Ants Learn Geometry and Features

Abstract

Rats trained to relocate a particular corner in a rectangular arena systematically confound the correct corner and the diametrically opposite one--this rotational error demonstrates the use of the geometry of space (i.e., the spatial arrangement of the different components of a visual scene). In many cases, geometric information is preferentially used over other spatial cues, suggesting the presence of a dedicated geometric module located in the parahippocampus and processing only geometric information. Since rotational errors were first demonstrated in 1986, the use of the geometry of space has attracted great interest and now seems to be widespread in vertebrate species, including humans. Until now, rotational errors have only been considered in vertebrate species. Here, for the first time, rotational errors are demonstrated in an insect. Our results, similar to those obtained with vertebrates, can be parsimoniously explained by a view-based matching strategy well known in insects, thereby challenging the hypothesis of a "geometric module" located in the animal's brain. While introducing a new concept of flexibility in the view-based matching theory, this study creates a link between two major topics of animal navigation: rotational errors in vertebrates and view-based navigation in insects.

Out of the box: how bees orient in an ambiguous environment

“…Thus, as shown for many other animals in similar experimental environments (e.g. Cheng, 1986;Kelly & Spetch, 2001;Sovrano, Bisazza, & Vallortigara, 2002;Wystrach & Beugnon, 2009), including recently for bumblebees in a different and five times smaller box (Sovrano et al, , 2013, bees also made rotational errors in our rectangular box. Note that bees did not always land in the corner where they first crossed the decision line.…”

“…In recent years, model simulations and behavioural experiments in ants have shown that rotational errors in rectangular arenas can be parsimoniously explained by view-based strategies without explicit representation of geometry Wystrach et al, 2011;Wystrach & Beugnon, 2009). The choice behaviour of the bees in our experimental set-up agrees very well with a simple image-matching model (see Fig.…”

“…Image matching is very successful in describing insect homing behaviour (e.g. Dittmar, Stürzl, Baird, Boeddeker, & Egelhaaf, 2010;Wystrach & Beugnon, 2009;Wystrach, Cheng, Sosa, & Beugnon, 2011;Zeil, 2012). It is based on the idea that homing insects move in such a way that the current visual input matches the 'snapshot' that was stored on previous visits (e.g.…”

“…Inner numbers show the mean percentage of the first corner choices, i.e. the corners where bees crossed the decision line for the first time (8 cm in front of the food hole, see dashed line) and outer numbers show the mean percentage of first landings in each corner (data presented are similar to those in Wystrach & Beugnon, 2009; mean percentages may not sum to 100% owing to rounding). Bees were trained to one corner in the rectangular box, which was marked by a blue paper stripe on the right and a yellow stripe on the left side of the food hole; the diagonally opposite corner had the same colour distribution.…”

Rudiments of mind: Insights through the chick model on number and space cognition in animals.

“…All species which have been investigated appear to mistake the target location with this rotationally equivalent position (reviewed in Vallortigara, 2009a). Very recently, even ants (Gigantiops destructor) have been proven capable of reorienting in a rectangular enclosure, displaying the same rotational errors of vertebrate species (Wystrach & Beugnon, 2009). The authors of this last study explained their findings within a view-based matching strategy, arguing that ants would match a view-based representation, stored and captured while moving, to the current view without encoding any geometry.…”