Criticality as a Determinant of Integrated Information Φ in Human Brain Networks

Integrated information theory (IIT) holds that consciousness arises from information that is both highly differentiated and integrated across a system's parts. Using a large-scale brain network model and high-density EEG recordings from people under general anesthesia, this work found that network criticality—a balance between flexible functional configurations and structural constraints—coincides with maximal integrated information (Φ). As consciousness diminished, both criticality and Φ decreased. The conscious resting state showed the largest Φ and criticality, while the balance between variation and constraint broke down as behavioral responsiveness declined. The findings suggest network criticality is a necessary condition for large integrated information in the human brain.

Study at a glance

Abstract

Integrated information theory (IIT) describes consciousness as information integrated across highly differentiated but irreducible constituent parts in a system. However, in a complex dynamic system such as the brain, the optimal conditions for large integrated information systems have not been elucidated. In this study, we hypothesized that network criticality, a balanced state between a large variation in functional network configuration and a large constraint on structural network configuration, may be the basis of the emergence of a large Φ, a surrogate of integrated information. We also hypothesized that as consciousness diminishes, the brain loses network criticality and Φ decreases. We tested these hypotheses with a large-scale brain network model and high-density electroencephalography (EEG) acquired during various levels of human consciousness under general anesthesia. In the modeling study, maximal criticality coincided with maximal Φ. The EEG study demonstrated an explicit relationship between Φ, criticality, and level of consciousness. The conscious resting state showed the largest Φ and criticality, whereas the balance between variation and constraint in the brain network broke down as the response rate dwindled. The results suggest network criticality as a necessary condition of a large Φ in the human brain.