How to decide between visual guidance and path integration?
Most ant navigation papers, at some point in their introduction, will refer to the ant’s primary navigation systems of Path Integration (PI) and visual landmark use. Often overlooked are the questions of how these systems interact. Routinely, we all refer to demonstrations that, when in stark conflict, familiar visual cues will seem to be preferred over PI information. This suggests that the output of the ‘visual module’ suppresses entirely the output of PI. However in many of these experiments, the directions indicated by PI and visual guidance are ~180deg apart, thus it would be almost impossible to see any residual influence of PI on the directions taken. In order to address the question of whether PI and visual guidance interact in a more subtle way, Matthew Collett devised a simple experiment. Ants ran along a channel before an abrupt turn towards a feeder. At the turn point ants have information from their PI system and information learnt about the visual panorama. The two could be put in varying conflict by using a cart to passively transport ants along a portion of the channel, so that the state of their PI varied at the turning point. Matthew found that the direction of paths at the turn point was a compromise between PI and vision, rather than visual route memories suppressing PI entirely. From one perspective the result might be suggestive of ants possessing the cognitive sophistication to combine cues, perhaps in an optimal Bayesian way. Another perspective is that ants are demonstrating a cognitive shortcut, whereby they have a navigation system that doesn’t need a complex cognitive architecture to choose between guidance systems, because their outputs can be applied simultaneously. Further studies on the interaction of navigational cues promise to be fascinating.
Collett, M (2012) How Navigational Guidance Systems Are Combined in a Desert Ant, Current Biology, doi:10.1016/j.cub.2012.03.049