Many-to-One Matching with Temporal and Hedonic Samples in Pigeons

Pigeons were initially trained with either temporal samples (2- and 8-s keylights) or hedonic samples (food and no food) mapped to line-orientation (horizontal and vertical) comparisons. Delay testing revealed marked retention asymmetries in both groups (i.e., a choose-short effect with temporal samples and a choose-no-food effect with hedonic samples). Next, while both groups continued training on the original task, a second set of samples was added, hedonic for birds originally trained with temporal samples and temporal for birds originally training with hedonic samples. For all birds, food and short samples were associated with one comparison, and no-food and long samples were associated with the alternative comparison (many-to-one, MTO, mapping). This time, delay testing revealed symmetrical retention functions in both groups with both sets of samples, and mediated-transfer testing revealed positive transfer. It was concluded that (1) a common code was used to represent samples associated with the same comparison in the MTO mapping and (2) the content of the codes was unrelated to the identity of the samples.