Internal States vs. External Cues in Crawling Insect Larvae*

Organisms that can move will seek out conditions that improve their chances of survival. Most prominently, they explore and search for food, but also respond to spatially or temporally varying external environmental stimuli. They will move up or down gradients of temperature or chemical concentration, for example. Simple animals, such as the Drosophila larva we deploy here, often move probabilistically using modified versions of the classic random walk, combining random exploration with directed navigation by modulating components of their behavior in response to stimuli. Our investigations here focus on the internal states of crawling animals at the individual and population levels, and how those states manifest as physical exploration and navigation.

 

First, larvae crawling on a vertically vibrating platform alter their behavior very quickly in a form of adaptation to the stimulus. A Markov model of behavioral states shows a switch to weaker responses after repeated stimulus pulses, and we characterize distinct time constants for de-sensitization and re-sensitization, and other features that build a framework for studying behavior in strongly adaptive animals.

 

Second, we characterize larvae whose internal states have been modified through associative conditioning. Larvae can be trained to move towards an otherwise neutral temperature by pairing the temperature with a gustatory reward or punishment.

 

Third, we investigate how crawling behavior changes across the animals’ physical development, as larvae grow rapidly and change speeds and navigation strategies. An automated transport robot also lets us observe behavior continuously for many hours, and we find that individuals fall into distinct groups for thermal navigation strength, for example.

Finally, we examine “handedness,” a trait that decides right vs. left turning or drifting. We find strong evidence that crawling larvae exhibit individual direction preference, that the curved shape of their trajectories must be included accurately model their overall movement, and most interestingly that their direction bias can almost completely outcompete their normal response to external temperature gradients during navigation.