Archive for the ‘Papers from 2015’ Category

The emergence of habitual routes

One of the interesting features of insect navigation systems is the way that innate guidance mechanisms (such as Path Integration or beacon aiming) interact with route learning. One such innate behaviour, which is essential in complex environments is obstacle avoidance, which is very rarely considered in accounts of route navigation. Bertrand et al. take a modelling approach, firstly looking at how a parsimonious model of obstacle avoidance can be built, but an interesting property of their testing is the emergence of habitual routes through clutter. Here is an extract from the author summary: “…. Inspired by the abilities of insects, we developed a parsimonious algorithm to avoid collisions in challenging environments solely based on elementary motion detectors. We coupled our algorithm to a goal direction and then tested it in cluttered environments. The trajectories resulting from this algorithm show interesting goal-directed behavior, such as the formation of a small number of routes, also observed in navigating insects.”
Bertrand, O. J., Lindemann, J. P., & Egelhaaf, M. (2015). A Bio-inspired Collision Avoidance Model Based on Spatial Information Derived from Motion Detectors Leads to Common Routes. PLoS Computational Biology, 11(11).
Categories: Papers from 2015

Visual resolution and view-based navigation

For any organism there are increased metabolic costs that come with increased resolution within a sensory system.  This simple fact can precede a line of thought where one think there must be a trade-offs between the cost of increased sensory resolution and the inherent value that comes from a greater amount of information . In this paper we look at the performance of hypothetical agents performing view-based navigation. In simulation we vary the visual resolution of agents and measure the consequent navigation performance. The interesting thing is that agents with low resolution perform better than agents with high resolution when asked to recapitulate a visually defined route. This simple demo shows how the sensory systems of animals need not always be a trade-off of performance and utility.

Wystrach, A., Dewar, A., Philippides, A., & Graham, P. (2015). How do field of view and resolution affect the information content of panoramic scenes for visual navigation? A computational investigation. Journal of Comparative Physiology A, 1-9.
Categories: Papers from 2015

Straight lines from coloured skys

Over the last few years, the concerted work of the Lund vision group has given us a great insight into the mechanisms of orientation that allow the plucky little dung beetle to roll his ball in a straight line. These beetles are able to use a range of orientation information from the sky (sun, moon, polarisation, even the milky way) and now a new cue is shown to provide orientation information for beetles. The sky has a chromatic gradient and beetles are shown to use such a gradient in order to roll their dung balls in a straight line.

Basil el Jundi, James J. Foster, Marcus J. Byrne, Emily Baird, Marie Dacke (2015) Spectral information as an orientation cue in dung beetles. Biology Letters, doi: 10.1098/rsbl.2015.0656

Categories: Papers from 2015

Interactions between behaviour and learning

Here is another nice review from the new journal Current Opinions in Insect Science, which promises to be an interesting read.
Abstract: “Flower patterns are thought to influence foraging decisions of insect pollinators. However, the resolution of insect compound eyes is poor. Insects perceive flower patterns only from short distances when they initiate landings or search for reward on the flower. From further away flower displays jointly form larger-sized patterns within the visual scene that will guide the insect’s flight. Chromatic and achromatic cues in such patterns may help insects to find, approach and learn rewarded locations in a flower patch, bringing them close enough to individual flowers. Flight trajectories and the spatial resolution of chromatic and achromatic vision in insects determine the effectiveness of floral displays, and both need to be considered in studies of plant–pollinator communication.”
de Ibarra, N. H., Langridge, K. V., & Vorobyev, M. (2015). More than colour attraction: behavioural functions of flower patterns. Current Opinion in Insect Science.
Categories: Papers from 2015

Different brain areas for different types of learning

Just a short note on an interesting new paper from PNAS and whilst this is not strictly a navigation paper, it is certainly very relevant. Devaud et al have investigated the role of the insect mushroom bodies (MBs) in non elemental learning. So called higher order learning might involve learning that the components A and B are negatively reinforced, whilst their mixture (AB) is positively reinforced. Intact MBs were necessary for the higher order learning tasks, but not simpler ‘elemental’ learning. This suggests that different brain areas might be needed for different styles of learning. Something that resembles the organisation of the vertebrate brain.
Devaud, J. M., Papouin, T., Carcaud, J., Sandoz, J. C., Grünewald, B., & Giurfa, M. (2015). Neural substrate for higher-order learning in an insect: Mushroom bodies are necessary for configural discriminations. Proceedings of the National Academy of Sciences, 201508422.
Categories: Papers from 2015

Memories of home vectors

PI can be used by ants to guide homeward and foodward memories, however it is unclear how distance and direction information becomes part of the long term memories of ants. Fernandes et al trained wood ants to find food at a fixed distance along a channel, with training locations dissociated from visual cues. One can test for vector memories by taking fed ants directly to the channel and observing homing behaviour that isn’t driven directly by PI. In this test ants walk in an appropriate distance and direction, these vector memories are primed simply by being in a fed state. A further batch of ants were trained (on alternate trials) to two feeder distances in opposite directions. Fed ants directly placed in the channel will choose a direction randomly, but then travel the appropriate distance for that direction. Thus distance and direction components are bound into insulated memories.

A.S.D. Fernandes, A. Philippides, T.S. Collett and J.E. Niven (2015) The acquisition and expression of memories of distance and direction in navigating wood ants. J Exp Biol doi:10.1242/jeb.125443

Categories: Papers from 2015

Cue integration in ants

For the control of behaviour, agents must take information from multiple sources and somehow merge them or decide between them in order to optimally control behaviour or develop accurate knowledge of the state of the environment.  Navigating insects provide a nice model for asking questions of this type as we can put learned visual information at odds with directions from path integration. Matthew Collett showed that both of these information sources can drive ant navigation at the same time. However, how they are averaged is unclear. Wystrach et al, used a nice experimental trick to help pick apart the weighting of these information sources. An insect’s path integration system has an increasing error with increasing length of the outward journey. That is to say, the region of the world within which the origin of the route might be increases in size. However, the error in the angular component of path integration decreases with distance away from the nest. So after a long journey, an ant far from the nest will have a more accurate departure bearing (compared to a short journey) even though the positional error in Path Integration is greater than for the small journey. Observing ants (with different length home vectors) when there is a conflict between visual information and PI, shows that ants’ resultant headings take into account the variability in the estimate of direction from the PI system. So far, so good, though an extra condition leads to an intriguing further result. The weighting of the PI component depends not on the total journey length (as would be optimal) but the current length of the home vector and therefore it seems that ants have an economical heuristic when deciding how to weight different cues.

Wystrach, A., Mangan, M., & Webb, B. (2015, October). Optimal cue integration in ants. In Proc. R. Soc. B (Vol. 282, No. 1816, p. 20151484). The Royal Society.
Categories: Papers from 2015