Science Friday - Predicting Heart Disease From Chest X-Rays With AI | Storing New Memories During Sleep

Episode Date: April 8, 2024

Dr. Eric Topol discusses the promise of “opportunistic” AI, using medical scans for unintended diagnostic purposes. Also, a study in mice found that the brain tags new memories through a “sharp wave ripple” mechanism that then repeats during sleep.How AI Could Predict Heart Disease From Chest X-RaysResearch on medical uses for artificial intelligence in medicine is exploding, with scientists exploring methods like using the retina to predict disease onset. That’s one example of a growing body of research on “opportunistic” AI, the practice of analyzing medical scans in unconventional ways and for unintended diagnostic purposes.Now, there’s some evidence to suggest that AI can mine data from chest x-rays to assess the risk of cardiovascular disease and detect diabetes.Ira talks with Dr. Eric Topol, founder and director of the Scripps Research Translational Institute and professor of molecular medicine.Neurons ‘Tag’ New Memories For Storage During SleepAll day long we’re taking in information and forming memories. Some stick around, others quickly fade away. But how does your brain push those memories into long term storage? And how does our brain recognize which memories should be kept and which should be discarded?This topic has been debated for decades, and a recent study in mice may help scientists understand this process.Researchers found that during the day, as the mice formed memories, cells in the hippocampus fired in a formation called “sharp wave ripples.” These are markers that tell the brain to keep those memories for later. Then, while the mice slept, those same sharp wave ripples activated again, and locked in those memories.Ira talks with Dr. György Buzsáki, professor of neuroscience at the NYU Grossman School of Medicine, about the findings of the study, which was published in the journal Science.Transcripts for each segment will be available after the show airs on sciencefriday.com.  Subscribe to this podcast. Plus, to stay updated on all things science, sign up for Science Friday's newsletters.

How exactly do memories get stored in our brains? Sleep has a lot of important processes, but one of them is making memories. It's Monday, April 8th. Happy Eclipse Day. You're listening to Science Friday. I'm Sciafri producer Shoshana Buxbaum. It's impossible to remember everything that happens to us. So how does your brain decide which memories to keep and which to throw away? In a bit, we'll talk with a neuroscientist about his latest research on how memories

form in our sleep. But first, how AI can use chest x-rays to predict the risk of cardiovascular disease. Here's Iroflato. The last time Dr. Eric Topal came on the show, he talked about the exciting new ways that artificial intelligence might help physicians make better diagnoses, like using the retina to predict disease onset. Well, now there's some new evidence to suggest that AI can mine data from chest x-rays too. So he's back to fill us in on this. Dr. Eric Topal, founder and director of the Scripps Research

Translational Institute, professor of molecular medicine based in La Jolla, California. Welcome back to Science Friday. Always great to have you. Oh, thanks so much, Ira. In your newsletter this week, you wrote about a study that showed that AI could better gauge a patient's

risk for heart disease by looking at a chest x-ray as compared to standard ways doctors gauge risk. How do cardiologists typically assess heart risk for patients, and how might this be different? Yeah, this one is really wild, Ira. There's been several of these so-called opportunistic studies where you get information from a scan that, you know, we would never have envisioned as possible. The normal way we look at it is there's nine variables that include blood pressure and cholesterol and several others. And that predicts a person's

10-year risk of having a heart attack, a stroke, or cardiovascular death. So that's based on Framingham and other, many other studies. And that's what's been used all these years. Who would have ever thought that you could get better information from the chestaches? Well, tell us about that. What kinds of information does AI pull out of it that's so useful? I wish we knew. This is the problem. You know, some of these opportunistic studies have really dealt into explainability. Like, for example, another chest x-ray study that diagnosed diabetes, which again, I would never have thought that was possible, right, from the chest x-ray. It basically did the so-called occlusion or masking where it would look at the chest x-ray and block out various regions to find out what

was the source of the information that we can't see? And it turned out it picked up the fat pads in the chest that was providing this diabetes possible diagnosis. But for the risk of heart attacks and stroke, we don't know how it was so powerful and better than our standard. So this is part of that X factor that we still need to learn. Well, I'm still not quite understanding this. I mean, how did the AI, what did it see in the chest x-ray that told doctors or it found that the patient's risk for heart disease was greater? Well, we can tell from other studies using the chest x-ray for this opportunistic detection, that it can pick up the calcium score, what people can undergo a CAT scan to see how much calcium they have in their coronary arteries. But that can be derived from a chest x-ray. and that is a indicator of risk.

Also, the chest x-ray can be, through AI, determine the heart strength, the so-called ejection fraction. So it's picking up a bunch of things, as seen in other studies, that are very predictive of a person's risk. And it must be the composite of these things, but we really don't know,

because although the study really was extraordinary, it didn't do enough as far as the explainability side, of things. So that's where some of the work that needs to be done. And it needs to be replicated before this becomes a standard way to predict risk of a person's harder stroke in the 10 years ahead. And one of the things you point out in your newsletter is that there are so many chest x-rays done each year, right? And you could repurpose them without having to take an additional one. Exactly. The fact that there are over 70 million chest x-rays in the United States each year alone, it's incredible. So there's all this free added information in those chest x-rays that could help us because, you know, most people don't know their cardiovascular risk. And that's important, especially, you know, for determining whether a person should take a statin and whether there should be a intensive statin type of medical. and dosage. So there's lots of things that can be done to mitigate risk. It isn't just that

you're going to have this bad future forecast. There are things that can be done to change it. So, yeah, as you point out, Ira, there's just so much chest x-ray data out there that isn't being used today. And what about, you know, you talked about the retina last time. Now you're talking about chest x-rays and lungs. What about picking up some other organ? Could we, could we, scan some other organ and find other stuff from it. Well, another example from the chest CT is the ability to pick up pancreatic cancer because radiologists don't look at the pancreas in a chest CT. So cancer risk that it's already being done for a scan otherwise.

But the other thing that's been noteworthy, as you mentioned, the retina, the electrocardogram has lots of information and give credit to the Mayo Clinic. They are the first to now use that opportunistic electrocardogram to tell all their doctors whether the patient has a low ejection fraction that is diminished heart strength and other diagnoses that are not done anywhere else in this country. So we're starting to see that transition of this being able to detect more and providing that to the patients through their doctors. Yeah, because now that you say that, that stimulates me to think.

think about could we be picking up something erroneously by using AI, right? That may not be important, but AI thinks it is. That's a really important point. It's just like when you do scans for a person that doesn't have symptoms and you pick up these incidental findings, that could be the problem, too, the false positives. So that's why before these are put into clinical practice on a routine basis, we've got to know whether you're going to get false signals. And, you know, so far, the data look encouraging, but it's still early.

Well, if you say it's still early, and it's a good point, how do we make it not so early? I mean, what needs to be done? Well, ideally, we'd have at least one more study to confirm the new one for the chest x-ray and picking up the risk of a person. And we'd have a much better handle on how did it do this? How did the AI figure out this risk so precisely? better than all of the studies have been done for the last many decades, right? So we just got to confirm this and understand it better.

But if it really holds, then someday, Ira, this will become a routine output, a readout from chestex range. Well, we'll wait for that day. Eric, thank you for taking time to be with us. And come on back when you've done some more snooping around. Okay, we'll do. Thank you. You're welcome, Dr. Eric Topol, found out. and director of the Scripps Research Translational Institute,

professor of molecular medicine that's in La Jolla, California. Some memories stick around and others quickly fade away. But how does your brain push those memories into long-term storage? And how does your brain recognize which memories should be kept and which should be discarded? You know, this topic has been bashed around for decades. And now scientists are beginning to better understand this process. and you know what? Sleep. Sleep may be the key. Here's the explanation in brief. During the day, as we are forming memories, cells in the hippocampus in our brain fire in a formation called sharp wave

ripples. These are markers that tell the brain to keep those memories for later. And then when we're asleep, those same sharp wave ripples activate again and lock in those memories. Pretty cool. Joining me now to explain more about the findings of his recent study published in the journal Science is my guest, Dr. Yuri Bezaki, Professor of Neuroscience at NYU Health based in New York City. Welcome to Science Friday. Thank you. Good meeting you, Iro. You're welcome. Let's start at the beginning. What's going on in our brain when we form memories throughout the day? What's going on in there? What I would say is that we interact with the world constantly and continuously during the day, but those are not memories yet.

In order to deposit them and to make memories, you have to go through some kind of a selection process. We have known it for a long time that sleep is a special state because it helps either rehears our memories or do something about experiences. and then the sleep itself is useful for consolidating the experiences that we had. And as you mentioned, the key thing is that how many or what fraction of the many, many things that we experience during the day will be deposited for long-term storage. And how does that work in the brain? What determines what gets deposited for long-term storage? What physiologically is going on there?

There are structures in the brain. hippocampus is known about to be a critical structure in the brain that forms our episodic memories. The mechanism by which it happens is a particularly interesting pattern called hippocampal sharp wave ripples. Now, this hippocampal sharp wave ripple is a fast oscillation, and it lasts for about 100 milliseconds, and it is the most synchronous pattern in the mammalian brain, at least, and it's a great tool to reach the vast part of the neocodex and tell something to the neocortex. Now, what is it telling?

These sharp wave ripples are representations, if you want, of what happened to us during the day, but the discussion that we are having with each other now will be chopped up into 100 milliseconds segments mixed together and mixed with our previous experiences.

And that's called the sleep consolidation. This is what we have known. So the question was, what process in the brain, what is the mechanism that allows to put tags to our experiences, they say, one goes to trash can, the other one could be retained? And that mechanism, of course, it sounds like a decision-making process. And for very long, we thought there is a, just jokingly speaking, there is a little man or homunculus that decides for us. You mean there isn't?

There isn't one? Well, there is one, but it's not a little man. It's called the Sharp Wave Ripple. There are three steps. The first one is experience, and it's a much slower process. It goes in real time. And then when the brain switches from this exploratory attentive state, that allows this other algorithm to kick in,

and this is when the sharp phase occur. And if they occur, and depending how many of them occur in the waking state, then those experiences in the visceral state, of this sharp phase will be, will have a priority of replaying that part of the daily experience what we have during sleep. Now, can I say to myself, oh, you know, I'm having an experience. I want to remember this wonderful conversation I had or here in a beautiful countryside. I don't want to remember that. Can I activate these sharp wave ripples? Not really. Sharp waves is the opposite of your.

will, if you want. Shalfaicil occurs exactly when you don't think hard, when you are not consciously attending to things. These happen in the absence of our effort. But you can try to maintain your mind throughout this conversation. If your brain is not allowed to switch into the sharp wave ripple stays, you don't learn a lot. For example, if you take amphetamine, which I did when I was a student, it's known that all known subcortical. neural modulators such as acetylcholine, norapinephrine, serotonin, they all suppress sharpways. So, then you maintain this attentive state, but you don't switch to the other state, you will not remember anything later.

So it's true then that taking a break from learning something and then maybe going back to it can actually help you? Exactly. You got it right. Just coming in and coming in and coming in is not enough because there is a phenomenon called the interference. So if you go to a French class after the Italian class, then most of the knowledge of the French class will be erased. It's much better to have sleep or idling brain state before that. Take a walk. Do you actually have to be asleep to do this, or can you just be

relaxed? The biggest change, of course, that we have is sleep waking. But the other changes happen all the time. You cannot attend a one-hour lecture to be attentive all the time. But you said that the Sharp waves will be reactivated during sleep. Is that only during sleep? I'm trying to understand if just the resting period also reactivates those waves in the brain, and that's why you remember it. So if the resting period is previously preceded by something important or some experience, then that particular experience will be marked also, and that will be replayed during sleep.

So there are many of these events occurred during the day, and those marked events are the only ones that will have the chance to enter in long-term storage because the unmarked ones will be not retained. So if you would allow me to record from your brain, from the depth of the brain, and I would see whether you have sharp-face after some of the sentences that I utter, and I could tell with some precision which parts of the conversation you would remember. Does that mean if I don't get sleep at night, I'm not actually being able to consolidate some of those experiences? Correct. That's why that's important. Yes. Sleep has a lot of important processes, but one of them is making memories.

And how can we use this now for people who are not good at making memories or have, you know, memory problems? Is there something useful here? Well, every psychiatric disease is associated with some kind of memory problems. Every animal model that has been investigated from autism to schizophrenia to Alzheimer's disease has altered sharp waves. So we have known that. Then how can we exploit this observation that we just made?

We can perhaps manipulate sharp phase and it can be done with drugs and so on. the easy thing at the moment is to erase them. There are several subcultical neuromodulators that have corresponding drugs and we can erase memories. Is it useful? Well, I can think of one disease, which is post-traumatic stress disorder, PTSD, where you don't want to remember things and maybe that will be an immediate avenue, potentially useful application. So where do you go now? What's next for your research? There are many, many ways. One of them would be going into this applications. The other one is manipulate the sharp phase, manipulate the dopaminergic system, for example, which is also known to be related to reward, how reward is facilitating the occurrence of the sharp phase, why the switches

are important, which brain systems are involved more and less in this process. We can manipulate them separately. So these are directions that will entertain us for the next several. years. Well, I hope you'll get entertained and come back and talk to us more about it. I would be happy to. Thank you for taking time to be with us today. Thank you. Dr. Yuri Busaki, Professor of Neuroscience at NYU Health based in New York City. And if you want to learn more about the brain, join Science Friday's Free new Hack Your Brain Neuroscience program, which kicks off on April 15th. You can solve fun puzzles in our family-friendly escape room as you investigate the mysteries of the mind, just go to science friday.com slash hack your brain to sign up.

That's science friday.com slash hack your brain. That's it for today. Lots of folks help make the show happen, including Felissa Mayors. Kathleen Davis. Jordan Smudjick. Charles Bergquist. Tomorrow, how trees help cities cool and why it's important to preserve urban tree canopies. Catch you next time.

I'm sci-fi producer, Shishana Bucksbaum. Thank you.