What Is Attention Schema Theory?
Attention Schema Theory (AST) offers one of the most mechanistically specific accounts of why consciousness exists and why it feels the way it does. Developed by neuroscientist Michael Graziano at Princeton University, AST proposes that consciousness is the brain's simplified internal model of its own process of attention. Just as your brain maintains a body schema — an internal model of your body's position, posture, and boundaries — it maintains an attention schema: a model of what your attention is doing, what it is focused on, and what it is like.
The twist is that this model, like all schemas, is a caricature. It captures the essential information while omitting the messy physical details. When your brain models its own attention, it does not represent neurons firing — it represents an abstract essence: "I am subjectively aware of this thing." And because the model omits the physical machinery, the experience of consciousness seems non-physical, mysterious, and irreducible. AST says this mystery is not a feature of consciousness itself but an artifact of the brain's simplified self-model.
The Core Framework
AST begins with the observation that attention is a real, well-studied neural process. When you attend to something, certain signals in the brain are enhanced while others are suppressed. This is competitive signal processing, implemented by known neural circuits. But the brain also needs to monitor and control this process — it needs to know what it is attending to and how to redirect attention.
The brain accomplishes this by constructing an internal model of attention — the attention schema. This model describes attention in simplified terms: there is a subjective essence (awareness) that is directed at an object, that is experienced by a self, and that has a certain quality. The schema does not describe the underlying physical process because, for control purposes, it does not need to.
Graziano draws a direct analogy to the body schema. The body schema models your arm as a solid, bounded object — it does not represent individual cells, blood vessels, or muscle fibers. This simplified model is useful for controlling movement. Similarly, the attention schema models attention as a non-physical inner experience — useful for controlling attention but fundamentally inaccurate about its true nature.
When you introspect and conclude that you have a non-physical conscious experience, you are reading out the attention schema. The schema says: "I have awareness, and it is directed at this stimulus." Because the schema lacks any reference to physical mechanism, the experience it describes seems inherently non-physical. AST predicts that the hard problem of consciousness is generated by this gap in the model — not by any actual gap in nature.
Who Proposed It
Michael Graziano is a professor of psychology and neuroscience at Princeton University. His earlier work focused on the body schema and peripersonal space — how the brain models the space immediately surrounding the body and uses this model to guide action. He extended this framework to consciousness in a series of papers beginning in 2011 and in his books "Consciousness and the Social Brain" (2013) and "Rethinking Consciousness" (2019). His lab combines theoretical work with experimental studies of attention, awareness, and social cognition in both humans and monkeys.
Key Evidence
AST draws on several lines of evidence. Studies of hemispatial neglect — a condition caused by damage to the right temporoparietal junction (TPJ) — show that patients lose awareness of the left side of space while retaining some ability to process left-side information unconsciously. AST interprets this as damage to the attention schema: the brain can still direct attention to the left, but it has lost its model of that attention, so the patient is unaware of being aware.
The TPJ and adjacent superior temporal sulcus (STS), which AST identifies as key areas for constructing attention schemas, are also the brain regions most consistently associated with theory of mind — modeling others' mental states. This convergence supports AST's prediction that self-awareness and other-awareness share a common mechanism.
Graziano's lab has conducted experiments showing that people readily attribute awareness to simple animated objects when those objects behave as if they are directing attention. This suggests the brain's awareness-attribution system (the attention schema) is promiscuous — it applies its model liberally, explaining why we intuitively attribute consciousness to other people, animals, and even inanimate objects under certain conditions.
Key Objections
Critics argue that AST, like other deflationary theories, does not truly explain phenomenal consciousness. Even if the brain builds a model of attention and this model lacks physical details, there is still the question of why introspecting on this model feels like something. The model might explain why we believe we are conscious, but does it explain why there is an experience of believing?
Ned Block argues that AST conflates access consciousness (information available for report and reasoning) with phenomenal consciousness (subjective experience). AST may explain the former while leaving the latter untouched.
Some philosophers question the analogy between body schema and attention schema. The body schema works because there is a real body to model. But if consciousness is just the model, then what is the model modeling? If attention is the real phenomenon and consciousness is just the brain's (imperfect) representation of it, has AST actually explained consciousness or explained it away?
Why It Matters
AST matters because it provides a concrete, neurally grounded mechanism for why consciousness exists and why it has the properties it does. Unlike theories that start with the hard problem, AST starts with known neuroscience — attention, internal models, social cognition — and builds outward. It makes specific predictions about which brain regions are involved, what deficits should result from damage, and even how artificial consciousness could be engineered. If AST is correct, consciousness is not a fundamental mystery but an engineering problem: build the right kind of self-model, and awareness follows.
