Abstract
Purpose. Using a flight simulator for Drosophila melanogaster, Ernst & Heisenberg (1999) reported that several parameters like size, color, and vertical compactness could be used as cues for visual discrimination in tethered flies. However, a puzzling result was found in their original study—flies failed in conditioned discrimination between a cross (+) and a circle (○), but showed spontaneous preferences towards the circle. Using a T-maze like apparatus (Tully & Quinn, 1985), we previously re-examined Ernst & Heisenberg's (1999) finding and found that flies could in fact learn to discriminate between a circle and a cross (Chien, Lin, Chen, & Lai, APCV 2008). In the present study, we further investigated whether flies can recognize luminance-defined shapes and generalize them to the same shapes that were defined by the law of continuity.
Methods. The same T-maze like device was used as the main training/testing apparatus. The stimuli were printed paper and were back illuminated by LED lights. We first trained flies to discriminate between a circle and a cross sequentially with one being associated with heat (aversive conditioning). In the test phase, flies were presented with the two stimuli located on the opposite arms. The rationale was that if the flies did learn which stimulus was unsafe (i.e. the heated one), we expected to see significantly fewer flies would stay in the unsafe arm than in the safe arm.
Results. Tested with about 500 flies in each experimental condition, our data indicated that, when trained with luminance-defined shapes, flies were able to show statistically significant learning behavior for the same pair of shapes but were made of dots (the law of continuity) instead of solid lines. This finding supports the notion that free flying Drosophila can learn to recognize and generalize simple shapes. Further control experiments are in progress.
This project was supported by Grant# NSC 97-2410-H-039-006 to Dr. Sarina H. L. Chien.