What the theory claims
IIT starts from experience itself rather than from the brain. Tononi argues that any conscious experience has certain essential properties, or axioms: it exists, it is structured, it is specific, it is unified, and it is definite. The theory then works out the physical requirements a system must satisfy to account for those properties. On this view consciousness is not something the brain produces as an output; it simply is the intrinsic cause-and-effect structure of a system considered from its own perspective.
The amount of consciousness a system has corresponds to phi, a measure of how much a whole generates more integrated cause-and-effect power than its parts taken independently. A system with high phi forms an irreducible whole. Where phi is zero, as in a set of disconnected components, there is no experience at all. Notably, IIT locates the physical substrate of consciousness in posterior regions of the cortex rather than in the front of the brain.
Who developed it, and how far it reaches
Giulio Tononi first proposed the framework in 2004, building on earlier work with Gerald Edelman, and has refined it through several versions, most recently IIT 4.0. The neuroscientist Christof Koch has been its most prominent champion. One striking implication is that consciousness comes in degrees and could exist, in small amounts, wherever integrated information exists, a stance close to panpsychism. IIT also predicts that a digital computer simulating a brain neuron by neuron could behave intelligently while having almost no experience, because its physical architecture would be poorly integrated.
Evidence and controversy
IIT has inspired real clinical tools. The perturbational complexity index, which pings the brain with magnetic stimulation and measures the complexity of the response, reliably separates wakefulness and dreaming from dreamless sleep and anesthesia. Critics counter that this practical success does not confirm the deeper theory. In 2023 more than 120 researchers signed an open letter arguing that, as commonly presented, IIT is untestable and should be labeled pseudoscience; defenders including Koch and Anil Seth replied that untestable in current practice is not the same as unfalsifiable, and many neutral observers found the label too harsh.
In 2025 the Cogitate consortium published a large adversarial study in Nature pitting IIT against global workspace theory. The results were mixed. Posterior cortex carried more of the conscious content than the rival theory expected, supporting one of IIT's claims, but the sustained synchronization within posterior cortex that the theory predicted did not clearly appear. The debate over whether phi captures consciousness, or only tracks it, remains open.