Understanding how consciousness arises and fluctuates remains a major challenge for neuroscience. The Global Neuronal Workspace Theory (GNWT) proposes that consciousness depends on the global broadcasting of information across large-scale cortical networks. Competing theories, like the Integrated Information Theory (IIT), offer alternative accounts, such as emphasizing intrinsic system properties. Sleep-wake transitions provide a natural model for studying shifts in conscious states. This review synthesizes current evidence on neural activation patterns, functional connectivity, and neurophysiological markers during transitions between wakefulness, non-rapid eye movement (NREM) sleep, and rapid eye movement (REM) sleep. The present review includes six studies yielded by a literature search taking place in the Web of Science and Scopus databases. Findings reveal that wakefulness is characterized by sustained frontoparietal integration, consistent with GNWT predictions, while NREM sleep shows a progressive disconnection of the global workspace, correlating with diminished conscious access. During REM sleep, localized gamma synchrony suggests partially reconnected cortical networks supporting internally generated conscious experiences, such as dreaming. This pattern is consistent with updated interpretations of the GNWT, which allow for localized or graded workspace ignition in the absence of full frontal engagement. Comparing these findings with predictions of the IIT highlights the need for adversarial research paradigms that directly test divergent theoretical claims. Sleep-wake dynamics thus offer critical insights into the mechanisms underlying consciousness and emphasize the value of multi-theory approaches in consciousness science.