[Analysis of the effect of a training regime for the correct choice of matched learning in gray crows]

It was shown that a large set of training stimuli promotes abstract concept learning. These experiments were designed to assess whether an application of a large set of training stimuli would facilitate matching learning in crows. Four hooded crows were trained with a set of 72 unique combinations of stimuli in two-alternative simultaneous matching tasks with stimuli of three different categories: achromatic color (white, light-grey, dark-grey, and black), shape (Arabic numerals from 1 to 4 used as visual shapes only), and number of elements (heterogeneous graphic arrays from 1 to 4 items). Although the performance of all crows was significantly above chance (p < 0.01) in some 72-trial blocks, birds were unable to establish matching and to reach the criterion of learning 80% correct or better over 72 consecutive trials) in 5184 trials. Thus, the modified training procedure was less efficient than the training technique previously used (successive cyclic repetition of three small sets of training stimuli), which allowed four of six crows to acquire the matching rule after 1780, 2360, 3830, and 5260 trials [4,9].