Transfer of Anolis locomotor behavior across environments and species

Arboreal animals must learn to modulate their movements to overcome the challenges posed by the complexity of their heterogeneous environment, reduce performance failure, and survive. Anolis lizards are remarkable in the apparent ease with which they conquer this heterogeneity, demonstrating an impressive ability to modulate their locomotor behavior to maintain stable locomotion on widely disparate surfaces. Significant progress has been made towards understanding the impact of substrate structure on the behavioral plasticity of arboreal species, but it is unclear whether the same strategies employed to shift between substrates in one context can be employed to shift between those same substrates in a new context. Is the kinematic shift between broad and narrow perches achieved in a similar way on different inclines? Do all species within an ecomorph make similar adjustments? Here, we analyze the limb movements of two trunk-crown Anolis ecomorphs, A. carolinensis and A. evermanni, running on 6 different surfaces (3 inclinations × 2 perch diameters), from the perspective of Transfer Learning. Transfer learning is that field of machine learning which aims at exploiting the knowledge gained from one task to improve generalization about another, related task. In our setting, we use transfer learning to show that the strategies employed to improve locomotor stability on narrow perches are transferred across environments with different inclines. Further, behaviors used on vertical inclines are shared, and thus transfer well, across perch diameters whereas the relationship between horizontal and intermediate inclines change on different perch diameters, leading to lower transfer learning of shallow inclines across perch diameters. Interestingly, the best incline for transfer of behavior differs between limbs: forelimb models learn best from the vertical incline and hind limb models learn best from horizontal and intermediate inclines. Finally, our results suggest both that subtle differences exist in how A. carolinensis and A. evermanni adjust their behaviors in typical trunk-crown environments and that they may have converged on similar strategies for modulating forelimb behavior on vertical surfaces and hind limb behavior on shallow surfaces. The transfer of behavior is analogous to phenotypic plasticity, which likely plays a key role in the rapid adaptive evolution characteristic of Anolis lizards. This work is an example of how modern statistical methodology can provide an interesting perspective on new biological questions, such as on the role and nuances of behavioral plasticity and the key behaviors that help shape the versatility and rapid evolution of Anolis lizards.