From peripersonal space to cognitive maps: An evolutionary perspective.

Abstract

Our brains support at least two major spatial representations: an egocentric, short-horizon representation of peripersonal space (PPS), and an allocentric, longer-horizon cognitive map system. Here, we speculate on their relationship from an evolutionary perspective. We argue that an ancient proto-PPS system for contact prediction likely emerged early in vertebrate evolution as a fight-or-flight mechanism for managing imminent threats, supported by evolutionarily conserved midbrain circuits such as the optic tectum (superior colliculus in mammals). The subsequent transition of many species from aquatic to terrestrial environments dramatically expanded sensory ranges and environmental complexity, arguably creating selective pressures for longer-horizon navigation, prediction, and planning. These pressures likely contributed to the emergence of hippocampal and entorhinal systems supporting allocentric spatial representations. Next, as navigation and foraging behaviors became more sophisticated and required dexterous multi-joint movements, it is possible that a cortical PPS system emerged within neocortical networks, including posterior parietal and ventral premotor cortices. Rather than mediating coarse fight-or-flight responses, we suggest this cortical PPS system now supports fine-grained hierarchical sensorimotor control, reference-frame transformations, tool use, and social interactions. In parallel, the allocentric mapping systems seemingly expanded beyond spatial navigation to support relational, conceptual, and abstract cognitive structures. We suggest that PPS and cognitive maps instantiate shared computational principles operating at different spatial and temporal scales and in different reference frames; a view supported by recent reinforcement-learning frameworks linking body-centered and allocentric predictive value. Together, this perspective points to a putative scaffolded evolutionary trajectory in which cognition may have expanded from near-body safety to a multiscale predictive architecture supporting flexible, goal-directed behavior across space and time.