Dynamic walking behavior during odor trail-following in locusts

One of the important subsets of odor sources used in olfactory navigation is surface-bound sources, which can broadly take the form of point sources or trails. Odor trails, in particular, have been observed to be highly relevant components of olfactory-based navigation for species as broadly distributed as dogs and ants. Here, we present an automated treadmill setup capable of dynamically printing odor trails of arbitrary lengths and configurations, and with closed-loop control of speed based on the subjects movement. We used this setup to characterize trail-following behavior in locusts (Schistocerca americana). The free-moving behavior of the locusts is more naturalistic and is richer in plumbable data than many traditional assays. We reveal broad classes of behavioral walking motifs and their dynamic transitions as locusts pursue or avoid an odor trail. Furthermore, we show how these motifs vary across individuals, with the identity of the odorant and with respect to the sex of the organism. Our dataset and analyses provide a first demonstration that this model organism is capable of robust odor trail following, and provides a comprehensive analysis of the dynamic motifs that underlie this behavioral capability.