Monkey Research

Cross-Modal Numerosity Matching from Brannon Lab on Vimeo.

Surely a monkey can discriminate two bananas from three bananas in order to choose the bunch with the greater amount of food. But when they make their choice, do they rely on number or some other stimulus attribute such as amount of yellow? If a monkey were able to make pure numerical discriminations, what would this say about her conceptual understanding of number? Do monkeys manipulate their representations of quantity using mathematical operations such as ordering, addition, and subtraction?

The Brannon lab uses psychophysical tools to investigate these questions and to probe the mechanisms by which monkeys represent number without language. Monkeys are trained in touch screen tasks to respond to pictures for juice reward.

We have used an ordering task that has demonstrated that monkeys not only discriminate numerosities but also appreciate their ordinal relations. Further analysis of the monkeys’ accuracy and speed of responding revealed that monkeys exhibit the same numerical distance and magnitude effects observed when adults compare Arabic numerals. In other words, both humans and monkeys are more likely to get an answer wrong, and it takes them longer to choose, when the two answer choices are closer in magnitude (a phenomenon called Weber’s law). This similarity between humans and monkeys suggests that numerical abilities are a phylogenetically ancient mechanism that evolved to organize the natural world.

In our lab, we have also used a match to sample (MTS) task that requires monkeys to match physically dissimilar pictures based on number (i.e., 3 oranges = 3 apples). We have used this paradigm to demonstrate that monkeys can match the number of sounds they hear to the number of shapes they see. We have also used this paradigm to demonstrate that monkeys can perform approximate addition. In a current study using the MTS paradigm we are exploring whether monkeys have a conceptual precursor to concept of “zero”. Our initial inquiries reveal that monkeys treat empty sets as values along an ordinal continuum.

Finally, in a collaborative project with Dr. Michael Platt’s lab, we are investigating the neural bases of numerical representation. These varied experiments will contribute towards the central goal of our research program which is to specify the nature of the computational processes that nonhuman primates use both in forming numerical representations and in making numerical computations.