Encoding panoramic scenes

The idea that insects can use low-resolution scenes for navigation is well established. That they can do this without necessarily having to care about the specific objects (landmarks if you will) that make up that scene is also an idea that has recently gained traction. However, this doesn’t mean that insects have to use scenes as raw views akin to pixelated images. It very well might be that insects store and use a low dimensional encoding of panoramic scenes. In this paper Collett et al continue a research program looking at such possible encodings. Previously they have shown how ants might be guided by a strategy which sets directions based on the proportional amount of simulated terrestrial objects that are on the ants left and right visual field. Here, for shapes that are away from the mid-line of the visual system, ants seem to put the centre of mass of a shape at a fixed visual angle. Currently it is hard to think about the mechanisms underpinning these encodings, but the implications are clear – that visual guidance can proceed without regard for the fine details of shapes and panorama components.

Woodgate, J. L., Buehlmann, C., & Collett, T. S. (2016). When navigating wood ants use a shape’s centre of mass to extract directional information from a panoramic skyline. Journal of Experimental Biology, jeb-136697.