Pattern recognition in tufted capuchin monkeys (Cebus apella): the role of the spatial organisation of stimulus parts.

We report four experiments aimed at characterising the role played by the encoding of the spatial relationship between stimulus parts in pattern recognition in capuchin monkeys, as assessed by a matching to sample task. The results of the first experiment, which were also reliably replicated at different stages in the course of the study, indicated that the simultaneous rotation and/or translation of the four parts into which the stimuli were divided, but not a global rotation of the entire stimulus, impaired matching performance in capuchin monkeys. Experiments two and three showed that matching performance was not impaired following similar manipulations of a subset of one, two or three parts. In experiment four, the same task was presented to human subjects. The same pattern of results emerged for humans and monkeys in trials where all the four stimulus parts were presented. However, the matching performance of humans was affected more than that of capuchin monkeys when only a subset of stimulus parts was featured in the task. These results support the conclusion that the matching performance of capuchin monkeys is affected by the rearrangement of stimulus parts and, as such it seems to rely on global properties of the stimulus such as the spatial relationships of the component parts. However, the remarkable ability of capuchin monkeys to identify a stimulus on the basis of a subset of parts suggests that the reliance on the global properties of the stimuli may not be pervasive as it is in humans.