Abstract:Relational cognition—the ability to infer relationships that generalize to novel combinations of objects—is fundamental to human and animal intelligence. Despite this importance, it remains unclear how relational cognition is implemented in the brain due in part to a lack of hypotheses and predictions at the levels of collective neural activity and behavior. Here we discovered, analyzed, and experimentally tested neural networks (NNs) that perform transitive inference (TI), a classic relational task (if A > B and B > C, then A > C). We found NNs that (i) generalized perfectly, despite lacking overt transitive structure prior to training, (ii) generalized when the task required working memory (WM), a capacity thought to be essential to inference in the brain, (iii) emergently expressed behaviors long observed in living subjects, in addition to a novel order-dependent behavior, and (iv) expressed different task solutions yielding alternative behavioral and neural predictions. Further, in a large-scale experiment, we found that human subjects performing WM-based TI showed behavior inconsistent with a class of NNs that characteristically expressed an intuitive task solution. These findings provide neural insights into a classical relational ability, with wider implications for how the brain realizes relational cognition. The ability to infer underlying relationships in the world—or "relational cognition"—is fundamental to both human and animal intelligence. Despite this importance, it remains generally unknown how relational abilities are realized in the brain. Addressing this lack of knowledge requires generating testable hypotheses regarding the activity of neurons when relationships are inferred. In this study, we discovered artificial neural networks that perform transitive inference (TI), a classic relational task that has the form "if A > B and B > C, then A > C." We found that these neural networks show behavioral patterns that match those seen in humans and animals performing TI, and, further, performed TI in different ways, thereby providing different testable hypotheses regarding neural activity. We also conducted a large-scale behavioral study in humans, which suggested systematic differences in biological accuracy across the neural networks. These findings provide new hypotheses and insights regarding how TI, and relational abilities more generally, can be realized in the brain.

Cognitive representation in transitive inference: a comparison of four corvid species

Emergent neural dynamics and geometry for generalization in a transitive inference task

Reasoning by exclusion in food-caching Eurasian jays (Garrulus glandarius)

A mathematical theory of relational generalization in transitive inference

Investigating the relationship between physical cognitive tasks and a social cognitive task in a wild bird

Inferior parietal cortex represents relational structures for explicit transitive inference

The Cognitive and Neural Basis of Transitive Inference

Modeling within-session dynamics of categorical and item-memory mechanisms in pigeons

Reasoning by exclusion in the kea (Nestor notabilis)

Spatiotemporal variation in cognitive phenotype, social network position, and distribution of social associations in a food-caching bird

Comparative Cognition Research Demonstrates the Similarity between Humans and Other Animals

Transitive inference as probabilistic preference learning

Individual differences in co-representation in three monkey species (Callithrix jacchus, Sapajus apella and Macaca tonkeana) in the joint Simon task: the role of social factors and inhibitory control

Evidence for a general cognitive structure in pigeons (Columba livia)

Representation of the inferred relationships in a map-like space

Abstraction of Reward Context Facilitates Relative Reward Coding in Neural Populations of the Macaque Anterior Cingulate Cortex

Changes in the dopaminergic circuitry and adult neurogenesis linked to reinforcement learning in corvids

Abstract representations emerge in human hippocampal neurons during inference

The Role of Cognition in Social Information Use for Breeding Site Selection: Experimental Evidence in a Wild Passerine Population

Asymmetric learning and adaptability to changes in relational structure during transitive inference

A cortical circuit mechanism for structural knowledge-based flexible sensorimotor decision-making