During sleep, functional connections between brain networks shift in a directional manner. In non-REM (NREM) sleep, positive wake-like correlations often become negative and strengthen in the opposite direction, while in REM sleep they trend back toward positive correlations. This pattern supports the idea that NREM sleep involves altered, not merely reduced, functional connectivity. Many of these connections involve higher-order networks linked to cognition and consciousness, such as the default mode network, suggesting possible accompanying changes in cognitive and conscious states.
Propofol, the most common general anesthetic, induces loss of consciousness through mechanisms that remain poorly understood. Using the generalized Ising model (GIM) to analyze fMRI data from healthy subjects listening to an audio clip, brain activity was modeled during wakefulness, mild sedation, deep sedation, and recovery. A novel inter-subject correlation technique captured common synchronization across participants. The GIM, modified to incorporate the naturalistic external stimulus, successfully fitted empirical task fMRI data at a temperature well above the critical temperature. This is the first mathematical modeling of human brain activity responding to real-life stimuli at different conscious levels, potentially aiding assessment of consciousness in patients with disorders of consciousness.