I'm particularly interested in understanding how animals assess and make decisions across contexts. I have studied this by focusing on assessment during contests and mate choice, and by integrating across behavior, biomechanics and physiology.
One major goal of my research is to understand how animals assess ability and safely resolve contests. During my Ph.D., studied this in the mantis shrimp Neogonodactylus bredini, which compete over access to territories in coral rubble. During contests, N. bredini exchange visual displays and high-force strikes with spring-powered raptorial appendages. These strikes are some of the fastest movements in the animal kingdom, able to crack open snails and dismember crabs. How do competitors use these weapons to resolve contests safely?
I found that N. bredini resolve body length-matched contests by ritualistically exchanging strikes on each other's tailplates (telsons). Contest winners did not have greater peak strike force than losers; instead, winners struck a greater number of times. This study showed how mantis shrimp use deadly weapons to resolve contests while minimizing injury. This work was published in 2015 in Biology Letters. Find out more on the publications page.
In addition to understanding how contests are resolved, behaviorists also want to know how contests proceed--how do competitors make the decision to continue or give up a fight? Theoretical models of contest resolution describe the rules animals follow during contests, but the predictions of these models are often not met by experimental studies. To resolve this mismatch between theory and experiment, I developed new methods to match N. bredini contest dynamics to model predictions. I tested these models at both the level of the contest (how long do contests last?) using correlation analysis, and at the level of contest behaviors (how do behaviors themselves progress through contests?) by developing a network analysis-based sequential analysis technique. I found that mantis shrimp assess their own ability as well as that of their competitor - a similar "mutual assessment" strategy as red deer. This work was published in 2018 in Proceedings of the Royal Society B. Find a copy of the paper on the publications page. The network analysis techniques I developed for this study were also used to study mutualistic interactions between cleaner shrimp and client fish (Caves, Green & Johnsen, 2018, Proc. B., see the publications page.)
Contests among social-living groups
As a Postdoctoral Fellow with the Human Frontiers Science Program, I'm currently working to extend this work to understand intergroup contests--contests between groups of social-living animals.
Social groups like colonies of ants, packs of mongooses, and even human soceities regularly compete over access to territory and resources.
How do groups assess their own fighting ability, and that of their competitors, to decide who should win a contest?
I'm working with Michael Cant at the University of Exeter Centre for Ecology and Conservation to test these questions using the banded mongoose system.
Banded mongoose groups have some of the highest rates of intergroup conflict of any mammal. Prof. Cant has studied a population of mongooses in Uganda nearly 25 years,
collecting data on over 600 of these contests. I am analyzing these data to determine how factors like group size, composition, and reproductive status predict
a group's chances of winning a contest, and how contests escalate (in duration and chance of injury or death).
Though these analyses are common to studies of dyadic contests, they have been greatly understudied in intergroup contests and can influence our understanding of
animal contests and social evolution.
Energetics and kinematics across contexts
A major goal in organismal biology is understanding how animals power myriad tasks using limited energy stores. I investigated this by testing how mantis shrimp vary strike kinematics and energetics across behavioral contexts.
I filmed mantis shrimp striking competitors and prey items at 40,000 frames/second, then used a MatLab-based mathematical model to analyze strike kinematics and energy use. I found that strike velocity and energetics scaled differently according to behavioral context: energy scaled positively with size during sparring, but did not vary with size during feeding. Velocity decreased with size while feeding, but did not vary with size while fighting. Mantis shrimp likely achieved energy variation through differentially compressing their exoskeletal spring. Finally, within a contest, individuals varied the energy and velocity of their strikes according to the relative size of their competitor. This work (published in 2019 in JEB) shows how and why animals power movements differentially. Learn more by reading the paper on the publications page.
Perception and assessment
Classic models of sexual selection assume that receivers of signals perceive and respond continuously to continuous variation in signal form. However, this may not be the case. In categorical perception, continuous variation is perceived as lying in in discrete categories, and stimuli compared across a category boundary are more easily discriminated than those within a category. Categorical perception of signaling traits may fundamentally alter models of sexual selection. With the Nowicki Lab at Duke, I've been testing hypotheses related to categorical perception. In zebra finches, we've found that females show categorical perception of carotenoid-based coloration used in mate choice. These results may necessitate a shift in our current understanding of sexual selection--see our recent publication in Nature (link on the publications page). We've also shown categorical perception extends to non-signaling stimuli (blue-green colors), and I've synthesized research in this field to understand the role categorical perception plays in animal communication and decision-making (see publications page). Currently, I'm testing how male zebra finches perceive carotenoid-based color and how this matters to the formation of dominance hierarchies.