RAD-Behavior (Recombining Atomized, Discretized, Behavior): A new framework for the quantitative analysis of behavioral execution

Abstract

The ability to precisely describe and numerically evaluate organismal phenotypes is a prerequisite for addressing most questions in evolutionary biology and ecology. The quantification and comparison of behavior, loosely defined as an external response to stimuli, is particularly challenging because the myriad axes of variation that exist make comparisons, both within and among species, difficult. Such evaluations often boil down to comparisons of time-budgets (e.g. relative investment in courtship displays) or probabilities (e.g. likelihood of engaging in a class of behaviors in a particular context) - which we refer to as behavioral strategies. A focus on variation in behavioral strategies underlies most research in evolutionary and ecological studies of behavior. Equally important, however, is perhaps the question of how animals are actually performing the complex motor sequences that comprise behaviors (i.e. behavioral execution). What are the patterns of movement, the relative transition rates, and kinematics underlying the behaviors exhibited in particular contexts? Understanding how behavioral execution differs among individuals, populations, and species has the potential to provide new insights into the factors shaping variation in behavior and the processes shaping behavioral evolution at different scales. Here, we propose a broad framework for comparing behavioral execution (RAD-behavior: recombining atomized, discretized behavior) that leverages string-matching/bioinformatic tools to understand phenotypic variation in behavioral execution and which holds the potential to yield novel insights about the evolutionary ecology of behavior at multiple scales.