Scientists have been studying the nature of human consciousness for 170 years, but still haven't fully explained how it works. A new study used fMRI technology to scan the brains of 17 healthy adults who were gradually sedated through four levels of consciousness, including wakefulness, light sedation, deep sedation and recovery. The researchers found that anesthesia doesn't act as an on/off switch but alters oscillatory modes across the brain. This led them to create a machine-learning algorithm that correctly identified a person’s conscious state 72 percent of the time. The study also revealed that as anesthesia takes hold, sounds are still detected in the primary auditory cortex, but these signals are not moving into higher-order regions of the brain, preventing the brain from progressing further. The findings could potentially lead to novel diagnostic and monitoring tools in this field.