Kirill V. Nourski, Mitchell Steinschneider, Ariane E. Rhone, Hiroto Kawasaki, Matthew A. Howard, Matthew I. Banks

The systems-level mechanisms underlying loss of consciousness (LOC) under anesthesia remain unclear. General anesthetics suppress sensory responses within higher-order cortex and feedback connections, both critical elements of predictive coding hypotheses of conscious perception. Responses to auditory novelty may offer promise as biomarkers for consciousness. This study examined anesthesia-induced changes in auditory novelty responses over short ("local deviant", LD) and long ("global deviant", GD) time scales, envisioned to engage pre-attentive and conscious levels of processing, respectively. Electrocorticographic recordings were obtained in human neurosurgical patients (3 male, 3 female) from four hierarchical processing levels: core auditory cortex, non-core auditory cortex, auditory-related and prefrontal cortex (PFC). Stimuli were vowel patterns incorporating deviants within and across stimuli (LD and GD). Subjects were presented with stimuli while awake, and during sedation (responsive) and following LOC (unresponsive) under propofol anesthesia. LD and GD effects were assayed as the averaged evoked potential (AEP) and high gamma (70-150 Hz) activity. In the awake state, LD and GD effects were present in all recorded regions, with AEP effects more broadly distributed than high gamma activity. Under sedation, LD effects were preserved in all regions except PFC. LOC was accompanied by loss of LD effects outside of auditory cortex. By contrast, GD effects were markedly suppressed under sedation in all regions and were absent following LOC. Thus, although the presence of GD effects is indicative of being awake, its absence is not indicative of LOC. Loss of LD effects in higher-order cortical areas may constitute an alternative biomarker of LOC.SIGNIFICANCE STATEMENTDevelopment of a biomarker that indexes changes in the brain upon loss of consciousness (LOC) under general anesthesia has broad implications for elucidating the neural basis of awareness and clinical relevance to mechanisms of sleep, coma and disorders of consciousness. Using intracranial recordings from neurosurgery patients, we investigated changes in the activation of cortical networks involved in auditory novelty detection over short ('local deviance') and long ('global deviance') time scales associated with sedation and LOC under propofol anesthesia. Our results indicate that while the presence of global deviance effects can index awareness, their loss cannot serve as a biomarker for LOC. The dramatic reduction of local deviance effects in areas beyond auditory cortex may constitute an alternative biomarker of LOC.