Interactions occurring within the space near the body are studied in a wide range of disciplines, from ethology to philosophy. Indeed, many behavioural responses are enhanced when stimuli occur near certain body parts. This makes evolutionary sense: a predator within striking distance is more relevant than one farther away. Neuroscientific studies in primates have suggested a physiological foundation for such behavioural modulations, leading to the concept of peripersonal space (PPS). But what is precisely meant when referring to PPS? Predominant conceptual frameworks describe PPS as a single, distance-based, in-or-out zone within which stimuli elicit enhanced neural and behavioural responses. In this talk I will first show that this intuitive framework is contradicted by neurophysiological and behavioural data. I will then argue that the so-called PPS measures do not represent stimulus proximity, but rather the value of actions aiming to create or avoid contact between objects and the body and that for this reason they should be referred to as bodypart-centred response fields. This reconceptualisation of PPS as a set of graded egocentric fields describing the value of contact actions takes into account mainstream theories of action selection and behaviour. I will finally demonstrate, using reinforcement learning in artificial neural networks, that bodypart-centred response fields arise naturally from two simple and plausible assumptions about living agents: 1) they experience reward when they contact objects in the environment, and 2) they act to maximise reward. This perspective reproduces multiple foundational findings in the peripersonal space literature (seamlessly reconciling a number of contradicting empirical observations), provides testable predictions, and subsumes existing formal models of the so-called peripersonal space (PPS).
[mehr]