Anaesthetized brains can still process podcasts

0:23Welcome back to the Nature Podcast. This week, what the unconscious brain might be taking in. I'm Benjamin Thompson.

0:41There's a question in neuroscience that has been tough to solve. Namely, how much is consciousness required to make sense of the world around us? Like, do we need to be alert and have all sections of our brain on deck to figure out what's going on? There is evidence that when someone is not conscious, the brain still has some ability to process information from a person's surrounding. For example, the brain region that processes sounds has been shown to activate when an unconscious person hears something, but how much of that information is passed on and processed in other parts of the brain related to cognition isn't as well understood.

1:22Ultimately, do people need to be conscious to process and understand complex information? This week, a team have tried to shed a little bit more light on this by taking neuronal readings directly from people undergoing brain surgery. These folk were, understandably, in anaesthetic-induced unconsciousness, and the team took this opportunity to see how their brains responded when hearing complex sounds. The team report their findings in Nature this week. One of the team is researcher and practicing neurosurgeon Samir Shef from Baylor College of Medicine in the US. I called him up to find out more and he gave me a bit more background about what's known about when an unconscious person hears a sound and some of the motivations for this work.

2:09So the auditory cortex will activate in an unconscious person when sounds are playing. But even beyond that, this has been known for quite a while with EEG and fMRI, when you have what's called an oddball paradigm, where you have a series of beeps and then the occasional boop, the brain is trained to be aware of and pay attention to these oddballs because they're relevant for real life. They might be threats, they might be opportunities for food, etc. And so we had done some recordings initially where we played sort of this oddball paradigm. And sure enough, we saw that even under anesthesia, these individual neurons in this regional brain called hippocampus responded quite robustly to these oddball sounds.

2:51That wasn't super surprising, but it did suggest that the brain is paying attention to statistical properties of the environment. So then we said, well, the natural language, of course, has a lot of statistics, You know, how often there are verbs or adjectives, etc. So what if we just played natural speech? Would these anesthetized neurons in the brain pay attention to the statistical properties of natural speech like they did to the basic sounds of this oddball beeps and boops paradigm? You mentioned that you're interested in the activity of the hippocampus specifically. Now, this is something that folk often typically think about as being important for memory.

3:29Why were you looking at this part of the brain in relation to speech? Absolutely. I mean, classically, the teaching in graduate school and medical school is that this is the part of the brain that's important for memory, because it is. But also, more recently, and a little more kind of in a niche part of the field, is that the hippocampus, of course, involves memory, but that's a subset of what it's doing. Really, it's doing something that is much more broad, which is taking attention to just the hierarchical abstract structure of the environment. So second to memory is place navigation. What is that?

4:02Again, it's just paying attention to kind of, you know, how physical space is arranged and what its structure is. And if you go even higher up in terms of abstraction, you know, how is the world organized? The hippocampus may be involved in that as well. And getting direct and specific measurements of the hippocampus activity can't be an easy thing to do. These structures are, of course, I don't need to tell you, very deep in the brain. Tell me about your experimental setup and how you went about recording the activity of the hippocampus. So for decades now, there's been the opportunity to record from single neurons in the hippocampus in individuals who are undergoing seizure monitoring where we have electrodes in their brain for several days.

4:37So we are certainly not the first to record from the hippocampus. What this, I think, is the first example of is doing so at the scale that we did it. And that's because we took advantage of a device called the Neuropixels probe. The probe is one centimeter long, but it's only like 70 microns wide. So it's the size of a human hair, but it has over a thousand micro contacts. This probe allowed us to record from scores of neurons, you know, 100, 120 neurons at a time. And what's key here is that you could do these tests in people who are undergoing surgery, seven of them, I think. Yeah, these are individuals who had severe and medical refractory epilepsy, meaning they're having seizures.

5:14They've tried many, many medications and medications have not stopped their seizures. This is a common problem. And this particular subset had, in some ways, kind of the most amenable type of epilepsy because it's coming from the hippocampus. And we each have one hippocampus on each side of the brain. And it's coming from the hippocampus that is safely removable. So this is a common surgery, which is a standard of care procedure that is highly effective. So the good news is all these individuals were candidates for this very effective surgery and all did very well and are seizure-free. The patient is there for the purpose of this clinical procedure to have the hippocampus removed.

5:49With their informed consent, just before we're about to remove it, we place this Neuropixels probe within it. They're still asleep. There's no need for them to be awake. We played these podcasts, which took 15, 20 minutes. Once that was done, we took the probe out. We proceeded to surgery, which involved removing the hippocampus, etc. So that's the clinical opportunity that we kind of piggyback this research on top of. And we'll get to the podcasts in a bit. But earlier on, you talked about playing these oddball sounds to some of these folks. What did you see there? We saw two things. One is that these hippocampal neurons did care about the presence of the oddball tone, a boop instead of a bunch of beeps.

6:25The neurons overall just fired at a higher rate. As some indication of, hey, this is something different, pay attention. Number two, over the course of the 10 minutes that these beep sounds are being played, much to the entertainment of everybody else in the OR, the neurons kind of learned over time to pay more attention to this oddball sound. So the first few minutes of the recording, that difference in firing rate was not statistically significant. But over the course of the next several minutes, it became more and more significant. The neurons just figured out, of course, without being instructed, obviously this person was under anesthesia, that this is something to pay attention to.

7:02And so then every time there was the oddball sound, they fired at a faster rate, indicating a sort of plasticity or learning that's occurring within the hippocampus. And one of the things that you do in your work as well is take this further then. And as you say, investigate what happens when people listen to speech and you played them a selection of podcasts. I do know one of them wasn't the nature podcast, Samir, but maybe next time. Tell me about this experiment. Yeah, we made a decision to use the moth radio podcast. And next time we will definitely use the nature podcast. So what we saw was that much to our surprise, several things.

7:33One, for example, was that like with the oddball thing where the neurons fired more to the less common occurrence, words that occurred less commonly, like cosmos, when those less common words occurred, the neurons fired faster. Then we saw that these neurons actually cared about the syntactic properties of speech, the grammar properties. There'd be neurons that fired in particular when there was a verb that was being listened to at that moment versus a noun versus other parts of speech. And then finally, they were paying attention to the actual semantic content, meaning the meaning of the story.

8:12The best example of this is that the firing rate of the neurons would not just encode what the current word is that's occurring, kind of whether it's a noun or a verb or a pronoun, but it would actually encode a prediction of what the next word is going to be. So we do this in natural speech all the time. If I were to start a story about a person who was playing baseball and they hit the ball and it went over the moon, one would be surprised because that's not the word that you would expect. You'd think over the fence or over the wall. So we are always making a prediction about what word we're about to hear next and when there's a conflict to that prediction, we pay attention.

8:48So these neurons were indeed encoding not just what's being heard at the time, kind of keeping track of what's going on, but making a prediction about what the next word is going to be. And we could decode that from the neural data. Again, indicating a very sophisticated form of processing of the natural speech that they were listening to. And based on the evidence you've shown then, what do you think this work means broadly? Well, I think number one is that For people who consider this anesthetized, unconscious state as one of dormancy, I think that idea very much needs to be expelled in favor of one that conceptualizes the subconscious, unconscious, anesthetized brain as still actively processing the environment.

9:35I think that was the major takeaway. And now we have kind of a neural mechanism for how this is occurring in terms of at least this part of the brain that certainly is very involved and interested in the structure of the environment. It is paying close attention to what is happening in the outside world. So the people who are undergoing this procedure, I mean, did they have any memory of this? Because you say that the brain is paying attention. Does this information get used in any way once they're brought round? We asked all of them and explicitly they had no memory of it. in our next studies, we will be testing for any kind of implicit, you know, memory.

10:10So for example, like what if we had them listen to a bunch of podcasts, including yours, and just say, which is the one that you have some familiarity with, you know, and test it in, you know, kind of a psychophysical way. I mean, of course, it has to be said that this is one area of the brain that is responding in a small amount of people under one specific anesthetic drug, right? And I think it's fair to say that being under anesthetic isn't the same as being asleep, isn't the same as being in a coma, I suppose, or isn't the same as being unconscious after a head trauma, these sorts of things.

10:41So is this an N equals one? All of us on the study team want to be very clear that we're not trying to be overly broad in our implications that we're suggesting here. So whether this even applies to other anesthetics, we don't know. Certainly we don't know whether it applies to an altered state of consciousness because of head injury, sleep, I think very much motivates that exact question. What kind of similar differences are there in these other states of altered consciousness? And what do you think this work doesn't do? I mean, what questions did you not answer? Or what work would you like to have done?

11:11What are the gaps? We'd love to know, even just within the hippocampus, exactly which cells that we were recording from. It's a little tough because, you know, we had 20 minutes to do these recordings. It's hard to know exactly where we were. And this is just hippocampus. What about all the rest of the brain? Hoping that future studies can find that out. And then again, maybe the most important thing is, what can we do about this if there is somebody who is clinically in a state of altered consciousness? Can we bring them back into a more conscious state if they're there because of injury? And finally, Samir, listeners won't be able to see this, but it has to be noted, you are in full scrubs right now.

11:43I don't think I spoke to a surgeon about to go into the OR before. I mean, does your work make you reassess what might be going on with the people you treat, what they might be taking in, do you think? I think so. I hope so. I very much believe that the better we understand the brain, the better chance we have of healing it when it's damaged. So I try to relate to individuals, of course, as a human being dealing with the difficulty that they're going through. But also, I think the scientist part of me wants to know better what it is at the very basic mechanistic neural level that has been altered, because I think that knowledge empowers us to do a better job of treating those problems.

12:21That was Samir Sheth from Baylor College of Medicine. To read his paper, look out for a link in the show notes. Finally on the show this week, it's time for the research highlights, read by Dan Fox.

12:36A quick electrical jolt works as well as a taste test when determining the strength and roast of a cup of coffee. Researchers put electrodes in cups of coffee and found a linear relationship between the amount of current that passed through them and the strength of the brew. During the measurements, the electrodes became coated with molecules, including caffeine, and more molecules stuck to the probes in drinks made with beans roasted to a darker colour. These measurements allowed the researchers to infer both the strength and the roast level, which affect the flavour profile of the beverage. The test could enable coffee roasters to rapidly assess the quality of different batches of the same coffee, which could vary in flavour but otherwise look identical.

13:21You can taste test that research in Nature Communications.

13:30Decades of US forest data suggests climate change has exacerbated the damage caused by insects and disease. As temperatures rise and rainfall patterns change, some tree pests can expand their range and reproduce more prolifically. Meanwhile, warmer, drier conditions also cause trees stress, which might make them more susceptible to pests. Researchers looked at 20 years of data on pest damage in forests across the continental United States and modelled how damage caused by 30 species of harmful insects and fungi was affected by factors like extreme temperatures. They found that pest damage tended to be worse in places where the maximum temperature in the warmest month was moderate or rising, though there was a limit to how well pests tolerate heat extremes.

14:20And regions in which rainfall and temperature shifted away from historical trends experienced harsher consequences than regions with more consistent climates. You can find that research in full in Nature Ecology and Evolution.

14:40Dan Fox there. That's all we've got time for this week. Don't forget, you can reach out to us on social media. We're at Nature Podcast. And of course, we're on email to podcast at nature.com. I'm Benjamin Thompson. Thanks for listening.