Speaker A
Thanks very much, and I hope you're all having a good day. I'm on annual leave at the moment; it's half term here, so if you hear any noise, it's either my teenage son playing the music too loud
An in-depth session on processed EEG monitoring in anesthesia, covering basics, brain states, power spectra, and clinical applications.
Key Takeaways
- Processed EEG provides valuable insights into anesthetic depth and brain function.
- EEG frequency and amplitude patterns change predictably with anesthesia induction.
- Power spectrum analysis is critical for understanding EEG components and brain states.
- Not all EEG monitors display qualitative data equally; some hide important details.
- Understanding EEG changes helps improve anesthesia management and patient safety.
Summary
- Introduction to processed EEG and its relevance in anesthesia monitoring.
- Explanation of EEG as a stochastic signal with multiple frequency bands (delta, theta, alpha, beta).
- Description of EEG amplitude and frequency changes from awake to anesthetized states.
- Comparison of EEG signals in young versus older brains and implications for monitoring.
- Use of power spectrum and Fourier transform to analyze EEG frequency distribution.
- Discussion of spectral edge frequency and median frequency as EEG parameters.
- Demonstration of changes in EEG during induction with propofol and other anesthetics.
- Introduction to density spectral array and its use in visualizing brain activity.
- Challenges and limitations of processed EEG indices and monitoring devices.
- Clinical relevance of EEG monitoring for brain metabolism, maturation, and anesthetic depth.
Chapters
- 00:00Introduction and basics of processed EEG
- 04:02EEG changes during anesthesia induction
- 08:03Power spectrum and spectral edge frequency
- 12:18Qualitative EEG differences in brain age
- 16:55Density spectral array and advanced EEG visualization
- 22:02Clinical implications and EEG interpretation challenges
- 26:12EEG patterns with different anesthetic agents
- 31:24Summary and practical applications in anesthesia
Full Transcript — Download SRT & Markdown
Speaker A
or the youngest one nagging for more time on his iPad. So I do excuse any gate crashing from the kids. So what we're going to do is have a reasonably in-depth session looking at processed EEG monitoring. However, I
Speaker A
recognize that not everybody's up to speed with some of the basics of processed EEG, so we'll whiz through some basics to begin with before looking at the more interesting stuff and understanding how processed EEG is a monitor of anesthetic mechanisms, brain
Speaker A
maturation, and brain metabolism, which is the really exciting stuff. Okay, so just to bring people up to speed, what we're talking about when we talk about the EEG is that the EEG is a stochastic signal. That means it's
Speaker A
not repeating, and it is made up of multiple different frequencies. Each of the electrodes that's put onto the head in an EEG is picking up the signal from about 50,000 to 150,000 individual nerve cells, which
Speaker A
are oscillating and producing electrical signals at various different frequencies. We use the Greek taxonomy to describe those frequencies, which range from very slow frequency signals through to delta, theta, alpha, and beta at increasingly frequent, increasing
Speaker A
frequencies. These frequencies actually go all the way up to about 600 Hertz, but what we're interested in typically in anesthesia is between 0.5 and 30 Hertz. And when we look at the various different waveforms, we've got
Speaker A
these very slow frequencies moving towards these much faster frequencies. These faster ones, you'll notice, are very much lower amplitude. The height, the vertical deflection, is much lower as frequencies get higher. We talk about the frequency distribution of the EEG,
Speaker A
but also the power, and the power is how much of a particular frequency there is at any particular time. Now, when you're looking at the screen on a processed EEG monitor, important things to be mindful of are the time duration,
Speaker A
which is along the x-axis, and the amplitude, which is a vertical deflection around the zero point on the y-axis. If you're looking at a standard BIS Vista monitor, it's 4.4 seconds for the EEG to traverse the screen, and the amplitude is
Speaker A
plus or minus 50 microvolts when you look at a single channel. And really on BIS, you must always look at a single channel because it is a single-channel monitor. Sometimes it will show you two channels, but actually it's only
Speaker A
analyzing one in the first place, so it's much better to use it in single-channel mode. Now, one of the things that first struck me when I got interested in processed EEG, which is probably around 16 to 18 years ago, is being able to
Speaker A
identify differences in different brains. This is a nice illustration of that. On the left-hand side here, we've got a young brain, very high amplitude EEG. That's a 33-year-old just having an OGD, and on the right-hand side, an older brain.
Speaker A
We can see a marked difference between these two brains, with very much smaller amplitude oscillations. Now, unfortunately, the way some of our monitors work is to hide that information from us, and if you were to look just at index values, we'd lose
Speaker A
very much of the qualitative information that tells us so much about this very rich signal that the EEG provides. Now, when we're inducing anesthesia, we see some very significant changes in the EEG as we move from the wakeful brain
Speaker A
towards the anesthetized brain. Up here on the top left-hand side, we can see the awake brain. This is low amplitude, high frequency, and we'll often see blink artifacts and a lot of EMG signal overriding that. As we start to give a little bit of propofol,
Speaker A
this is just a minute after administering a bolus of propofol. We're seeing that the amplitude, the height of that EEG, has more than doubled, and the frequency is starting to slow down. These are the first signs that the EEG and the brain
Speaker A
are starting to be affected by our anesthetic medications. Shortly after, 15 seconds later, we're starting to see some vertical undulations or slow delta oscillations, and then these become very, very pronounced just a few seconds later on. At this point,
Speaker A
this is the point where the brainstem has really started to become anesthetized, would have lost cranial nerve function, and our patient would be profoundly unrousable. Then we settle into this anesthetized brain state with slow undulations with superimposed alpha oscillations, and if
Speaker A
we push things a little bit too far, we get an ECG. You can see here a reproducible pattern on this particular trace. Anything reproducible is cardiac generally, and here we can see the heartbeat once a second, so that's giving
Speaker A
us a frequency of 1 Hertz. That's the heart beating nicely at 60 per minute. So certainly the ECG, which is obviously a microvolt signal, can be picked up on the EEG, which is a microvolt signal. Now, let's think about some of the
Speaker A
other ways that we can look at the EEG, and we can visualize the frequency distribution using something called a power spectrum. The power spectrum is the very first stage in starting to process the EEG to produce some of the
Speaker A
indices that we're used to. It's almost like using a prism to split the signal into its multiple different component parts. That's the function of the Fourier transform. The fast is the bit that just makes it happen a little bit quicker.
Speaker A
When we're looking at a power spectrum, and this is taken from the Narcotrend, the Narcotrend is the only device that displays the power spectrum, which is bizarre because every device is creating a power spectrum. The Narcotrend is the only
Speaker A
system that displays it to us, and this updates every 5 seconds, which is fantastic for induction of anesthesia at any time point where the brain activity is quickly changing. So this power spectrum is the EEG transformed into a
Speaker A
histogram, and here we can see height is the amount of activity in any particular frequency band. We've got frequency along the x-axis, and here in this particular example, we can see a very tall peak in our slow and delta
Speaker A
band, not so much activity in theta, and a large peak here in alpha, and very little high-frequency activity. With a little bit of experience and a bit of underlying knowledge, we can immediately recognize that this brain is anesthetized, and
Speaker A
this is propofol or volatile at less than 0.8 of a MAC, there or thereabouts, would produce this sort of brain activity, very, very recognizable. That coincides with what we'd see on the density spectral array, which is this
Speaker A
diagram here, which we'll dive into in a little bit more detail. When we're thinking about power spectrums, people often talk about something called spectral edge frequency. The spectral edge frequency, there are two, well, various different terms that
Speaker A
one can use, but the spectral edge frequency 95 is what most monitors will display, and that is the frequency below which 95% of the EEG power is below. So this point here, 95% of the EEG power is below
Speaker A
this point. Typically, that's less than 16 Hertz under anesthesia. Median frequency or median edge frequency is the frequency below which 50% of the EEG power resides. So these are very simple, these are the earliest sort of descriptive processed parameters
Speaker A
that you'll find in EEG science. How are you? Sorry, being gate crashed. I'm busy, Oscar. I'm giving a talk to people in Scotland. You know, you need to go talk to Mummy about that. Please close the door. Thanks very much. Say bye-bye to
Speaker A
everybody. Say bye-bye, everyone. Yeah, there we are. So the median frequency is the power below which 50% of the EEG is below. Now, there's problems with this. It's not a particularly useful value.
Speaker A
Some people get rather hung up and think it's useful. I think people think it's useful because it's simple. The problem with simple is it's not very good. So one of the things you'll see with spectral edge frequency is it will
Speaker A
increase natu-
Speaker A
increase spectral Edge it doesn't differentiate between deep unitized brain States and surprisingly some uh devices such as the sedline for instance weight the spectral Edge frequency very heavily in the algorithm so as the spectral Edge frequency increases the
Speaker A
PSI value increases now we've just said that it increases naturally with age and that can lead to some paradoxical behavior in the Aging brain as we'll get to a little bit later on looking at more deeply antiz brains as we see an
Speaker A
increasing amount of slow wave or even down into suppression of these particular values we see that paradoxical behavior of spectral Edge and median frequency as those values 10 9 13 increase as the brain becomes increasingly antiz and and here we are here's a
Speaker A
lovely statement from a review article I've referenced on the bottom to try to convince the naysayers who still enjoy uh spectral Edge that actually none of the studies in this particular review article were able to demonstrate that the spectral Edge frequency uh was
Speaker A
associated with adequate anesthesia spectr frequency must be adapted to the anesthetic drug use the patient's age and brain State under general anesthesia so it doesn't really have any value and just to add um sort of uh layer of layer
Speaker A
of concrete over this principle we're trying to bury and get rid of the bis monitor has a real problem with spectral Edge and median frequency if you adjust the filters on the uh device you'll find that you will alter artificially the
Speaker A
value of the median and spectral Edge frequencies as you turn the filters on and off as I've demonstrated with these black arrows so those are the sort of simple basic values that you might see on a monitor power spectrums spectr frequency but
Speaker A
what about these index values that these monitors show us let's have a little think about those for a few minutes well when you create an algorithm you're trying to produce a a descriptor of how the EEG changes from a wakefulness down
Speaker A
to Full Burst suppression an isoelectric Flatline you do that by um using multiple different parameters to try to describe the EEG and these are these tend to be proprietary and then we need to Anchor are scale to Something in the
Speaker A
real world and typically what um equipment manufacturers do is use clinical signs and what the aspect did when they were designing the bis back in the late 80s and early 90s was to um use the modified observers assessment of
Speaker A
anesthesia and sedation and they anchored the device around this point here which is m.2 which they described as conscious which is responds only after mild prodding or shaking I would say that that isn't actually unconscious to respond only
Speaker A
after a mild prod or Shake very few of my surgeons are mildly prodding they tend to use a scalpel which is slightly more of a stimulus I would say but um with that they came up with this particular uh table which shows this
Speaker A
index values where various different clinical correlates occur so for instance we have a patient who's awake and alert generating bis index values of greater than 80 and as our patients become increasingly sedated they will produce different ranges of
Speaker A
numbers so there is some attempt to correlate these index values with clinical um underpinnings that we used to now unusually though there's quite an error range with this particular system um and I'll talk about that in a few
Speaker A
slides time now we need to remember that these monitors are not monitors of Consciousness there is no such thing as a monitor of Consciousness because we don't have an EEG descriptor of what a conscious brain looks like or various
Speaker A
different conscious states of the brain and certainly using the frontal cortex which is what we're doing with our processed EEG machines it's impossible to dis differentiate Consciousness from unconsciousness with absolute reliability as I'll show you in a moment so what these machines
Speaker A
do is function as probability of recall monitors so back in the day the B was calibrated following healthy volunteer studies first of all with that clinical information I've just shown you but also with a test of recall so 22 patients not
Speaker A
patients but volunteers 22 people volunteered for the anesthesia under isopor mazalan and propol and were um given doses to achieve different B index values whilst at that uh following that dose at a deeper state of uh plane of um hypnosis
Speaker A
they were read a list of words six words Apple table Penny whatever they might be and when they emerged from that brain State they were asked what do you remember and they were tested for explicit recall what do you remember oh
Speaker A
I remember you saying Apple table Penny or implicit recall here's a list of 20 words do any of these sound familiar and what the uh Team working at aspect discovered was on their scale a volunteer with an index value below 64
Speaker A
had a 95% chance of having no recall blue line here so they round down to 60 a little bit easier to remember they use the value of 40 supposedly as the point where you start to see suppression on the EEG but as
Speaker A
we'll see later that isn't really uh a particularly useful end point so we have probability of recall B 95 below 64 but unconsciousness that not responding to a prod occurs at 50 we are much better at switching off memory than we are
Speaker A
movement now just to add more weight to my argument that these are not monitors of Consciousness more recent studies using the isolated forarm technique have demonstrated that there is no processed EEG value or EEG morphology which will prevent a positive isolated forearm
Speaker A
response in the 2 minutes following laryngoscopy so patients that had an if positive response could have B index values anyway where from 20 19 all the way up to 80 or whatever the incidence of a positive isolated forearm test is
Speaker A
between 5 and 11% up to 14% in a female cohort and it is not prevented by higher doses of induction agents opiates mazalan and there's no processed EEG value which is predictive the only things that were preventative in these this Trio of
Speaker A
studies from Amy gasal and Jamie slle were ensuring volatile was delivered prior to incubation to avoid the pharmacokinetic Gap and also to use TAA which we all know is the way the truth and the light now thinking a little bit
Speaker A
more about these algorithms these comp computational algorithms do have a directional delay so depending on whether your patient is moving from wakefulness down to Anesthesia or from anesthesia to suppression or back the other way there are different computational delays which can apply and
Speaker A
with the uh setline this particular uh Super Paper from AAS ker demonstrates up to a 90c delay if you're just watching your index values we really ought to be moving away from looking at these index values and looking at the raw EEG to
Speaker A
give us real time information to brain activity this nice study demonstrated that the B index value um doesn't correlate really with the more apology of the EEG even these uh dark lines here are showing us what the EEG looked like
Speaker A
when patients became unresponsive what was the brain doing at the point where patients no longer responded to a noxious stimulus the majority of patients are producing Delta oscillations or slow Delta oscillations uh by the time that they are
Speaker A
unresponsive if we look at that Delta or subd Delta activity there's quite a spread of index values that are commencing with those EEG waveforms so EEG morphology and this index value don't really correlate things are slightly different though with the Nar Trend system Nar
Speaker A
trend is designed using a descriptor of EEG morphology it is designed to describe the EEG to you this is really helpful if we want to discuss what the EEG looks like we have this wonderful taxonomy which is designed by a guy
Speaker A
called cougler a German electrophysiologist in the 1980s Modified by the narcot trend team and what this actually does is describe the morphology of the EEG in the same way that an electrophysiologist would do so the complex algorithms in this system
Speaker A
are converting the EEG into a taxonomy we can use where we want to be during anesthesia is during it's a stage of Delta oscillation and increasing Delta power will will give us a different narr Trend stage from d0 down to
Speaker A
e0 this has been compared the algorithm in the machine has been compared with the skills of a human electrophysiologist and there's a 97.6% correlation between what the electrophysiologist thinks of the EEG and what the device thinks of the EEG so
Speaker A
as a system to interpret eegs narot Trend really does very very well and I should point out that that comes from a paper uh dated in the year 2000 and there's been a number of revisions to the algorithm still then since then so
Speaker A
maybe the correlation between the algorithm and the electrophys physiologist is even better another interesting thing about this taxonomy is how it correlates with changes in slow frequencies and Delta frequencies and what we see during this stage here as Delta moves from 30 to 80
Speaker A
to almost constant is a change in um narc Trend index value but no change in bis index value the B isn't really very useful tracking changes in Delta power this is perhaps one of the reasons why if you increased your dose of opiate for
Speaker A
instance you very rarely see a change in B index value but you will do in Nar Trend state age because this device accounts for how the morphology of EEG is changing and opiates will produce a slow wave dominance
Speaker A
EEG so hopefully we're starting to think that the EEG is about more than just these numbers and if we look at these particular traces it's very easy to look at this and think fantastic 42 that's the meaning of life the universe and
Speaker A
everything I can um go back to my sodoku at this point but if we look a little bit further we can see here that well this value is elevated this Sr suppression ratio 27 if we look at the E EG we can see
Speaker A
these reproducible morphologies here about one hertz frequency that's the ECG so essentially I've got a flatline EEG the brain is profoundly suppressed at this point yet the monitor is showing me a reassuring value this cannot be right similarly we can look at this side
Speaker A
here and see oh 70 that's a bit High I need to do something about that whacking a bit of pro profile bit of fence turn up the volatile whatever it might be well let's have a little look again well
Speaker A
suppression ratio is okay and my EEG well that doesn't look awake we've got huge great big Delta oscillations here lots of lovely Alpha oscillations oh but a bit of EMG well this monitor is is misinterpreting this lovely um e very antiz looking EEG as
Speaker A
being wakeful because of EMG muscle signal uh from the frontalis muscle so it is important that we look Beyond these numbers to get a real sense of what's going on beond in the brain and nowhere was that better demonstrated
Speaker A
than Peter Schuler's study looking at 10 of his colleagues who were awake yet paralyzed first with suxamethonium then two weeks later they came back for more with rocuronium they had two processed e uh two bis monitor strips applied to
Speaker A
their head with two different eras of bis system and they were able to communicate with their colleagues through an isolated forearm and I'll just show you that here this is the Su aonian arm this is an anst completely paralyzed but conscious
Speaker A
12 plus n what do you indicating complete Consciousness and ability to communicate and perform basic arithmetic was being ventilated by a faithful colleague but the index values are depressed paralysis with the B system and the entropy will produce
Speaker A
artificially low index values despite the EEG being demonstrably wakeful low amplitude high frequency that is a brain which is awake wake patient is paralyzed yet the index values indicate the veil manufactures we need to know the monitors we use and know their
Speaker A
CS now a even better way of looking at the EEG than the power Spectrum some of the other things I've spoken about is to look at the spectrogram so this is the way that we really should be looking at EEG uh
Speaker A
system or e G signals much easier to interpret than the raar EEG and much more colorful as well so what is a spectrogram well if we take that stochastic signal which is a blend of all those different frequencies split it
Speaker A
with our prism or or our forier transform into its different component parts plot them as a power Spectrum these power spectrums can update every five seconds or so if we look at this imagine it as some little mountains
Speaker A
we've got our um cyclist shooting up the mountains over the Foothills up the great big climb at the end this can be transformed into a spectrogram we take many of these power spectrums and put them vertically with frequency along the
Speaker A
y- axis and time along the x-axis each one of these uh Power spectrums can be plotted next to each other so we see the mountain here on the far left that becomes the mountain in range along the bottom here with the brighter colors
Speaker A
indicating a taller higher Peak so we've got some high activity along the uh base here very low frequency Comm in with this here we've got small amounts of activity these Bluer colors here then we've got another Peak around about 8 to
Speaker A
12 Hertz this is Alpha that's this band here and then very very little higher frequency activity so this is a spectrogram or density spectral array they're the same sort of things and these are a wonderful way of looking at
Speaker A
how our EEG power and EEG frequency is changing over time um I can look at this and say that this is a youngish patient who's receiving proor based anesthesia um just from looking at that picture so one way to think about this is to
Speaker A
imagine that you're on your holiday you're flying away we can look down out of the airplane window over the mountains as we go onto our holiday we're going to see the great big high mountain peaks and then we're going to
Speaker A
come over the Foothills and then out over the sea and this is exactly what we see when we look out of the window of our plane flying over the density spectral array tall Peaks little uh Foothills and then out
Speaker A
over the sea now let's compare how different an aesthetic drugs look with the density spectral array well imagine for a moment that you're a Pianist and what I want you to imagine is that different anesthetic drugs stop us from
Speaker A
being able to play particular notes on the keyboard in the awake brain you can play all 88 keys play whatever tune whatever Rhythm however you like but when we give propol propol prevents the brain from producing particular frequencies it stops particular parts of
Speaker A
the brain working notably slowing down activity in the thalamus and between some of the cortico cortical Communications across the surface of the brain and from the frontal cortex back to the parietal cortex and those are normally high volume high speed
Speaker A
motorways if you like allowing conduction of signals rapidly around the brain with propor we slow those signals down and we only allow the Bas notes and some these lower notes to be played and that produces a density spectral array
Speaker A
that looks like this two tram lines volatiles a little bit different as you start to exceed a Mac 08 of a Mac up to about a Mac you'll see this bit in the middle these Theta oscillations these notes of the keyboard suddenly
Speaker A
start to fire up and be activated Dex metat tomine depending on how much you're giving if you're giving a lot you'll get a sleepy looking very sleepy looking EEG that's predominantly um these low frequency notes if you back
Speaker A
off and give an analgesic dose you'll see a pattern with a little bit more like this um perhaps slightly higher frequency there some sleep spindles in it but none of these patterns really indicate that a patient are neoti they
Speaker A
can be Bly rousable for Dex metat in sedation but look at ketamine ketamine does something completely different to All the Other Drugs it's an excitation drug of course so we're seeing high frequency oscillations produced by ketamine I'll show you some video of
Speaker A
this in a few moments time and this is what typically confuses our processed EEG monitors which are expecting to see the brain operate in a slower um state with uh typically the gabic related drugs and are volatiles ketamine is an
Speaker A
NVA receptor antagonist is starting to produce higher frequency oscillation the brain is still becoming antiz yet through a different mechanism now our different monitors produce a DSA in different ways so here's a comparison between the setline and Nar trend for instance so the set
Speaker A
line has a very rapid update so you'll see a new vertical line of pixels plotted every 1.24 seconds it's got two different types of DSA handing and multi taper um multitaper has been licensed from Emory Brown's team at Massachusetts General
Speaker A
it's a little bit like a bit more speckly well this is multitaper so you get a little bit more speckle for your money it's like a um not quite high definition but like watching a uh a VHS you've got you've got the DVD version
Speaker A
rather than looking at the VHS version so perhaps a little bit more um granularity should we say uh there's no scaling on this particular monitor you have to do that manually yourself and it's a bit more fiddly to scroll and
Speaker A
review your way through the DSA what I've Illustrated though that very very rapid update 1.24 seconds these These are the this is the same patient who's being neozed and induced at exactly the same time and I've got both monitors
Speaker A
applied to the same patient so this piece of the DSA on the Narco is that entire screen worth of uh setline based data this Scrolls across the screen very very quickly and you'll see it's also showing us both sides of the brain
Speaker A
simultaneously and whether there's a value in looking at both sides of the cortex independently outside of some extremely specific um th Neurosurgical or Interventional Radiology vascular surgery that is looking at the frontal cortex and prefrontal cortex is Up For
Speaker A
Debate personally my neuronist colleagues don't find any value in looking at um uh bifrontal uh d sa so we tend not to do it Although our Nar Trends will do that as well as our abys monitors and uh of course the um B also
Speaker A
does a DSA which is extremely simple very very basic there's no ability to change the scale um it can be exported as a PDF but there are some printing issues with that it prints it back to front for some slightly strange reason
Speaker A
it's a dimensionless value on this it just says high and low it doesn't give us any actual um ability to window or change the system so when we're looking at our uh different processed EEG devices and deciding what to get it's a really nice
Speaker A
idea to compare them together so I've been fortunate to be able to play with every device on the market and a few things that aren't yet on the market and we can see that our monitors perform in different ways on that presents dimy my
Speaker A
B monitor is picking up a lot of interference whereas the fing on the kn trend is very good that's not causing any interference ketamine um which are given uh to the patient being monitored with these systems increase the value on
Speaker A
the abys but the narc Trend it's a bit more stable and steady so it's showing me there's still a lot of slow wave activity despite the higher frequency activity that's evident on the DSA this particular monitor um which isn't
Speaker A
available in the UK at the moment it's very poor at um Computing vers suppression so we can see that we picked up just seven uh% burst suppression on the narc Trend this moment is showed an index value of 59
Speaker A
which is at the top of its working range that's indicating telling me to give more anesthetic the time when really I should be perhaps easing off a little bit and here's me using the set line and getting some correlation between the two
Speaker A
different systems um in their density spectral oay and E so it can be very useful to compare devices simultaneously and get a feel for what they offer I did say at the beginning once we'd had a skip through some of the
Speaker A
basics that we think about how process EG is a monitor of mechanism maturation and metabolism and this is perhaps the more interesting stuff and at the end I'll show you how it can also be a monitor of inoperative
Speaker A
Mischief so as we mentioned at the beginning the classical changes associated with induction take us from a low amplitude high frequency signal through to a very very high amplitude slow signal uh when we hit slow wave saturation which is where we knock out
Speaker A
the brain stem now in this particular set of Graphics I've done that nice and slowly but many times we'll be giving a rapid sequence induction and the Brain State can change very quickly and I want to illustrate for those of you who are
Speaker A
less used to looking at processed eug that um this is readily identifiable so I've got a little bit of video here at the bottom we've got the I've taken the screenshot of the awake brain low amplitude high frequency and the top is
Speaker A
going to be a little bit of video this is a 18yearold having an RSI uh for an appendicectomy um I do TBR rsis uh so she's going to get 160 milligrams of proy on a backgrounded four of Remy
Speaker A
fenil and let's see what happens so drugs are going in brains already changed huge great B Delta oscillations there and we've settled into the pattern of anesthesia that's all happens really quite quickly I've sped that up five times because it's uh
Speaker A
uh not desperately exciting to watch these things in absolute real time but the message there is that things do change in a way that you can pick up just by looking at the EEG and we can see here that this brain looks very
Speaker A
different in the antiz state to in the awake state where we've got this much higher amplitude and one of the beauties of working as an anst with EEG is the signal to noise ratio is amazing because we use neuromuscular blocking drugs
Speaker A
frequently we've taken out the EMG which works like a background static uh confounding our ability to uh interpret the signal but we also by the way our drugs work which hugely increases the amplitude we increase the amplitude of
Speaker A
the signal between 20 and 25 times we have the best signal to noise ratio of any clinicians using processed EEG our neurologists would kill for the signals we have now if we give drugs a bit more slowly we'll go through those stages of
Speaker A
uh anesthesia a little bit uh in a bit more of a controlled way here's a young patient having a slower titrated induction of propon we're going to see a big peak this is our power Spectrum this is really nice to be able to see this
Speaker A
during induction we're going to see as the patient becomes a bit more relaxed their eyes aren't going to Blink quite so much so we'll see that Delta artifacts fall down like that and the surface of the brain is now starting to
Speaker A
become activated this is called beta activation this is the brain starting to become paradoxically aroused as patients become sedated and we can track that frequency falling as the patient starts to uh move towards an unconsciousness unconscious State and we'll see those
Speaker A
huge great big Delta oscillations in a few moments time about now which is where their brain stem has become inactivated so the concentration of drug in the brain stem is high enough to prevent signal propagation of the spinal
Speaker A
cord into the thalamus and cortical arousal so this patient is now antiz but we took our time to get there nice slow smooth anesthesia and when we look at those if I put all of those bits of EEG together for that
Speaker A
patient we can see nice and smoothly how we go from that high frequency low amplitude through to that beta activation the brain becoming a little bit more roused there those oscillations get a little bit larger ever larger before we start to get to very very
Speaker A
large slow waves here commen it with brain steming activity you can perform a jaw thrust at that time or popular your masking or perform laringoscopia you won't get any response at that point because we've Switched Off brain stem
Speaker A
function and we can see that on the density spectral array that lovely sweeping down of activity or on this screen this is called relative band activities which is something on Nara Trend we can see this big stall ties of
Speaker A
beta activity the white bit here the beta frequencies as they uh as the brain becomes increasingly increasingly increasingly aroused before it becomes inactivated Delta Peaks up and we settle into the nice pattern that we expect from proor based anesthesia which is
Speaker A
that Alpha Delta pattern this nice thick magenta line and our um Delta frequencies as demonstrated by the blue bit on the screen there really very very nice to watch and what we see on the density spectral arrate depends on how quickly
Speaker A
we give the prop profile they broadly speaking two different ways of inducing anesthesia the slow titrated bonus which is a top down mechanism we um make the cortex hyperactive before the concentration in the thalamus and brain stem builds up afterwards so we NE tize
Speaker A
from the top down to the bottom of the brain or we deliver a fast bolus like you might do with the rapid sequence and we get a enormous delivery of drug to the brain stem and um Thalamus which
Speaker A
produces uh perhaps a little bit of birth suppression sometimes and something called Alpha collapse before we um um before anesthesia settles down into a similar pattern just a heck of a lot quicker because a larger dose has been given now I showed a couple little
Speaker A
pieces of information about ketamine showing how it produces a difference in our processed EEG and ketamine is a fascinating drug and I use it quite a lot for additional antios deception and what you see with ketamine is approximately 2 minutes after giving a
Speaker A
dose and on the naron I can put little markers in to indicate where I've done various things and these are fully customizable like on other unlike other systems so I have a m a flag here for ketamine and two minutes after I've
Speaker A
given a ketamine we can see an increase in frequency so the brain becomes um aroused and it's delivering uh OS producing oscillations at around about 15 Hertz at that point here we've got another example we've got that dog leg of acceleration 2
Speaker A
minutes following the dose of ketamine lasts for between 15 and 45 minutes typically depending on the brain and then that Alpha Delta pattern that we're so used to seeing with TAA recommences again so we can visualize the effect of
Speaker A
our drugs we can understand any change in the uh index values that are in front of us and we can visualize the effect of the ketamine we can also and what is interesting some of the work coming out
Speaker A
of the um elderly anesthesia is that if the brain responds like this with an increase and an acceleration in activity means brains are much less likely to suffer acute Recovery Room delirium than patients which cannot respond patients which can't respond when the accelerator
Speaker A
pedal is pressed in their brain unight to have a vulnerable brain and to have postotic delirium which is a using ketamine as a probe could be a fantastic proposition in the future let's have a look that happening again in slightly
Speaker A
sped up time we're going to administer a small bolus 30 milligrams to a 55y old man um he's under stable anesthesia at the moment this is that fantastic power Spectrum updating rapidly ketamine's gone in and I want you to look at what
Speaker A
happens to that Peak that was here for Alpha it's moved up to 15 Hertz exactly where it should be for ketamine my EEG is looking a little bit faster my index values have increased but but I know that this is ketamine
Speaker A
because it happened 2 minutes after I gave ketamine and it's absolutely classical 15 Hertz and that's going to last for around about 15 to 45 minutes before it shifts back down to producing that Alpha Delta power Spectrum or
Speaker A
density spectral array that I expected to see now we can make things with ketamine more exciting can't we I do a lot of trauma and emergency anesthesia as well as the other bits and Bobs that I do and let's look at a ketamine RSI and see
Speaker A
what happens to the patient's brain as I say I use process DG for everything whether or not I'm using T volatile sedation there's no excuse not to monitor the brain so this poor patient uh so lady she's having U um surgical
Speaker A
management of a of a miscarriage she's bled a lot so she's shocked I've given 150 of ketamine bit of alfentanil and what we're going to see in a few moments time is a lot of blinking and moving around
Speaker A
because she's very very anxious and diaphoretic as you might imagine but the ketamine's in it's producing this high frequency low amplitude oscillation this peak here this is the ketamine working she's going to be unresponsive at this point it's nice and flat there's no EMG
Speaker A
we haven't got huge Delta power but I'm am reassured that she's an etiz because this is a pathological trace this is not a trace associated with wakefulness this is what ketamine has done it's excited the brain it's producing anesthesia
Speaker A
through a different set of mechanisms volatiles on board now and we're seeing Theta power start to build remember I mentioned that Theta builds with a end tile of greater than 08 of a Mac that's exactly what we've drifted into C
Speaker A
Florine maintenance following ketamine induction and if we look at the power Spectrum for that that's that first little bit of ketamine excitatory stuff at around about 20 Herz get the volatile on board once the tubes in and we start
Speaker A
to see those Theta oscillations build before we turn off the anesthetic and we see an acceleration um in these high frequencies deceleration here as the brain starts to W drop so we can look at the density spectral Ray in even more detail and this is
Speaker A
really a dire of travel that we should be using so this is a from a colleague of mine and again time along the x-axis frequency along the y- AIS I've labeled different frequency bands to make it a little bit easier because there's a lot
Speaker A
going on here this is a critically ill 66y old lady who came for an emergency laparotomy with four quadrant feal peritonitis pretty pretty poorly she's septic she's drowsy she's hypotensive and what we can see from this this point here where there's all
Speaker A
this high frequency stuff this is when she's having lines put in a Wake Central Line arterial line This is EMG but there's a band of Thea occurring here and that is because she's so drowsy as we start to become drowsy our brain
Speaker A
produces theer oscillations this is not anesthetic this is the fact that she is really very very poorly she's then induced with mazalan and ketamine both of those drugs will produce high frequency oscillations and she's maintained on a ketamine infusion
Speaker A
that's what this horizontal band here is so we can be we can see a big difference the transition point where she's become unconscious and where anesthesia has been made using a different mechanism to what we'd seen with protal so the
Speaker A
density spectral array is a great storytelling mechanism to understand the changes that have gone in this lady's brain we can also be reassured that given how septic and poorly she was that she must have a reason well profused
Speaker A
brain to be able to produce these forms of oscillations so there's the EMG where the lines are going in Theta she's drowsy um ketamine and MZ to go off to sleep ketamine maintenance bit of hypotension here we call this I Alpha collapse maybe it's
Speaker A
more beta collapse really perhaps a period of brain hypo perfusion before the um brain was restored back to good perfusion so thinking about how our anesthetics are working though density spectral array tells us a lot about that remember
Speaker A
we're only monitoring the frontal cortex prefrontal cortex this little bit here these are these monitors are cortical activity monitors not depth of anesthesia monitors there is no such thing as a depth of anesthesia anesthesia does not have the
Speaker A
depth what we're seeing though is how we are producing a state of sensory disconnection we do not produce um unconsciousness um we have no way of determining that but what we can see is how we're unplugging the cables from the
Speaker A
internet router if you like thinking of the phus as the router we're unplugging cables that lead to other parts of the brain and this study demonstrated that really nicely every green line here is an auditory stimulus um and a little dot
Speaker A
at the top of it is when a vol uner pressed a button they will play The Sounds they had to press a button every Green Dot is a response the purple lines indicate an increasing propile concentration as the patients became
Speaker A
increasingly sedated they weren't able to press the button in response well no surprises there but the EEG patterns were very specific when we saw that lovely Alpha Delta pattern as we've seen in the example examples I've shown so
Speaker A
far as well the patients were able to process any sensory input this is the point at which the thalamus has become inactivated to sensory input sensory disconnection this is what we're producing on beneath this SI of people sensory disconnection and that's exactly
Speaker A
what we see when we um produce anesthesia that looks like this this is a full anesthetic from induction all the way through to emergence I've turned the drugs off here and the Brain started to accelerate and I can see that all on my density
Speaker A
spectral array so what's happening uh neurophysiologically well here's a brain here's the thalamus the thalamus is exquisitly sensitive to the effects of propool and what we're doing in um neurophysiological terms is closing the thalamic gate and that's preventing
Speaker A
propagation of signals from the spinal cord into the particular activating system from getting up into the cortex so we've closed off these thalamocortical Pathways preventing the uh normally rapid transition of information at 30 to 50 htz we've slowed
Speaker A
the signaling down down to 8 to 12 Hertz exactly what we see on the monitor with our alpab band and these connections between brain areas um have been slowed down towards Delta frequency so we're seeing a breakdown to theortical
Speaker A
communic ation and a breakdown of cortico cortico communication a way I like to think about integrating hypnosis and antinos setion is to think about the phalus as a castle and how we need to defend the cortex from Attack here's our castle castle phalus
Speaker A
here's our village the cortex Cottages which are connected to each other via the the C pathway our patients are being attacked by the Dark Nights of NOS acception intubation and surgical stimulus we have the nights of hypnosis preventing the
Speaker A
attack on the cortex we can provide a mo of analgesia to slow these guys down or make it more difficult for them to get to us if we've got inadequate hypnosis then it's pretty easy for our Dark Nights to cross over our analgesic
Speaker A
blockade stimulate the thalamus producing thalamic breakthrough and potentially cortical arousal and what does that look like well during um covid when we had to uh use tho rather than propol certainly my hospital we were very short uh performing rapid sequence inductions
Speaker A
with sodium thop penel rather than using propal gave a wonderful opportunity to be able to use processed e G and to see what this looks like so this patient is going to have um 200 milligrams of tho bit of alfentanil very standard RSI
Speaker A
we're going to fall asleep become antiz incredibly quickly tho Works fantastically quickly but has a very very rapid offset I ventilate my patients during um the gap between incubation between loss of consciousness and incubation with SEO so he's got SEO
Speaker A
on board but following laring op y the um EEG is becoming very flat we're shifting to higher frequencies and this is the brain becoming stimulated even though we've got we've started to ventilate the patient would see a Florine before the tubes in we see the
Speaker A
brain becoming aroused and that Delta power takes a while to recover and this has been well documented by many other EEG um researchers and it has a term beta arousal and this is what the Cort this is what the EG looks like pre-
Speaker A
laryngoscopy looks reassuringly antiz lovely great big Delta oscillations but following laryngoscopy we're starting to see a real flattening of the EEG and a breakthrough of stimulus fortunately this patient had no evidence of um any uh awareness following that brief
Speaker A
stimulus but it shows very nicely the pharmac kinetics and the phac Dynamics of a legacy drug such as San fire pent now as well as ensuring that we've got adequate hypnosis if we ensure we've got adequate antios setion our Knights have
Speaker A
been busy and they've dug a much much bigger moat so we're preventing the Dark Nights of no reception from Crossing and attacking the thalamus and that can be a really useful thing to do and this is an 80-year old patient she's got a thoracic
Speaker A
epidural and she's having a um laparotomy for um bowel perforation and we're running a really very very small amount of anesthesia um I use elel these days so she's got about 12 Ms of propol running and 12 mph propile and her Remy
Speaker A
fenil is running at 2.5 which is under .1 mics per kilogram per minute so very very low doses but my processed EEG monitor is showing me that her brain is producing a very slow wave dominant EEG as the narren stag is indicating and
Speaker A
backed up by my density Spectra array so this is a brain that does not need very much anesthetic there's very little NOS deceptive information getting past the epidural blockade I'm probably running some metamol to maintain a good profusing pressure but I'm not
Speaker A
overdosing her brain and she will wake up nicely at the end and be able to return to the ward afterwards there are three typically three different forms of um arousal that can can Ur with the EEG which are um
Speaker A
beta arousal and we see like we saw with that sodium Fire penel case acceleration in the brain Alpha Dropout which is probably mediated by vagal nerves um so this is vaguely mediated that we see a fall in our Alpha power here typically you see
Speaker A
that with laparoscopies so when you're performing insulation of the abdomen creating num peritoneum or washing out and then there's paradoxical Delta arousal which can occur and the Delta power gets higher you can just about make it out there I think I've only ever
Speaker A
seen that once or twice to be honest and the EEG values can paradoxically fall with that even though the brain is more aroused quite difficult to interpret now as our patients age the density spectular AR a changes nothing's
Speaker A
easy in life but what we do find is that the alpha power for as patients age and we get to around about 65 here we've got age and years and this is a little snapshot taken at a stable point of
Speaker A
anesthesia and at 65 years of age the density spectral Ray is starting to become much different the alpha power is falling and that fall in Alpha power is associated with um anatomical changes in the brain thinning of the frontal cortex
Speaker A
for instance particularly areas of the brain that are associated with executive function and patients who are multi comorbid patients who have um delirium or dementing syndromes will have much lower Alpha power than their age match peers who are cognitively normal one of
Speaker A
the problems with many of the processed EEG systems is that they are age agnostic they don't take this into account they're calibrated for a brain of 45 years old with a reasonable amount of alpha power they're not so good in
Speaker A
patients who are very young and certainly not very good in patients who are more elderly and not all of these indices are created equal so for example with bis if you take three cohorts of patients um the 30y olds 50 to 70 year
Speaker A
olds and the over 70s and you give them all one age adjusted Mac the elderly patients paradoxically will produce higher index values than the younger patients for the same Mac value of anesthesia and what's been found across a range of
Speaker A
different devices sedine bis entropy and qon is that they overread in the elderly patient so as the patient ages typically these values become increasingly inaccurate by two index units per decade which doesn't sound like very much but that means your 40-year-old and your
Speaker A
90-year-old would have a 10 unit difference so if your 40-year-old produced an index value of 50 your or 90-year-old would produce a value of 60 what do people do when they see 60 they give more anesthetic so there's a real
Speaker A
potential to overdose the elderly the Narco that has an age adjusted algorithm does something different and produces what you might expects that the elderly patients produce a lower index values than the younger patients and that's what should happen because elderly
Speaker A
patients produce less Alpha but paradoxically a little bit more bet but certainly more slow wave with much more delicate brains now I'm going to skip on just a little bit talk about birth suppression because that's an important topic that's
Speaker A
often um not well understood and this comes under metabolism so how can we use these monitors as monitors of metabolism well as cereal blood flow Falls our EEG becomes increasingly flat and suppressed as metabolic activity Falls as shown in
Speaker A
these um pet scans on the left hand side as we move from warm colors to cold colors um are EEG indices for as one might expect so we can use these monitors as monitors of metabolism as well as a drug response what do we mean
Speaker A
by birth suppression well birth suppression is periods of suppression and ISO electric EEG Plus or minus5 microvolts and then intermittent burst then a period of iso electricity then a burst complete suppression is when we don't have any bursting activity at all
Speaker A
this is really if we were going to use the term depth of anesthesia profound anesthesia this would be a profoundly antiz brain State and this is something we really need our monitors to tell us about when our patients brains are not
Speaker A
generating a normal EEG and that can be because of anesthetic overdose but there's a range of sign ific problems which I'll show you in a moment that can lead to suppression unfortunately not all monitors are created equal and the B
Speaker A
monitor has real um I have a real bug bear with how the B monitor integrates suppression into its algorithm and this graph here shows us B index value by suppression ratio and you can see here that these suppression ratios between 0
Speaker A
and 50 don't really influence the B index value very much at all that's why you might have seen satisfactory index values but very highl looking suppression ratios and even isoelectric eegs like this the bis needs a lot of
Speaker A
suppression to produce low index values so this point here these index values between 30 and 40 describe an enormous range of brain States from optimal through to profoundly suppressed remember a suppression ratio of 50 is over 30 seconds of iso El electricity
Speaker A
per minute so it's not very well coupled the Nar Tren does something a little bit different we go right back to just two seconds of suppression to produce a low stage on the system and a low number between two and 20 seconds moved us into
Speaker A
the next stage over 20 seconds to the deepest most profound um stayed on cuda's taxonomy and this is all indicated to the user in bright red numbers indicating that there certainly something a miss so it's very tightly coupled to
Speaker A
suppression um entropy and sedline again they do something slightly different too sedline needs about 20 seconds of suppression before its values start to fall entropy looks quite favorable and that it's tightly coupled it uses a separate algorithm for suppression but
Speaker A
unfortunately in studies of humans versus site humans versus entropy humans outperform entropy significantly it's not very good at detecting suppression in the first place so I mentioned that U multiple problems can cause birth suppression it's not just anst here on the left hand side
Speaker A
we've had a hemorrhage in um our HPB theater with patient lost about six lers of blood fairly quickly um and then required a massive transfusion obviously we you can see suppression ratio increasing and the uh uh index value
Speaker A
falling which is appropriately um filled in with red there here's a patient who became acutely hypoxemic following a surgical complication SATs fell to about 30% suppression ratio shot up this is a patient who's had a cardiac arrest and
Speaker A
been resuscitated there bonus of adrenaline get the heart started Cardiac Arrest uh occurs again but fortunately is recovered and brain perfusion is restored and the patient goes off to itu but there's multiple other causes of suppression hypothermia typically core
Speaker A
temperatures less than 17 de hypotension hypoglycemia one always to remember vascular brain injury which we've seen and brain death of course so just to pull these things together I want to show you this super little case one of
Speaker A
my colleagues very kindly contacted me about this and because we've got access to narcotrend systems I have the full Telemetry from the case this is a patient 77y old male who presented to our emergency spines list with unstable
Speaker A
thoracic spine fractures so full care uh log rolling inline stabilization meticulous care with videoendoscopy Etc to try to prevent this guy's um thoracic spine fractures becoming uh a permanent paralysis he's had a very carefully titrated TAA TCI induction with standard
Speaker A
monitoring so blood pressure cuff cycling every two and a half minutes and processed EEG because we use that a lot where I am so so he's going to have an anesthetic induction with that carefully titrated propol and when he loses
Speaker A
Consciousness he's going to now he's going to get a dose of his neuromuscular blocking agent some rocuronium he's elderly so it's starting to look a little bit featureless it's being handbagged at the moment blood pressure cuff cycling and
Speaker A
cycling and the blood pressure Shu cycling again so what do we do we just check it's connected and uh have a fiddle with it and my colleague notices that the EEG is now becoming isoelectric this is now very very flat
Speaker A
indeed suppression ratio is increasing so he decides that the patient actually profoundly suppressed and this is cardiac arrest and the patient gets an immediate intramuscular bolus of adrenal in and we're able to restore perfusion very very quickly and we can
Speaker A
see that on the density spectr aay here this is the period of profound hypotension so he's G he didn't actually arrest sorry he had profound hypotension cereal hypop perfusion adrenaline to treat the um anaphylaxis and then a discussion with the surgeons in critical
Speaker A
care about what to do obviously gets no arterial line and central line but actually it was in this man's best interests once he'd recovered from that immediate hypotension um from the anaphylaxis to go on to have surgery and
Speaker A
here we can see the Alpha Delta pattern indicating really good brain perfusion so my colleague emailed me afterwards and said that uh that he waited a few more minutes fiddling around with the tube I think he would have arrested and
Speaker A
if he'd have had CPR in the context of his unstable thoracic spine fractures he would have been paralyzed of course and it was fantastic to get the information that the EEG provided because that saved time a real time monitor of anesthetic
Speaker A
mechanism brain maturation and metabolism and of course it prevented some intraoperative and anesthetic Mischief so we've talked a little bit about processed EEG but it's so much more than what I've explained today of course it's a monitor that can give us
Speaker A
an indication of awareness or rather sensory dissociation of deception brain responses to what we're doing we can see physiological responses to uh stimulus but also um hypop perfusion we can see how the aging and maturing bra brain um
Speaker A
is affected by illness vulnerability and birst suppression we can see changes in pharmac kinetics as I showed you with fast and slow probor boluses as well as the effects of our different drugs and we've even been able to pick up
Speaker A
epileptic forms which are very very very common epileptic forms occur in about 75% of children who are having U volatile um inhalation inductions so very very common indeed and of course I'd like to invite you all to Siver if
Speaker A
you can make it at the end of the ne at the end of November where we have special workshops um on processed EEG but also as well as that pediatric T and uh Tony Absalom and team from gragan
Speaker A
will be running to eleval worksh shops to take a much deeper dive into that fascinating new Pharmaca kinetic model thank you very much indeed I'll take any questions with pleasure
Topics:processed EEGanesthesia monitoringEEG frequency bandspower spectrumspectral edge frequencypropofolbrain metabolismdensity spectral arrayanesthetic depthbrain maturation
