For years, sleep researchers have characterized physiological sleepiness observed by EEG through one measure: slow wave activity. Slow wave activity is fairly 
straightforward to identify and represents EEG power at slower frequencies that essentially look like a roadside view of the Rocky Mountain Range. The amount of slow wave activity is directly proportional to the length of time that an animal is awake meaning that slow wave activity dominates the EEG after long periods of wake or sleep deprivation. While slow wave activity has helped determine how a host of environmental, physiological, and genetic factors influence the ability to recover from sleep loss, it is limiting because it does not accurately quantify how an animal feels during sleep loss. And so, our lab teamed up with another electrophysiology lab that studies epilepsy to provide a means to quantify physiological sleepiness as it occurs, not after.
This was achieved through a period-amplitude analysis which looks at an individual EEG within a specific time frame, say 10 seconds, rather than all EEG waveforms within a specific time frame. With this period-amplitude analysis, the number of slow wave peaks were counted across sleep loss with the mice being deprived of sleep for up to 24 hours. Early into the sleep deprivation, there were slow wave peaks here and there.
But as time passed, particularly after the animal had been deprived of sleep for 6 hours, slow wave peaks began to dominate the EEG and persists until the end of 24 hours of sleep deprivation.
We also found that the number of slow wave peaks was negatively correlated to the time it took the animal to fall asleep, in particular NREM sleep, which makes up 85-90% of total sleep and is the first type of sleep that an animal has unless the animal is narcoleptic. This relationship between slow wave peaks and time to NREM sleep corroborates the previously identified negative relationship between slow wave activity and time to NREM sleep.
Finally, we found a time-dependent difference in the quantity of slow wave peaks based on whether the animal got sufficient sleep or was recovering from sleep loss. With sufficient sleep, slow wave peaks were more common during the light or rest/sleeping period of a nocturnal rodent. There were also more slow wave peaks during the middle of the night when most nocturnal rodents nap. This distinctive rhythm in slow wave peaks was absent in a mouse recovering from sleep loss. This is largely because there were more slow wave peaks during the night than that present with sufficient sleep.
This study provides a means to better characterize changes in sleep and wake with inadvertent or voluntary sleep deprivation.
Ehlen JC, Jefferson F, Brager AJ, Benveniste M, & Paul KN (2013). Period-Amplitude Analysis Reveals Wake-Dependent Changes in the Electroencephalogram during Sleep Deprivation. Sleep, 36 (11), 1723-35 PMID: 24179307
