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

Area of Science:

• Neuroscience
• Cognitive Science
• Computational Psychiatry

Background:

• Humans possess a sophisticated ability to infer the mental states of others.
• The neural mechanisms underlying the distinction between self-attributed and other-attributed mental states remain largely unknown.

Purpose of the Study:

• To investigate how the brain assigns fundamental neural learning signals to distinct agents (self vs. other).
• To determine if learning signals are encoded in agent-specific neural patterns or if agent identity is encoded separately.

Main Methods:

• Subjects learned two models of an environment, one attributed to self and one to another agent.
• Magnetoencephalography (MEG) and computational modeling tracked neural representations of prediction errors and beliefs for self and other.
• Decoding analyses assessed the neural encoding of agent identity.

Main Results:

• Neural patterns of prediction errors reliably predicted the agent (self or other) to whom they were attributed.
• A weaker neural self-other distinction correlated with reduced behavioral self-other differentiation and increased subclinical psychopathological traits.
• Agent identity decoding was reduced in a nonsocial control task, indicating the influence of social context.

Conclusions:

• Self-other distinction is achieved through the intrinsic encoding of agent identity within fundamental neural learning signals.
• The fidelity of this neural encoding may serve as a neurocomputational biomarker for psychopathological traits.