摘要：Learning styles in capuchin monkeys were assessed with a computerized reversal- learning task called the mediational paradigm. First, monkeys were trained to respond with 90% accuracy on a two-choice discrimination (A+B–). Then the authors examined differences in performance on three different types of reversal trials (A–B+, A–C+, B+C–), each of which offered differing predictions for performance, depending on whether the monkeys were using associative cues or rule-based strategies. Performance indicated that the monkeys mainly learned to avoid the B stimulus during training, as the A–C+ condition produced the best performance levels. Therefore, negative stimuli showed greater control over responding after reversal and reflected a more associative rather than rule-based form of learning.
