Explanations concerning the poster: Learning of Drosophila at the Flight Simulator: Classically Conditioned Visual Pattern Discrimination (mushroom bodies and cAMP signaling) Reinhard Wolf, Marcus Dill, Dirk Eyding, Tobias Wittig and Martin Heisenberg Theodor-Boveri-Institut für Biowissenschaften der Universität Würzburg, Lehrstuhl für Genetik, Am Hubland (Biozentrum) 97074 Würzburg, Germany

Explanations

• Contents:
• Yaw Torque
• Flight Simulator
• Closed Loop
• Open Loop
• Preference Test
• Operant Training
• Replay Training
• Performance Index PI

Yaw Torque:

Physical dimension of the power the fly has to generate to overcome its rotational momentum of inertia and the air friction, when it changes its flight direction in the horizontal plane.

Flight Simulator:

In contrast to free flight, where the fly is moving in the stationary world, in the flight simulator the fly is attached to a torque meter and is flying stationarily. The angular movements of an artificial visual panorama around the vertical body axis are calculated from the continuously recorded yaw torque and fed back to the fly via angular motion of the panorama as visual motion stimulus (Fig.1 and top of Fig.2). As seen from the position of the fly, clockwise directed yaw torque results in counterclockwise rotation of the panorama (and vice versa).

Closed Loop:

Flight simulator mode. The feedback loop [torque - visual motion stimulus] is closed (Fig.2). This situation enables the fly to have control of the visual stimulus.

Open Loop:

The feedback loop [torque - visual motion stimulus] is disconnected. Angular motion is not under control of the fly's yaw torque.

Preference Test:

Closed loop flight (without punishment) prior to operant or classical training. Serves to determine, whether the fly has a spontaneous preference for one of the pattern shapes. During this procedure, the fly may practice the control of the panorama in the flight simulator.

Operant Training:

During operant training the fly is in closed loop, like in the preference test. Depending on the fly's choice of flight direction with respect to the visual patterns at the arena wall, a beam of infrared light (heat) is switched on (or off) by the computer (see block diagram in Fig.2). For this, the entire panorama (360°) is subdivided into quadrants (invisible for the fly), each of them containing one of the T-shaped patterns in its centre. Whenever, caused by the fly's voluntary rotatory steering manoeuvres, one of the panorama's imaginary quadrant-boundaries crosses the fly's longitudinal body axis, heat is switched either on or off.

Replay Training:

During replay training, the whole sequence of pattern motion during the training session ('master training') is precisely recorded and afterwards played back to the same (or another) fly in open loop. As seen from the fly's point of view, the sensory stimuli (pattern motion and heat) in both types of training follow the same time schedule and have the same quality. The only difference between both training methods is, that the fly has no control of the stimuli during replay training (see block diagram on top of Fig.3). This makes the replay experiment a classical conditioning experiment.

Performance Index PI:

The performance index (PI) for the flight direction with respect to the 'non-heated' pattern is calculated as:

where tc indicates the time during which the fly is flying into the quadrant containing the 'cold' pattern, and th the time towards the 'hot' pattern, respectively. During the preceeding preference test, PI is a measure for the fly's spontaneous preference for a certain pattern type, during training it indicates the level of preference of the 'unpunished' pattern and during the following learning test it represents the degree of learning.

Go back to Poster | Go to Table of Contents