Applying movement ecology models to comparative cognition experiments: a field test in hummingbirds.

Traditionally, evidence for spatial learning in animals involves experiments; however, tools developed for analysing tracking data are increasingly being used to infer the use of spatial memory and other cognitive processes in animal movements. In this study, we combined these statistical models with field experiments to analyse how patterns of hummingbird movements change as birds learn a rewarded location. Using hidden Markov models (HMMs), we used both changes in two movement states as well as experimental behavioural measures of spatial memory to determine how searching performance and behaviour changed as birds gained experience and how important local landmarks were for guiding search. Regardless of whether birds had a single training trial to learn a flower's location or 12 repeated trials, hummingbirds learned the location of the flower although performance suffered when we removed the local landmarks. While the hovering locations suggested that removing landmarks led to a slight decrease in accuracy compared with when landmarks were present, the HMMs suggest that this was part of a larger shift from a memory-led search strategy to a more systematic searching process. Our results suggests that models used in movement ecology could provide a valuable tool for experiments in comparative cognition.