Spatial learning and visual navigation

Navigation requires bees to select the relevant cues amongst a multitude of visual features and then to choose the appropriate action amongst several behavioural options. Despite the large number of studies on navigation and pattern discrimination in bees (honeybees and bumblebees) it is still controversial what type of information is learnt by the bee during exploration behaviour and later used to localise the goal when returning to it. By using high-speed cinematography we analyse the flight manoeuvres of bumblebees during homing behaviour in indoor flight arenas, allowing us to compare the flight style, the learning progress and problem solving strategies of individuals. The basic design of these experiments is to allow the bees, initially unfamiliar with their environment, to become accustomed to seeing distinct objects close to a place of interest, i.e. their inconspicuous nest hole, and then displacing or modifying these objects or their background. It can thus be analysed where and by what flight and gaze strategy the bee searches for the goal. This enables us to identify the visual cues that are relevant during the last phase of homing. For understanding the mechanisms and algorithms underlying the navigation capabilities of bees computational models are developed at a variety of abstraction levels. These models are tested in computer simulations whose performance can be compared with the behaviour observed. One prerequisite is to know the visual input bees have experienced during their complex learning and searching flight manoeuvres. Therefore, we attempt to resolve both body and head movements of the animals as accurately as possible.

 

Research issues that are currently being investigated comprise:

  • Coordination of head and body movements during learning and return flights in the vicinity of a goal (i.e. the nest of a bumblebee) and how head-body coordination affects the behaviourally generated retinal image sequences.
  • The characteristics of learning behaviour in the vicinity of a goal and how learning progress affects the visual information available.
  • The information about the three-dimensional environment around a goal that is memorized during learning flights and how such information is used to ensure a successful return to the goal.
  • Interaction between mechanisms that provide information about the goal direction and collision avoidance during navigation in cluttered environments.