sophisticated planning and control of move-
ments while moving at high speed. Detailed behavioral quantification and compar-
ison to a control-theoretic model demonstrated that mice do possess such planning
skills, allowing them to execute rapid and efficient trajectories to a goal. Population-
level extracellular recordings of neural activity during goal directed locomotion was
also used to begin uncovering the neural bases of planning during locomotion. Altogether, my work combined accurate quantification of animal movements with the-
oretical models of optimal behavior to understand behavior at a computation level,
aiming to provide crucial information to inform future studies on the neural bases
of innate behaviors and aid in the development of novel artificial learning systems