What is the functional relationship between sleep and anesthesia? We could talk about that in a couple of different ways. Sleep, or your sleep profile, could have an effect on how you respond to anesthesia. And in the other direction, anesthesia could have an effect on your sleep profile. So let's start out thinking about how sleep affects anesthetic sensitivity. This was a study that was performed in rats by Dr. Avery Tung's group at the University of Chicago. And these animals were deprived of sleep for 24 hours. And then they were induced with either the anesthetic propofol or isoflurane. And the measure of that anesthetic induction is the time to the loss of righting reflex. Now the righting reflex basically means that rodents like to be on all fours. They don't like to be on their back. And so if you're giving a rodent an anesthetic and you keep trying to turn it on its back, it's going to keep turning around and getting on all fours. So when it loses the ability to do that, we use that as a marker or a surrogate for unconsciousness. So looking at this graph, you can see in the black bars, these are the animals that were sleep deprived. And you can see the time it took for them to lose that righting reflex is significantly decreased compared to the non-deprived counterparts. Now in the other end of the experience when they're waking up or recovering, the time to recovery is prolonged in both the propofol and the isoflurane groups, even more so, relatively speaking, for isoflurane. So sleep deprivation both sensitizes the animal to the effects of anesthetics and it delays the recovery. This actually might have a genetic basis. So the graph I'm showing you now is not with rodents, but comes from Drosophila, or the fruit fly. And the fruit fly is used in experiments because it can be very easily manipulated from the genetic perspective. So I'm showing you now a dose response curve for isoflurane, for both wild type, that is normal Drosophila, as well as a mutant Drosophila. And the mutant in this case is referred to as shaker minisleep. Shaker because that's the name of the voltage gated potassium channel that's mutated, and minisleep because these Drospohila seem to spend more time in a quiescent state. That is, not moving around, thought to be an analog of sleep in the fruit fly. So let's take a look at this dose response curve. This is the wild type, and this is the shaker mutant. And what you can see is that there's a rightward shift to the curve. That means that these mutants that tend to sleep less, need more anesthesia than their wild type counterparts. And that's really interesting because it suggests, in this model, that a single gene controls sensitivity to anesthesia as well as sleep regulation. So showing an underlying genetic relationship between these two states. So now let's think about it in the other direction. If you're sleep deprived, for example, how does anesthesia affect the brain? Does the brain see anesthesia as a kind of sleep? And in thinking about how anesthetics could affect sleep homeostasis, there are really three general options. First, anesthesia could satisfy sleep debt in the same way that sleep itself does. Number two, anesthesia could, like the waking state, lead to the accrual of sleep debt. Or three, anesthesia could be neutral with respect to sleep debt, that is sleep debt neither gets dissipated nor accrues. So again, Avery Tung's group did several important studies. This one looked at how the anesthetic propofol affects recovery from sleep deprivation. So these animals were deprived from sleep for 24 hours and then they were either allowed to sleep with a control infusion, or they were given an infusion of propofol, an intravenous anesthetic. And then the authors looked to see what happens in terms of rebound. So rebound basically means, if you're deprived of sleep and then you go to sleep, you're going to have more kinds of sleep than you might normally have. So in this case, the rebound profiles of propofol in the control look the same statistically. Now you see some stars there, for example, suggesting a statistical difference. That's really just from the baseline circadian control that was not deprived. But in this study for non-REM sleep, anesthesia and natural sleep looked very similar. The same was true with REM rebound profiles. So again, these were different from the baseline but not really different from one another. So these studies suggested that propofol can satisfy both components of sleep, non-REM and REM, as well as sleep itself. So in figuring out how propofol fits in with these three options, it appears to satisfy sleep debt, very much like recovery sleep itself. So we ask the question, well what about inhaled anesthetics? Propofol is used often to induce anesthesia, but we're usually using anesthetic vapors to maintain anesthesia. And so we conducted two studies. The first study looked at isoflurane and REM sleep deprivation. As I showed you before, non-REM sleep, slow-wave sleep, looks very similar to anesthesia in terms of anesthetic slow waves. This isn't really the case for REM, we don't have clear REM correlates during a general anesthetic. So we wanted to focus in first on REM. And we had four conditions. Condition 1 and 3 was 24 hours of the animal doing what it wanted. And then condition 2 and 4, we specifically or selectively deprived the animal of REM sleep. And what we're looking at, what we're really comparing here, what's really of interest is comparing conditions 2 and 4. We were comparing the effects of isoflurane for those first four hours after deprivation, that's in condition 4, with what's going on for the first four hours of natural sleep after condition 2. And we put an electrode not only on the surface of the brain, but deep to the surface of the brain and the hippocampus. And what you see at the bottom part of the slide is a theta wave. And that's very much characteristic of REM sleep. We wanted to see what was going on with theta in these two conditions. And the animals, which you can see here, they're inhaling isoflurane through a nose cone and then we have EEG recordings coming out of their brain, were titrated to unresponsiveness without burst suppression. And the significance of that will become clear later. So these animals had a slow-wave anesthetic. So what we found is that isoflurane does not satisfy the homeostatic need for REM sleep. So if you look at the top of the slide, that's just comparing normal circadian sleep to REM deprivation. And you can see with the gray bars that the animals have REM and it increases as the time go on. But look at the black bars at the top of this panel. There's a big spike of REM sleep that happens after REM deprivation. That's what's called REM rebound. And then it seems to dissipate in those first four hours, because in the second four hours, hours 4 to 8, there's no difference between the deprived group and the non-deprived group. Now we go down to the bottom part of this figure. You can see, instead of sleeping, we let the animals have four hours of isoflurane. And let's look what happened afterwards. From hours 4 to 8, there's clearly an increase in REM sleep rebound. Now if you look at hours 4 to 8 at the top part of the slide, you see there's no difference. So what's really happened here is that there's been almost a circadian shift. The animals, during the anesthetic, don't have a discharge of REM sleep debt, the brain seems to have a memory for how much REM sleep it still needs. And when they come out of the anesthetic, you see a rebound. And that circadian shift is interesting, especially with respect to a very clever study performed and published last year in bees, showing that general anesthesia results in a phase shift in their circadian clock. It's almost as if it induces jet lag. So we also saw that there was no effect of REM deprivation on hippocampal theta activity during isoflurane. So whatever's driving theta activity during sleep, it seems to be different compared to that which is driving theta during isoflurane anesthesia. So in thinking about the options here, what we found for isoflurane in REM sleep, is that it appears to be neutral with respect to sleep debt. REM sleep debt didn't seem to accrue, but it also didn't seem to get discharged during that anesthetic. So that was the first study. Now we wanted to see, well, is this true for another anesthetic? And what if we have total sleep deprivation? So in this set of experiments, we had 36 hours recording which you could see on the left part of the slide, both in the light phase and during the dark phase. The first condition was the animals just got to sleep or awake as they wanted to in ad libitum fashion. With condition 2, we deprived them of sleep for 12 hours during the light phase and then we let them recover however they wanted. And then finally, in condition 3 they were deprived, but now instead of sleep, you can see that they had sevoflurane for six hours. And that was followed by ad-lib recovery sleep. Now, in this set of experiments, sevoflurane was titrated to 50% burst suppression. And you can see on that recording screen, that there are periods in which the EEG is flat and periods in which there's a lot of activity going on, or a burst of activity. And that's referred to as burst suppression. And the reason why we chose burst suppression is because burst suppression is something that is specific to anesthesia. You never see burst suppression during normal sleep. That's important because we wanted to make sure, with some of these old experiments conducted by Avery Tung, as well as some of our own experiments, that when we were seeing slow wave activity on the EEG, that it was really anesthesia going on and that the animal hadn't really just been sleeping. So by seeing burst suppression, we know the animal is under anesthesia rather than sleep. So what we're doing here is we're comparing that first period of ad-lib sleep to six hours of sevoflurane. So let's look at what happens with non-REM sleep. And we're looking at a rebound phenomena just like the past study with propofol. So leading up to the graph that you're seeing has been 12 hours of total sleep deprivation. Then the animal goes into the first six hour block where it's allowed to either sleep in recovery, or it's getting anesthesia. Now you can see the comparison between the red bar and the blue bar in that first six hour block. The blue bar is significantly higher. That's the non-REM rebound, and that continues into the next block. But look at what happens in the animals that got the anesthesia instead. When they get to that second hour block, the green bar, which is the sevoflurane group, you see those pound signs above that green bar. That's showing a difference from the sleep deprivation group, but there's actually no difference between the ad-lib group. So what I'm telling you is, for natural sleep recovery, the animal needed a full 12 hours to recover that non-REM sleep. But it just needed six hours to have complete recovery of non-REM sleep during an anesthetic. So by the time the animal gets out of the anesthetic, it has no more non-REM rebound. And you can see on the right part of this figure, for the last 12 hours, there's really no rebound in any of the groups. This was also demonstrated by a decrease in delta power. And this is an important marker for an animal that's been sleep deprived. So look at the first six-hour block, 6 PM to 12 AM. You see the red bar, which is a normal delta power, you see the blue bar, there's a significant increase in that first block of the animal who's been sleep deprived. And then we have the other group that's getting the sevoflurane. When we move to the next blocks, we see that the blue bar is decreased compared to the ad-lib group. That's what's called a negative rebound, and that's been reported in the literature. But in the sevoflurane group, it's never above the control condition. And in fact, it's the lowest of all conditions, suggesting that the animal had really satisfied its need for that delta wave or slow-wave sleep. Now the situation is different for REM sleep, sevoflurane does not satisfy the need for REM sleep. You can see, in the first six-hour block, the blue bar is significantly elevated above control. That's showing a REM rebound and that extends to the next six-hour block, and then it dissipates. So the animal needs 12 hours of recovery sleep to satisfy the need for REM. Well look what happens with the sevoflurane group. If you go to the block of 12 AM to 6 AM, the first part, you see there's still an increase in the blue bar, that is the sleep deprivation. But there's also a significant increase in the sevoflurane group. So not only is it different from the sleep deprived group, but it's elevated above the normal control. And it keeps being elevated. So if you go now to the 6 AM to 12 PM block, you see that the sevoflurane group is still increased. It's significantly elevated now above both the control condition in red and the sleep deprived group in blue. That means the animals still needed 12 hours of REM sleep when they were coming out of the anesthetic, in a way that's quite similar to what we found with isoflurane. Now there's some converging findings from other investigators. We know that isoflurane and desflurane satisfy slow-wave sleep homeostasis after sleep deprivation. And we also know that sevoflurane, desflurane can result in a selective REM sleep rebound. So what we're seeing here is a confirmation that these inhaled anesthetics, isoflurane, desflurane, sevoflurane, tend to satisfy slow-wave sleep but they don't satisfy REM sleep. So the picture we got from propofol was a very balanced picture, in which anesthesia could affect sleep homeostasis and sleep could affect anesthetic sensitivity. But what we're seeing for these inhaled anesthetics, in particular halogenated ethers, isoflurane, sevoflurane, desflurane, anesthesia seems to have a profound effect on non-REM sleep. But the REM sleep debt is not recovered, it's not discharged. So let's think about this clinically, what this could mean. I think we normally think of a patient coming in in a certain brain state, where they can get either propofol or sevoflurane, and then they're coming out back to normal with respect to sleep homeostasis. But what we can infer from these animal studies that I've shown you, is that if a patient came in who is in a sleep deprived state and they got propofol, they might actually be recovered and restored from the perspective of both non-REM and REM sleep. The same patient just getting sevoflurane might actually come out still in a REM deprived state. And that's important for patients who have sleep apnea, for example. When they go into REM sleep where we have paralysis, they can lose muscle tone, they have an increased propensity to having airway closure, and they can actually get hypoxic. Now, this is just the picture with anesthesia. Of course, we're not taking into consideration the fact that a patient, unlike these animals, is having surgery. They're getting pain medicines, they might have pain. So a lot of other things, such as circulating stress hormones, that can affect sleep, but this slide summarizes what could be happening, purely from the perspective of anesthesia.