"experiment 1 tested if honeybees might have sufficient visual"
Face recognition by bees: non-specialised brains demonstrate an ability to both bind and integrate image features. AG Dyer PhD, Centre for Brain and Behavior, Physiology Dept., Monash University, Clayton 3800, Vic., Australia. Email: firstname.lastname@example.org Institut fur Zoologie III (Neurobologie), Johannes Gutenberg University, Mainz 55099, Germany. Abstract: How organisms learn to recognise faces is of considerable interest, but can be confounded by the difficulty of selecting appropriate controls for ontogenetic history. In this context, studies of social insects can produce insights on what is considered to be a high-order visual task. We use free-flying honeybees (Apis mellifera) as a model to understand the mechanisms a non-mammalian brain uses to recognise human faces. Individual bees were trained with differential conditioning to achromatic target and distractor face images presented on a rotating screen (J Exp Biol v208p4709), then bee performance was evaluated in non-rewarding transfer tests using image manipluations that have previously been useful for understanding face processing by primate brains. When face images excluded either the inner (eyes, nose, mouth) or outer (hair, ears) salient features (for human e.g. see Child Dev v77p297) bee choices for target stimulus were significant from chance suggesting bees bind both internal and external features in face recognition. However, if faces were vertically devided into five sectors containing all features but disrupting configuration bee choices dropped to random (humans can solve this task, but with some degree of impairment; Perception v29p893). Bees could not recognise stretched face stimuli, an image manipulation that does not impair human face processing (Perception v31p1221). Bee face processing appears to be independent of image brightness, and bees reliably recognise faces that have had brightnes inverted (negative faces; a problem humans are poor at solving; Perception v26p1289). Finally, we evaluated the ability of miniature brains to process rotated face stimuli (a major imaging problem that has recently been shown to be solved by newborn human infant brains; Turati et al. 2008 Cognition, 106p1300). Bees could not recognise novel 30 deg. rotated faces after differential condition to either plane (0 deg.), or 60 deg. rotated faces (independent groups); however, bees that received conditioning to both (0 deg.) and 60 degree faces were able to reliably recognise target face presented at a novel 30 deg. angle. This finding points towards image integration by honeybee brains to solve a novel task (PLoS ONE 3: e4086 http://www.plosone.org/article/info:doi/10.1371/journal.pone.0004086).