Circadian rhythms and epilepsy Part I: Dr. Mark Quigg

Listen or read: Circadian rhythms and epilepsy, Part II: Dr. Maxime Baud

Reported by Laurent Sheybani, PhD | Edited and produced by Nancy Volkers

---

Cite this article: Sheybani L. Circadian rhythms and epilepsy Part I: Dr. Mark Quigg. Epigraph 2023; 25(3): 1-7.

---

Listen below or download the episode.

Find Sharp Waves episodes on Spotify, Apple Podcasts, Google Podcasts, Amazon Music, iHeart Radio, and Stitcher.

Though the cyclic properties of seizures have been known for more than 100 years, much is still unknown. Does everyone with epilepsy have a seizure pattern? What can be learned from seizure diaries, RNS data, and animal models, and how can this information guide further research and clinical care? In the first of a two-part series on circadian rhythms, Dr. Laurent Sheybani talks with Dr. Mark Quigg.

Sharp Waves episodes are meant for informational purposes only, and not as clinical or medical advice.

Podcast Transcript

[00:00:00] Dr. Laurent Sheybani: Hi everyone. Today's podcast focuses on circadian rhythms in epilepsy. I'm interviewing Mark Quigg, professor of neurology from the University of Virginia, who is one of the leading scientists in the field of chronobiology of epilepsy. 

In the recent years, there has been a significant and growing interest in chronobiology of epilepsy. In a following episode, I will discuss this topic also with Professor Maxime Baud from Bern University, who has published groundbreaking findings in this topic in the last few years. Enjoy the podcast and don't forget to visit our website to check for other interviews.

So, Dr. Quigg, thank you for discussing with me today about circadian rhythms in epilepsy. To begin with, may I ask you to introduce yourself? 

[00:00:50] Dr. Mark Quigg: My name's Mark Quigg. I am the T.R. Johns Professor of Neurology at the University of Virginia, which is in Charlottesville, Virginia, USA.

My clinical practice is in epileptology and sleep. And I have kind of a broad-based research career, but it's centered for the purposes of today's talk in the chronobiology of epilepsy. 

[00:01:23] Dr. Laurent Sheybani: So as a specialist in epilepsy, sleep, and circadian rhythms, could you explain what are circadian and multi, multi-day rhythms in epilepsy and why is that important?

[00:01:34] Dr. Mark Quigg: Okay. Well, I think if we take a step back and just propose a very broad hypothesis that seizure occurrences, whether electrographic or clinical, are events that occur not randomly, but as an emergent property of all the endogenous and exogenous influences that affect a patient. And those exogenous influences can be facilitatory or inhibitory, as well as the endogenous rhythms.

And for the purposes of today's questions, those influences can be cyclic. So just as an EEG can have a variety of rhythms on one page — you have a slow wave and there's a medium frequency on top and a very fast frequency on top of that, and that combination of very simple rhythms creates a very complex picture once all the nodes and nadirs add up. So seizures and the rhythms that they have are the emergent property of the summation of all those different rhythms that affect the patient. 

So the rhythms are, are we'll use circadian as a beginning. "Circa dian" meaning about one day. So that period is, is roughly 24 hours. Rhythms that are faster than circadian are ultradian, faster than a day. And ones that are slower than a day are infradian.

So that's kind of the range of those things. A great example of an ultradian rhythm that has important properties for sleep regulation is the alternation between REM and non- REM sleep. And examples of an infradian slow rhythm would be, for example, the lunar cycle or catamenial seizures to the monthly menstrual cycle.

You can imagine that a patient with medically intractable epilepsy perhaps is under the influence of very many additive cycles. And these, the cyclic property of seizures and electrographic discharges and so forth, has been known for a long time, 150 years. The difference is that we've gotten better and better at the math and at ways of isolating the different forces, the different cycles.

That's kind of the treetop overview. 

[00:04:28] Dr. Laurent Sheybani: Could you give an example of one of these endogenous factors that can affect the risk of epilepsy? 

[00:04:34] Dr. Mark Quigg: Well, I think the sleep-wake cycle is an important aspect of that. The sleep cycle, how we all tend to feel sleepy and fall asleep at the same time every day, is a property regulated by the circadian timing system. Sleep is facilitated or inhibited according to the internal circadian clock. So similarly, seizures can be facilitated or inhibited by the underlying sleep-wake state. The complicating factor that I should have mentioned at the beginning of the treetop view is that particular epilepsy syndromes may be more or less susceptible to different kinds of rhythms or exogenous facilitators or inhibitors.

So the syndromes are not equal in their susceptibility. A good example of an epilepsy that is nicely tied to the sleep-wake cycle is classically frontal lobe epilepsy, or perhaps more broadly, depending upon who you read, non-temporal-lobe epilepsy, cortical epilepsy. 

And you know, a classic patient who typically has their seizures only during sleep, well, one can think, “Hmm, maybe that is a frontal lobe or at least a non-temporal corticoepilepsy.”

[00:06:05] Dr. Laurent Sheybani: Right. So this leads to the next question. Do all patients have such fluctuations of epileptic seizures? 

[00:06:13] Dr. Mark Quigg: I'm going to have to fall back on the original observation by Gowers in 1880-something, who divided his observations of patients into three groups: day seizures, night seizures, and random.

What we're finding though is that the random group probably is also susceptible to a cyclic expression. It's just a matter of observing a single patient long enough so that we can collect enough seizures to discern a pattern, or having groups of homogeneous patients that can provide enough seizures for a long enough time to observe a pattern.

This is the big issue with this particular business; over the years, technology has gotten better and better at helping us register the number of seizures that are necessary to observe patterns. So I think it could be safely said that the better we get at measuring, the less of an incidence is this random, truly random group.

[00:07:25] Dr. Laurent Sheybani: How can we identify that patients have fluctuations of seizure risk? 

[00:07:32] Dr. Mark Quigg: Well, early in my career, I collected seizure diaries from patients and did lots of seizure diary analysis. Seizure diaries are the gold standard of epilepsy drug trials. They can be very unreliable when it comes to seizures that are confined to sleep. But the point is that it's kind of a built-in error that we have accepted. And so if you have long enough seizure diaries, that's a great way to determine from a patient whether or not their seizures are happening at a typical time of day.

So, a good bedside question I always ask is do they notice a particular time of day in which most of their seizures occur? And the typical answer I get is, “Yeah, lots of them occur during sleep,” or “Most of them are during the day.” Or they're all over the place. So going back to Gower's original observation. 

[00:08:29] Dr. Laurent Sheybani: And besides human beings, these rhythms, have they been observed in other mammals or other animal model of epilepsy? 

[00:08:37] Dr. Mark Quigg: Yes. Animal models of epilepsy are a great way to kind of break that limitation of humans, you know, trying to get enough homogeneous patients to make some conclusion about their rhythms or have enough seizures to discern a pattern.

Because you can observe tons of animals for very long periods of time, and with the proper EEG equipment determine that seizures are in fact seizures. So, for example, early in my career, I studied a rat model of limbic epilepsy, the self-sustaining limbic status epilepticus model developed by Ed Bertram and Eric Lothman. And I used their model in a rat lab that was basically a circadian lab, a chronobiology lab in which we could regulate light exposure and introduce random feeding so that there are no external time cues. Close it up, turn off the lights, and observe seizures occurring in a free running pattern, which is the fundamental property of circadian.

So maybe we should state what the official definition of circadian is. Circadian is a free running, approximately 24-hour cycle, uninfluenced by external forces. So it has to be self-sustaining in an environment that does not provide a timing cue. There are a variety of human experiments in which they use, like, old German bunkers for isolating patients in the dark, or patients or normal healthy subjects just to observe what the circadian clock was in humans.

So observe patients in a free running environment. Doing that for humans is incredibly difficult. Matter of fact, there's been only one published study that I know of for human free running seizure observation that was done by a woman named Pavlova in the early two thousands, late nineties, in which she was able to use a forced desynchrony protocol in observing the appearance of the ictal interictal epileptiform discharges in patients with JME and proved that they were truly circadian. 

So the beauty of animals is that you can show that these are truly self-sustaining 24-hour rhythms. The critical experiment I did was compare the free running seizures of this limbic epilepsy rat model to humans with temporal lobe epilepsy recorded in an epilepsy unit and showed that both these rhythms occurred in phase. So in other words, patients with mesial temporal lobe epilepsy have a peak seizure occurrence in a normal epilepsy unit environment with seizures occurring mainly between 12 and 6 at night. Rat models in a free running environment maintain that circadian time identically.

And this is despite the fact that rats are nocturnal creatures and sleep during the day and are awake at night. And humans, well, most of our humans in the epilepsy unit are awake during the day, and sleep at night. So their sleep-wake is out of phase 180 degrees, but their seizure occurrence is in phase. Proving that at least in the case of temporal lobe epilepsy, circadian forces are the predominant influence on seizure expression in a circadian fashion that overwhelms sleep wake. 

[00:12:48] Dr. Laurent Sheybani: How would you make the difference between the impact of the circadian rhythm and the homeostatic sleep pressure?

[00:12:57] Dr. Mark Quigg: Yeah. They are truly different cycles. The problem is, is that when you look at studies that tried to look at sleep deprivation as an independent influence, it's very hard to do, right? Because you need to figure out how to isolate sleep and wake from the equation. Sleep deprivation is separate from sleep-wake cycle, right?

So you need to separate sleep deprivation from the sleep-wake cycle and from the underlying circadian rhythm. Maybe that's an animal experiment that I don't think has been done, but in humans, attempts to prove that sleep deprivation is something we use in the epilepsy unit all the time. Whether that's effective or not. Sleep pressure or sleep homeostasis really hasn't come to fruition. That's a long way of saying I'm not quite sure.

[00:13:52] Dr. Laurent Sheybani: That's true. It's a procedure that we very often use sleep deprivation to increase the risk of seizure and seems to work very well in some patients. While in other patients, it seemed to have no effect. 

[00:14:03] Dr. Mark Quigg: So for example, Beth Malow a long time ago tried to show, I believe this was a letter in Neurology, so it wasn't even a full-blown study, demonstrated that in the epilepsy unit in people with focal epilepsy, sleep deprivation really didn't seem to have an overwhelming effect, but clinically it's been well known that juvenile myoclonic epilepsy is very susceptible to sleep deprivation as a seizure trigger. Formalizing that observation to my knowledge hasn't quite been done yet, but it's a very common thing. So you're right. Maybe that's why we haven't proven it well, because we haven't figured out the epilepsy syndrome to best test it by. 

This is where animal research could really provide a great answer because, you know, the different mouse or rat models correspond to the epilepsies differently, and that would be the best way to do it.

[00:15:02] Dr. Laurent Sheybani: So if we are talking about subgroups, do you think, or do you know if there are subgroups of patient based on the localization of the focus or the duration of the disease or the gender that are most susceptible to having rhythms of epileptic seizures? 

[00:15:19] Dr. Mark Quigg: So if you look at, for example, catamenial, seizures are, it's a rhythm, it's lunar, meaning that the, the rhythm is roughly 28 days corresponding to the menstrual cycle. So it is a cycle. We think we know what is causing it, right? The menstrual cycle. 

Through the 1990s and, and through the early 2000s, people were still debating whether or not there's such thing as catamenial epilepsy. And I think it wasn't until Andy Herzog really sat down and categorized the different phases of the menstrual cycle in which different women may have susceptibility to epilepsy and discovered that there were different phases, and kind of cleared up that noise problem and demonstrated catamenial epilepsy very effectively. 

The other thing that you can use is these things all depend upon a reliable timing marker, a phase marker.

So if you use onset of menstrual bleeding as a phase marker and you stretch each cycle to a common period. So if some woman has 25 days, the other one has 31, and you adjust everything, so it's 28 days, those patterns are very, very robust. 

The ones that we can't explain are the infradian or multidien patterns that are revealed by implanted ECoG systems now. So this is kind of where the technology has led us. I told you that this business has been advanced by just better being able to collect the seizures in a reliable fashion.

The RNS system is basically an implantable ECoG, and you can observe fast circadian and slow rhythms in the various patterns detected by the RNS, and the Australian group with their implanted systems has been very productive in demonstrating seizure clustering, seizure periodicity, again, to a variety of rhythms, and explaining these rhythms is the next step.

I think that's the hard step to do, and also as a community deciding what is an accurate observation and what is not is going to need to be worked out. So for example, Maxime Baud's group and our group too, working with David Spencer from Oregon, we looked at RNS data to determine seizure cyclicity.

And one has to realize the RNS is not an ECoG recorder. I mean, it records ECoGs, but it's not designed to be an EEG system, right? It is designed to administer electrical discharges as detected by a human written program. So you can't necessarily use the detections as a surrogate marker for seizures, because the detections could be inaccurate, right? They're just zapping something that isn't there. What you have to do to be rigorous about it is to take that one step further and look at the actual ECoGs recorded with the detections. 

So we got a bunch of EEG readers, experts, and we kind of rated over 7,000 recordings as to whether or not they were actual seizures or not. Now, I didn't read 7,000 ECoGs. We divided it up into like, six different pairs and, we all read a section of these, but we demonstrated decent inter-rater reliability, and then we were able to provide a weighting or a reliability index to each one of these ECoGs.

And then in our subsequent paper, we then said, “Oh, let's use the high rated ones and determine the circadian and infradian or multidien patterns from those reliable recordings.” Both our work and the Baud group have shown—I think they use detections, but I think it largely agrees with the more in-depth ECoG view which is that patients will have rhythms. Many patients will have rhythms of more than 24 hours, but they're at odd, odd frequencies that don't really make sense. 17 days, 15 days 12 days. And to be honest, I can't remember if there is a distribution curve. I don't know where the maximum expression is in the slower rhythms, but I think the issue is trying to think through what we know biologically what could be the endogenous or the exogenous force that's occurring in these kind of odd, slow time signatures. I haven't really heard a valid hypothesis as to why that is. 

So you know, I think from a pathophysiological standpoint I think we really have to think hard now, like, okay, we have the ability to start looking at long recordings from individuals and try to match that up with biology. That's the promise ahead, but I don't have the answer as to which brain mechanism can account for a multidien or infradian expression. Notice that I'm using the terms multidien and infradian as synonyms. I don't really think they're different. I think it's just people using different words for the same thing. 

[00:21:36] Dr. Laurent Sheybani: And in line with this, are you aware of other biological fluctuations, other biological rhythms? 

[00:21:45] Dr. Mark Quigg: There are seasonal rhythms. So Marcus Ng from Winnipeg is studying cases of status epilepticus that are occurring in the far north of Canada. Way, way up there in these little isolated villages, and he's shown that the number of cases of status epilepticus that are flown in from these little villages down to the hospitals have a late spring phase, I may be misquoting him here, so forgive me, Marcus. But I'm going to say April, May, June. That quarter seems to be a strong rhythm that is apparent on a seasonal basis as to why that's happening. We don't know, but it could be as something as simple as enhanced light exposure above the Arctic Circle transitioning from something in which there's light and dark to when there's continuous or near-continuous light.

[00:22:53] Dr. Laurent Sheybani: So it could be also linked to a decreased sleep duration. 

[00:22:56] Dr. Mark Quigg: Exactly. Exactly. So it could be something as simple as that. A seasonal rhythm messing around with sleep homeostasis. 

One of the challenges of chronobiology is isolating the particular influence sufficiently for you to be able to say that that is it. So that's where we're at right now. We're getting much better at measuring we still haven't been great at explaining. 

[00:23:26] Dr. Laurent Sheybani: In terms of patient management, how do you think that this research can help?

[00:23:32] Dr. Mark Quigg: Well, I think by being aware that many patients, if not the majority, have a cyclic occurrence of their seizures, is to use some of that knowledge in identifying their epilepsy syndrome and helping to categorize which one. So we just, in this talk, we've said that temporal lobe epilepsy is a day problem for most. Frontal lobe epilepsy is a night problem. You can use that clinical information to help identify localization. 

From a safety standpoint, well, I mean, it's impractical to have implanted electrocorticography in all patients, but, you know, if we did, or if there are other ways to provide some sort of biological monitoring, identifying these multiplicity of rhythms can certainly lead up to the point in which you can provide prediction algorithms for individual patients. Perhaps say, between September 3rd and 7th, you have a 50% chance of having a seizure, but between the 7th and the 19th, you have a 10% chance of having a seizure.

So we haven't gotten to that point yet, but for example, the folks with Seizure Tracker, that app, which is basically a fancy phone-based calendar, they've done lots of population-based studies to kind of come up with these predictive algorithms for patients. So I think maybe, you know, as long as patients are reliable in reporting seizure diaries, maybe that can be something to provide patients kind of these windows of opportunity, if you will, or windows of risk.

Again, the key is individual patients observed long enough, having enough seizures.

So paradoxically, all these great observations were better, the sicker the patient is, right? So these risk window prediction things kind of are paradoxical because what we really want is to give a person with rare or breakthrough seizures ample warning. I think the patients who can benefit the most from this are paradoxically the ones who we have the hardest time coming up with a pattern for. So those are I think two practical issues. And the third I'll mention is what we haven't talked about, which is chronopharmacology, which I think is still a relatively understudied phenomenon in that by manipulating the dosing pattern of an anti-seizure medication, you can perhaps best meet the patient's needs as far as avoiding toxicity or having more drug aboard when they're most susceptible.

So this has been done from time to time, especially with some of the older medications. Dilantin, Tegretol, there's some older studies back there. I'll step out on a limb and say relatively safely that most of these liver-based metabolism of anti-seizure medications are accelerated at night and inhibited during the day.

So lots of these older studies have shown that if you can take a b.i.d. dosing schedule and weight it like 25% or 30% to the night dose, you can help better seizure control and avoid daytime toxicity. So, for example it's common with me using Lamictal, if someone's on that cusp of toxicity of 400 milligrams a day, it's common for me to a dose at 150 in the morning and 250 at night. Especially if they have a very robust pattern of early day or nighttime seizures. Those are kind of three practical areas. that chronobiology can help in patient management. 

[00:27:53] Dr. Laurent Sheybani: So adapting treatment in is one of the benefits of this research. And I expect that all patients, or most patients included in these studies were taking medications while we have identified their rhythms.

[00:28:07] Dr. Mark Quigg: You've gotten to the key point of the thing: All these observations are taken from patients taking medications, and the animal research can really help us. I mentioned quite early on my observation of temporal lobe epilepsy being a daytime phenomenon. This was an epilepsy unit, and I couldn't account for the fact that some of the seizures were occurring early on when they still had medications, versus late seizures that may be in the setting of no anti-convulsant medications because they've been tapered off as part of their clinical evaluation. And there's no way for me to account for that in a robust fashion. And that that is the same susceptibility that all these studies and humans have had. So it's still hard for me to explain seasonal, or every 17 days, or things due to pharmacokinetics.

[00:29:09] Dr. Laurent Sheybani: And do we know if other factors can influence these rhythms? 

[00:29:14] Dr. Mark Quigg: Well I'll give an example of lunar rhythms. This is something that we know something about. So it's been observed for a very long time, especially in emergency rooms. The suspicious nurses would say, “Oh, it's a full moon out. Bad things are going to happen.”

And turns out it's true that if you look at population-based admissions to emergency rooms for status epilepticus or severe seizures, that they do occur during the full moon. And this has been replicated by at least three publications that I can think of off the top of my head. Why that happens, I think was best revealed by, I believe it was a German study that went back and looked at the full moons that were obscured by bad weather. They demonstrated that the nights in which the full moon was obscured, seizure occurrence was normal baseline. 

So that's one example where an exogenous force, there's nothing mysterious, there's no weird gravity or tides or anything like that. It's just more or less light exposure, keeping people up or sleep depriving them. So that, that is a clear example of an exogenous rhythmic force resulting in a biologic rhythm, but it's basically keeping the lights on at night. 

SELECTED PUBLICATIONS

Electrocorticographic events from long-term ambulatory brain recordings can potentially supplement seizure diaries - Quigg M, et al., Epilepsy Research (2020)

Circadian and ultradian patterns of epileptiform discharges differ by seizure-onset location during long-term ambulatory intracranial monitoring - Spencer DC, et al., Epilepsia (2016)

Interrater reliability in interpretation of electrocorticographic seizure detections of the responsive neurostimulator - Quigg M, et al., Epilepsia (2015)

Variation of seizure frequency with ovulatory status of menstrual cycles - Herzog AG, et al., Epilepsia (2011)

Is there a circadian variation of epileptiform abnormalities in idiopathic generalized epilepsy? Pavlova MK, et al., Epilepsy & Behavior (2009)