Ants integrate proprioception as well as visual context and efference copies to make robust predictions

The relative landmark shift during free movement on a treadmill may enhance visual learning for ant foragers

Ants are known for their sophisticated navigational abilities and rely on various cues to locate food sources efficiently. Visual landmarks play a crucial role in guiding foraging behavior. However, the significance of walking ants perceiving the relative motion of a landmark to themselves during learning is still not fully understood. Here, Japanese carpenter ants were observed on a treadmill device where they could walk in place on the treadmill. Two types of conditions were set for the training process: the fixed condition involved a stationary landmark, and the moving condition featured a landmark oscillating at a constant speed from side to side after feeding, which was independent of the movements of the ants. In the Y-maze test, a significantly greater number of ants in the moving training condition associated the landmark with food (82% of the ants) than in the fixed training condition (56% of the ants). Our results thus suggest that perceiving the relative movement of landmarks during the learning process on a treadmill is important for ant foragers.

Is this scenery worth exploring? Insight into the visual encoding of navigating ants

Solitary foraging insects such as desert ants rely heavily on vision for navigation. Although ants can learn visual scenes, it is unclear what cues they use to decide whether a scene is worth exploring at the first place. To investigate this, we recorded the motor behaviour of Cataglyphis velox ants navigating in a virtual reality setup and measured their lateral oscillations in response to various unfamiliar visual scenes under both closed-loop and open-loop conditions. In naturalistic-looking panorama, ants display regular oscillations as observed outdoors, allowing them to efficiently scan the scenery. Manipulations of the virtual environment revealed distinct functions served by dynamic and static cues. Dynamic cues, mainly rotational optic flow, regulated the amplitude of oscillations but not their regularity. Conversely, static cues had little impact on the amplitude but were essential for producing regular oscillations. Regularity of oscillations decreased in scenes with only horizontal, only vertical or no edges, but was restored in scenes with both edge types together. The actual number of edges, the visual pattern heterogeneity across azimuths, the light intensity or the relative elevation of brighter regions did not affect oscillations. We conclude that ants use a simple but functional heuristic to determine whether the visual world is worth exploring, relying on the presence of at least two different edge orientations in the scene.

Desert ants (Melophorus bagoti) oscillate and scan more in navigation when the visual scene changes

Solitarily foraging ants learn to navigate between important locations by comparing their current view with memorized scenes along a familiar route. As desert ants, in particular, travel between their nest and a food source, they establish stable and visually guided routes that guide them without relying on trail pheromones. We investigated how changes in familiar visual scenes affect the navigation of the red honey ant (Melophorus bagoti). In Experiment 1, ants were trained to follow a one-way diamond-shaped path to forage and return home. We manipulated scene familiarity by adding a board on their homebound route just before the nest. In Experiment 2, ants were trained to travel a straight path from their nest to a feeder, and we removed the prominent landmarks on the route after they had established a stable route. We predicted that these scene changes would cause the ants to deviate from their usual straight paths, slow down, scan more, and increase their lateral oscillations to gather additional information. Our findings showed that when the familiar scene was changed, ants oscillated more, slowed their speed, and increased scanning bouts, indicating a shift from exploiting known information to more actively exploring and learning new visual cues. These results suggest that scene familiarity plays a crucial role in ant navigation, and changes in their visual environment lead to distinct behavioral adaptations aimed at learning about the new cues.

The neuroethology of ant navigation

Unlike any other group of animals, all ant species are social: individual ants share the food they gather with their nestmates and as a consequence they must repeatedly leave their nest to find food and then return home with it. These back-and-forth foraging trips have been studied for about a century and much of our growing understanding of the strategies underlying animal navigation has come from these studies. One important strategy that ants use to keep track of where they are on a foraging trip is 'path integration', in which they continuously update a 'home vector' that gives their estimated distance and direction from the nest. As path integration accumulates errors, it cannot be relied on to bring ants precisely home: such precision is accomplished by using views of the nest acquired before they start foraging. Further learning is scaffolded by home vectors or remembered food vectors, which guide a route and help in learning useful views experienced on the way. Many species rely on olfaction as well as vision for route guidance and the full details of their foraging paths have revealed how ants use a mix of innate and learnt multisensory cues. Wood ants, a species on which we focus in this review, take an oscillating path along a pheromone trail to sample odours, but acquire visual information only at the peaks and troughs of the oscillations. To provide a working model of the neural basis of the multimodal navigational strategies of ants, we outline the anatomy and functioning of major central brain areas and neural circuits - the central complex, mushroom bodies and lateral accessory lobes - that are involved in the coordination of navigational behaviour and the learning of visual and olfactory patterns. Because ant brains have not yet been well-studied, we rely on the work that has been done with other species - notably, Drosophila, silkworm moths and bees - to derive plausible neural circuitry that can deliver the ants' navigational strategies.

Ants integrate proprioception as well as visual context and efference copies to make robust predictions

Forward models are mechanisms enabling an agent to predict the sensory outcomes of its actions. They can be implemented through efference copies: copies of motor signals inhibiting the expected sensory stimulation, literally canceling the perceptual outcome of the predicted action. In insects, efference copies are known to modulate optic flow detection for flight control in flies. Here we investigate whether forward models account for the detection of optic flow in walking ants, and how the latter is integrated for locomotion control. We mounted Cataglyphis velox ants in a virtual reality setup and manipulated the relationship between the ants' movements and the optic flow perceived. Our results show that ants compute predictions of the optic flow expected according to their own movements. However, the prediction is not solely based on efference copies, but involves proprioceptive feedbacks and is fine-tuned by the panorama's visual structure. Mismatches between prediction and perception are computed for each eye, and error signals are integrated to adjust locomotion through the modulation of internal oscillators. Our work reveals that insects' forward models are non-trivial and compute predictions based on multimodal information.