Agent-specific learning signals for self-other distinction during mentalising.

Zeb Kurth-Nelson,Raymond J Dolan,Ben Seymour,Sam Ereira

doi:10.1371/journal.pbio.2004752

Abstract

Humans have a remarkable ability to simulate the minds of others. How the brain distinguishes between mental states attributed to self and mental states attributed to someone else is unknown. Here, we investigated how fundamental neural learning signals are selectively attributed to different agents. Specifically, we asked whether learning signals are encoded in agent-specific neural patterns or whether a self–other distinction depends on encoding agent identity separately from this learning signal. To examine this, we tasked subjects to learn continuously 2 models of the same environment, such that one was selectively attributed to self and the other was selectively attributed to another agent. Combining computational modelling with magnetoencephalography (MEG) enabled us to track neural representations of prediction errors (PEs) and beliefs attributed to self, and of simulated PEs and beliefs attributed to another agent. We found that the representational pattern of a PE reliably predicts the identity of the agent to whom the signal is attributed, consistent with a neural self–other distinction implemented via agent-specific learning signals. Strikingly, subjects exhibiting a weaker neural self–other distinction also had a reduced behavioural capacity for self–other distinction and displayed more marked subclinical psychopathological traits. The neural self–other distinction was also modulated by social context, evidenced in a significantly reduced decoding of agent identity in a nonsocial control task. Thus, we show that self–other distinction is realised through an encoding of agent identity intrinsic to fundamental learning signals. The observation that the fidelity of this encoding predicts psychopathological traits is of interest as a potential neurocomputational psychiatric biomarker.

Highlights

Social interactions are underpinned by an ability to infer the mental states of self and others, referred to as mentalising [1]
During MEG scanning, they observed a sequence of samples from a Bernoulli distribution, with a drifting Bernoulli parameter P
We show that a representation of a learning signal (PE or belief) is encoded with a different neural spatial pattern when the signal is attributed to self as compared to when it is attributed to another agent

Summary

Introduction

Social interactions are underpinned by an ability to infer the mental states of self and others, referred to as mentalising [1]. The false belief task [30], a standard test of mentalising ability, requires that subjects make inferences about an environment and selectively attribute one belief-state to self and a different belief-state to another agent for whom the environment is only partially observable These belief-states are not informed by the behaviour of the subject or the other agent but arise through passively observing the environment. In this case, neural coding schemes that discriminate between executed behaviour and observed or imagined behaviour cannot facilitate a self–other distinction, and a more fundamental computation for selectively attributing signals to different agents is required. Our task design required sparse probe trials and a larger total number of trials than would be possible to acquire in a single fMRI session

Results

Discussion

Ethics statement