What Is Theory of Mind in Autism and How It Works?

Theory of mind is the ability to understand that other people have their own thoughts, beliefs, and emotions that may differ from yours. In autism, this ability develops differently, making it harder to intuitively grasp what someone else is thinking or feeling in the moment. This doesn’t mean autistic people lack empathy or don’t care about others. It means the automatic, behind-the-scenes mental process of reading other minds works on a different track.

How Theory of Mind Works

Every time you predict how a friend will react to bad news, guess why a coworker looks upset, or realize someone is being sarcastic, you’re using theory of mind. It’s a set of cognitive skills that lets you build a mental model of another person’s internal state and use that model to navigate social situations. In typical development, children begin mastering this around age four or five, when they can understand that someone else can hold a belief that’s actually wrong.

The classic test for this is the “false belief” task. A child watches a scene where a character puts a marble in a box, then leaves the room. While the character is gone, someone moves the marble to a basket. The child is asked: where will the character look for the marble? Children who’ve developed theory of mind say the box, because they understand the character still believes the marble is there. Children who haven’t developed it yet say the basket, because that’s where the marble actually is. They can’t separate their own knowledge from what the character knows.

Before children reach this milestone, several building-block skills appear first: following someone’s eye gaze, establishing joint attention (looking at the same thing as another person and knowing you’re both focused on it), imitating actions, engaging in pretend play, and recognizing basic emotions on faces. These precursor skills form the scaffolding that theory of mind is built on.

What’s Different in Autism

Autistic people often find it genuinely difficult to automatically read what’s going on in someone else’s mind. This isn’t about intelligence. Even highly capable autistic adults can struggle with complex theory of mind tasks that require tracking multiple people’s beliefs simultaneously or reading subtle social cues in real time. The difficulty appears to persist from childhood through adulthood, particularly for the spontaneous, split-second kind of mental state reading that happens without conscious effort.

In practice, this can look like difficulty understanding why someone is upset when the reason hasn’t been stated directly, missing sarcasm or white lies, not realizing that a conversation partner has lost interest, or struggling to predict how someone will respond to surprising news. It can also affect the earlier building-block skills. Some autistic children show differences in joint attention, imitation, and pretend play well before the age when false belief understanding would typically emerge.

It’s worth being specific about what theory of mind difficulties are not. They’re not a lack of caring. An autistic person who doesn’t notice that a friend is sad isn’t choosing to ignore the friend’s feelings. The signal just doesn’t register the same way. Once an autistic person does understand what someone else is feeling, they often respond with deep concern.

What Happens in the Brain

Neuroimaging research has identified a network of brain regions involved in theory of mind reasoning. The medial prefrontal cortex, a region behind the forehead involved in thinking about the self and others, shows up in nearly all studies. Other key areas include the temporoparietal junction (active in about 58% of studies), which helps distinguish your own perspective from someone else’s, and superior temporal regions involved in processing social signals like facial expressions and tone of voice. The anterior cingulate cortex, amygdala, and temporal poles also contribute.

In autistic individuals, this network appears to function differently, though researchers haven’t pinpointed a single area responsible for theory of mind differences. The picture is more of a distributed network where the coordination between regions, rather than any one region’s activity, seems to be what’s altered.

The Double Empathy Problem

The traditional framing of theory of mind in autism treats it as a one-sided deficit: autistic people can’t understand non-autistic minds. But a more recent perspective, proposed by autistic researcher Damian Milton, challenges this view. Milton’s “double empathy problem” points out that the difficulty goes both ways. Non-autistic people are also poor at reading autistic people’s emotions, intentions, and communication styles.

The logic is straightforward. Theory of mind partly relies on shared experience. When two people have similar ways of processing the world, they can predict each other’s reactions more easily. When their experiences and communication styles diverge significantly, as they do between autistic and non-autistic people, both sides struggle to understand the other. Research has supported this: autistic people often communicate effectively with other autistic people, and non-autistic people frequently misread autistic people’s emotions and intentions.

Because the non-autistic way of communicating is treated as the default, autistic people are typically the ones expected to bridge the gap. Milton’s argument is that this expectation is unfair. If non-autistic people struggle just as much to understand autistic perspectives, the “deficit” isn’t located in one group. It’s a mutual problem rooted in difference, not dysfunction.

How Adults Compensate

Many autistic adults develop workarounds for theory of mind challenges, often without realizing they’re doing it. These compensatory strategies can be remarkably effective, sometimes masking the underlying difficulty entirely. Common approaches include memorizing social rules (“when someone says ‘how are you,’ they expect a short positive answer”), preparing scripts for predictable conversations, studying facial expressions intellectually rather than reading them instinctively, and using logic to deduce what someone might be feeling based on the situation rather than picking it up from nonverbal cues.

This compensation comes at a cost. It’s mentally exhausting to consciously process what other people handle automatically. An autistic person might navigate a work meeting successfully by actively monitoring everyone’s expressions, tracking the conversational flow, and selecting appropriate responses, but feel drained afterward in a way their colleagues don’t. The strategies also tend to be rigid and context-dependent. A rule that works in one social setting may not transfer to another, which can make unfamiliar situations especially stressful.

Building Theory of Mind Skills

Structured training programs can help autistic children develop stronger theory of mind abilities. These programs typically work through a developmental sequence, starting with recognizing basic emotions like happiness, sadness, fear, and anger in pictures and cartoons. From there, children practice connecting emotions to situations (understanding why a character feels a certain way), then move to understanding desires, and finally to understanding beliefs, including false beliefs.

One well-studied approach uses 15 sessions over five weeks, progressing through emotions, situational emotions, desires, beliefs, and combined desire-belief reasoning. Children who completed this kind of training showed significant improvements in social skills as rated by both parents and teachers, compared to children who didn’t receive the training. The gains weren’t just on lab tests; they translated into real interactions.

These interventions work best when they’re concrete and visual, using cartoon scenarios, role-playing, and step-by-step breakdowns of social situations. The goal isn’t to make autistic children think like non-autistic children. It’s to give them explicit tools for understanding perspectives that don’t come to them intuitively, much like learning a second language through grammar rules rather than immersion. The skill may never feel fully automatic, but it becomes functional.