A topic in the Open Knowledge Graph — a free, open map of 15,290 topics and the order to learn them in.
The mentalizing network, including temporo-parietal junction (TPJ), superior temporal sulcus (STS), and medial prefrontal cortex (mPFC), represents the mental states of others. These regions integrate behavioral cues and contextual information to infer beliefs, desires, intentions, and perspectives, enabling social prediction, cooperation, and moral reasoning.
From your prerequisite study of theory of mind development, you know that the ability to represent others as having distinct beliefs, desires, and intentions — mental states that may differ from your own and from reality — is a cognitive milestone that emerges in early childhood and undergirds much of human social life. From your study of social cognition, you know that mentalizing refers to the ongoing, often automatic process of reading and predicting others' minds. The question here is: what does the brain do when it mentalizes, and which neural structures implement this capacity?
The temporo-parietal junction (TPJ), a region at the intersection of the temporal and parietal lobes (roughly above and behind the ear), is among the most consistently activated regions in neuroimaging studies of false-belief tasks and perspective-taking. Its key role appears to be representing the distinction between one's own perspective and another's — specifically, it becomes most active when those perspectives conflict, as in a false-belief scenario where another person holds a belief you know to be false. The TPJ does not represent mental content in full detail; rather, it seems to act as a perspective-decoupling mechanism, flagging that the other person's representation of the world must be tracked separately from your own model of reality.
The medial prefrontal cortex (mPFC) makes a complementary contribution. Whereas TPJ supports the structural distinction of perspectives, mPFC is more involved in the content of mental states — particularly beliefs and desires with social and emotional significance. The mPFC is also active during self-reflection and during evaluation of social norms, consistent with its broader role connecting self-representation to other-representation. The superior temporal sulcus (STS) contributes earlier in the processing chain, extracting intention-relevant cues from biological motion and gaze direction — reading the building blocks of intentional action before full mental state inference occurs.
These three regions function as a coordinated network, not as independent modules. They receive input from sensory areas detecting socially relevant cues (gaze, gesture, facial expression, speech prosody), integrate that input with stored knowledge about people's goals and contexts, and generate probabilistic inferences about what others believe and intend. This is why mentalizing is computationally demanding and why it fails — or relies on heuristics — under cognitive load or time pressure. It is also why mentalizing is impaired when the network itself is disrupted, as appears to be the case in certain neurodevelopmental conditions like autism spectrum disorder, where reduced activation of TPJ and altered mPFC engagement during social tasks correlates with difficulties in spontaneous perspective-taking. Understanding the mentalizing network reveals that social intelligence is not a diffuse, mysterious faculty — it is implemented in identifiable neural circuits carrying out specifiable computations.