Graziano Lab

  Department of Psychology, Princeton University, Princeton, NJ 08544

Michael Graziano, Phone : (609) 258-7555, Fax : (609)258-1113, Email :


Consciousness Research

Past Research





Graziano's CV

Psych Home


at Princeton

The Attention Schema Theory

A neuro-cognitive theory of what consciousness is and how it evolved

The theory at a glance: from selective signal enhancement to consciousness.

About half a billion years ago, nervous systems evolved an ability to enhance the most pressing of incoming signals. Gradually, this ability to focus on selected signals came under a more sophisticated, top-down control and became what is now called attention.

In control theory, if a brain is to control something, it should have an internal model of the thing to be controlled. According to the "attention schema theory", to effectively deploy its own attentional focus, the brain needs a constantly updated simulation or model of attention. Otherwise the brain would not possess explicit knowledge about its changing state of attention, or about the consequences of attending to something. This model of attention is schematic and lacking in detail.

There is no adaptive reason for a brain to know that it has electrochemical signals passing through neurons, or that the signals compete in a complex manner that results in some signals becoming enhanced, or that the enhanced signals have more influence over the parts of the brain involved in decision-making, movement, and memory. Brains don’t need that detailed or accurate information about themselves in order to function. Instead, the simplified model of attention attributes to the self an experience of X -- the property of being conscious of something. In this theory, a brain attributes to itself, "I am aware of X, in the sense of mentally possessing X and being able to react to X," because that attribution is a good, if simplified, model of the much more complex process of paying attention to X. The model helps keep track of the ever-changing state of attention and helps to predict the consequences of attention. Just as the brain can direct attention to external signals or to internal signals, that model of attention can attribute to the self a consciousness of external events or of internal events. Self awareness, awareness of emotions, awareness of one’s own thoughts, awareness of sensory events, all of these types of awareness can be accommodated by this theory.

In this theory, a brain does not actually have awareness. Instead it has attention, a mechanistic process. It also has information, in an internal model, that tells it that it has awareness. The information describes a self that experiences something and that can choose to react to and remember that something. The reason for this information is that it is a useful, if approximate, description of attention. The brain is captive to that internal information. On introspection — when relying on internal data — the system will always conclude that it has awareness, because that is what its internal models tell it.

As the model of attention increased in sophistication through evolutionary time, we hypothesize that it came to be used for a variety of other cognitive purposes. It may have enhanced the integration of information in the brain. For example, if your brain is attending to an apple, a model of that internal state requires a model of the apple, a model of yourself, and a model of the act of attention. These disparate pieces of information must be linked together — much like color and shape information must be linked together to form a visual model of the apple. Your brain would then possess an internal model that says, in effect, “I am aware of the apple.” An internal model of attention therefore fundamentally links information across many domains, especially between information about the self and information about the outside world.

Another use of an internal model of attention is to model the attentional state of other individuals to gain better prediction of their behavior. We suggest that in the human brain, similar and partly overlapping mechanisms attribute awareness to oneself and attribute awareness to others.

It is not clear when in evolution the social attribution of awareness began to emerge. The accompanying diagram places it at the start of primate evolution, 65 million years ago (MYA), but it could have begun much earlier. Perhaps most birds and mammals have some ability to attribute awareness to each other. Another possibility is that the social use of awareness expanded much later with hominins, beginning about 6 MYA. Now, in humans, consciousness plays a major role in social and cultural capability. We paint the world with perceived consciousness. Family, friends, pets, spirits, gods, these are all suffused with attributions of consciousness.

In this theory, awareness, the ability of brains to attribute to themselves a subjective experience of something, emerged first with a specific function related to the control of attention. It continues to evolve, however, expanding its cognitive role, becoming the intricate lattice of cognitive and social properties we call consciousness.

The attention schema theory is entirely mechanistic and therefore scientifically testable. In this theory, awareness is not a fuzzy philosophical flourish, but a key part of the brain's machinery for processing data. My lab is currently focused on testing predictions of the theory. For further information, see my book (Consciousness and the Social Brain, 2013, Oxford University Press). See also the most recent publications from my lab, available on my Publications page.

Timeline: Hydras evolved approximately 550 MYA with no selective signal enhancement. Animals that do show selective signal enhancement diverged from each other approximately 530 MYA. Animals that show sophisticated top-down control of attention diverged from each other approximately 350 MYA. Primates first appeared approximately 65 MYA. Hominins appeared approximately 6 MYA. Homo sapiens appeared approximately 0.2 MYA.