Slow waves and sleep spindles are the two fundamental brain oscillations of NREM sleep, yet they have been mostly studied in vitro, under anesthesia, within few brain regions or with scalp EEG recordings. We examined intracranial depth EEG and single-unit activity recorded simultaneously in up to 12 brain regions in neurosurgical patients to better characterize regional diversity in these sleep oscillations. First, we found changes in spindle occurrence, frequency, and timing between regions and across sleep, reflecting anatomical projections and thalamocortical hyperpolarization levels that change with sleep depth. We further show that both slow waves (and the underlying active and silent neuronal states) and sleep spindles occur mostly locally, thereby showing that constrained intracerebral communication is an important feature of sleep. Next, we confirmed that in freely behaving rats, slow waves and silent periods in sleep likewise occur predominantly locally. Moreover, after a long period of being awake, while both EEG and behavior indicate wakefulness, local populations of neurons go offline, exhibiting "local sleep". We are now exploring whether such local sleep may lead to cognitive consequences, such as lapses of attention, in awake people who are sleep deprived
Another line of research focuses on disconnection from the external environment - conditions in which sensory stimuli fail to be incorporated into our perceptual stream. To this end, we are examining neuronal responses to sounds in rats across spontaneous vigilance states with an emphasis on comparing wakefulness with REM sleep. Responses of individual neurons in primary auditory cortex are comparable in wake and sleep, calling into question the proposal that the thalamus does not relay peripheral signals effectively to the cortex in sleep. Important differences between waking and sleep may lie in how signals propagate across cortical regions and layers.
