I will give an overview of some of my research on human memory, starting with an emphasis on the distinction between memory capabilities that allow for conscious experiences of recollection versus implicit memory functions that apparently unfold in an unconscious manner. I will then describe research on memory processing during sleep. Our work has made use of methods for modifying this memory processing during sleep through sensory stimulation that does not cause awakening. Finally, I will cover our recent studies in which we engage sleeping individuals in a two-way dialogue during their dreams. The many opportunities for bridging to Vajrayana practices are intriguing and worthy of discussion.
Speaker Bio: Ken Paller
Professor of Psychology, Director of the Cognitive Neuroscience Program, James Padilla Chair in Arts & Sciences, and Director of the Training Program in the Neuroscience of Human Cognition Northwestern University, Evanston, Illinois, USA.
- Thank you Maria for the invitation to this workshop. I’m looking forward to the upcoming discussions because I really enjoy opportunities to talk with people about the issues at the intersection of Buddhism and neuroscience. I’ve enjoyed that previously at events like the Mind and Life summer research institutes that I attended a bunch of times in the early days. I taught neuroscience through the Emory Tibet Science Initiative in India, Drepung and Ganden monasteries. And also there was a workshop in Berkeley put together by Cliff Saron and Bill Waldron, where David Germano and I first started talking about Dream Yoga.
- Now, my own research has looked at the neuroscience of meditation only a little bit. And yet the Buddhist science dialogue has influenced my work and my thinking, as you’ll see. So I’ll start by telling you a bit about where I’m coming from in my neuroscience research. My main science interests and driving forces to understand something about mental experience, the capacity that seems to stand in such contrast with the physical world as we know it. My and my training in neuroscience started at UC San Diego. When I arrived there for graduate school, I was trying to figure out how to learn to be a neuroscientist, but still to get to focus on subjective experience. Something regarded by some is beyond the territory of neuroscience. So what is the territory of neuroscience?
- Well, some of the same people who started that field also started the neuroscience program at UCSD. They founded the first Department of Neuroscience there and before that, there were the separate disciplines of anatomy, physiology, pharmacology, biology, pathology, and so on. And the origin story is that they realized that this was producing barriers to the work of understanding nervous systems and their functions. So they wanted to remove those barriers by creating a new field that would have space for more interactions among all the scientists studying nervous systems and things. So that’s how we got this relatively new field called neuroscience.
- One of the things I like about neuroscience is that it can be open to including anything that will be helpful for the goals but really more important than preserving the disciplinary boundaries that are reinforced in universities is working towards a better understanding. And neuroscientists have many different goals, of course, and there’s a big tent of neuroscience so we can take on questions about how brains develop environments influenced that development, how social and cultural factors have their impact, and even phenomenology can be part of neuroscience, and neurophenomenology, as Francisco Varela a and others realized. Participants in a workshop like this one might have different goals and methods, different things we want to understand different ways of going about our work. So when we’re talking about neuroscience, I hope this big tent metaphor seems acceptable, where we can all work together, we shouldn’t think of neuroscience like a dominating colonial power, where neuroscience can convert or gobble up any field by just appending a neuro to the front of it.
- We’re potentially facing this challenge of bridging the fundamental science humanities divide to cultures. And with some effort, maybe we can avoid neuro-hegemony. We can be collaborative and admit that there are many places where we can ask questions and look for answers. So my inclination is to focus on scientifically understanding the human mind and including these ideas of enhancing cognition and creativity. So let’s use whatever strategies and perspectives can be helpful. I’d like to start with memory. And in grad school, I ended up working in the lab with Larry Squire, an expert on memory and amnesia. And I found the science of memory particularly interesting because we got to talk about consciousness in the sense of distinguishing between conscious memory versus unconscious memory. And that way we gain a foothold on what’s special about conscious experience. When we ask what happens in the brain to allow us to consciously recollect a fully immersive experience of some earlier event versus a memory for something that remains implicit. Well, we may not even know that we know that. So let me try to explain that and I’ll start with this story about Clapper Ed.
- He was a Swiss neurologist, and he would see a patient every day and she was very amnesic. And so she would never recognize him. It was as if she was meeting him for the first time every time and one day he went to greet her with his hand to shake and in his hand he had a hidden pin, and she shook hands with him and it hurt. And then she tested her the next day to see if she could remember that. And he reached out to shake her hand and she refused to shake. And she didn’t say anything that gave away a memory she just said why I don’t want to shake hands. So it seems that she had some memory for that prior episode, but didn’t really know she had it and didn't know why she didn’t want to shake hands. And so that was one of the first step demonstrations of these different types of memory.
- And since then, we have studied amnesia a lot. And so we call this a case of search circumscribed amnesia where the patient has a very restricted set of deficits that are only on a certain type of memory. We call it declarative memory. That’s the recall and recognition of facts and episodes in their various categories of preserved memory function in these patients. Including this category of implicit memory, things that they don’t know that they know. And that can include skills and habits, conditioning, and something I’ll tell you about called priming. And so we have this very interesting contrast between some brain damage that leads to this syndrome of amnesia that disrupts certain types of memory that allow us to consciously recollect prior experiences but leaves intact various other types of memory.
- Now we can study that in a lab these days, and we don’t need to inflict pain on people. Instead, we can show them some words and ask them to memorize these words. And if we do that, we can then give them various types of memory tests. And one is this implicit memory test known as a perceptual priming test. And we show words very briefly, so the word piano might show on the screen just for a small number of milliseconds and we can adjust that to make it so difficult that people can see the word only about half the time. And in this experiment, when we did that, we found that their ability to read the word different systematically depending on it, whether that was one of the words they’d seen a few minutes earlier in the study phase, and so amnesic patients had this effect, they were more likely to read a word correctly if it had appeared earlier. And yet if we tested their memory with a recall or recognition test, this is a three choice recognition test. They have to pick which word they saw before they’re very bad at that, as would be expected. But they’re quite normal at the priming tests, and they can read words more accurately if it’s a word they saw before and something else interesting because we also asked them to rate how quickly that word flashed on the screen. Was it very quick, or even quicker than that, you know, really, really quick and they had to use a scale to say how quickly they thought the word had flashed. And even though the times didn’t differ, everybody thought that the words that had been seen earlier in the study phase had been on the screen longer.
- In fact, they maybe were more efficient in reading that word, because of the priming. So they inferred, perhaps unconsciously, that that word had been on the screen longer. And so this is another priming effect present in both the patients and the controls to about the same extent. So it seems like we have a way to measure memory of this implicit variety that shows the MEC patients have this quite preserved memory when it comes to implicit memory in these certain specialized tests. So going back to the list, there are these different types of memory we’ve been able to study and try to understand how they work in the brain. And it gets a little bit more interesting because sometimes, this implicit memory influences our conscious experiences and our conscious memories. So there is some crossover. And so what’s interesting is that you know, scientifically, we like to study one type of memory and study it uncontaminated from the other types of memory and amnesia allows us to do that we can see this implicit memory operating, even though the patients can’t remember that they tried to memorize that list of words earlier. And we think that’s of course happening and all of us and it takes a little more effort to disentangle it.
- But we want to study it separately so we can understand how it works. And yet, when we go back to people in real experiences in the world all these types of memory are working together kind of interactively, so there’s a lot more interaction. And so the way I might think about it is that sometimes you can look at something and know or have a strong feeling that you’ve seen it before but not be able to remember the details and we’ve studied this with these Kaleidoscope images and show that some people sometimes people can guess really accurately because of their memory. That’s only implicit and they don’t consciously remember having seen one of those before. Maybe you might remember you did actually see one of those before on my title site. And you could perhaps guess accurately about which one it was that’s our experiment. But, but moving on that we can we can think about how studying these different types of memory is an example of what scientists and people are always doing. We’re trying to study one thing at a time and understand how it works in isolation and sometimes that’s called carving nature at her joints and hopefully we can do that and understand what what are the different types of memory, how they work, but there’s some violence in doing that. So follow through on the metaphor. It may be misleading a little bit because we have to put the parts back together and see how these different memory systems interact in various circumstances where they’re not just isolated one type of memory, but there, there’s interdependence between them. And so that’s an interesting part of understanding how memory works and understanding that were many things.
- Another question you might ask of course, is, is that implicit knowledge always implicit? Or maybe there are practitioners that show different results and we could look at different people that that sort of remains to be an interesting question. Now, if we look at the big picture here, you can imagine there’s these two types of information, explicit knowledge and implicit knowledge, and they influence what we do. And you could say our thoughts, our decisions or actions can arise from one or the other of these types of stored information that we have. And then when we do that, we of course, influence the storage because, you know, these memories in our brain are malleable, and they keep changing as we use them every time we use them. And so these arrows going back show that there’s sort of a potential tendency to go ahead and change as a function of our thoughts, decisions and actions. You can think about this as karma in one sense that you’re going back and changing things. And then these two types of memory are not really quite so separate, perhaps. Perhaps they interact a little bit. So maybe there’s some interesting crosstalk between these differences, memory systems that are storing implicit and explicit knowledge. I’m going to leave the memory realm now and or expand the memory realm rather, and talk about how it connects to sleep and eventually to dreaming.
- So as we’re storing memories, we have countless memories every day. And yet, we forget most of what happens each day. You know, at the end of the day, you can remember a ton of things that you did that day, but try to do that for what you did 10 days ago and it’s pretty impossible to remember exactly the same level of detail. So we’re, we’re discarding a lot of memories. We’re keeping some, what is it that allows some of that information, some of those memories to be kept. Whereas lots of memories are just fading. And we’re not remembering everything we learned from every moment of the day. And so, memory researchers have talked about a process to try to explain that which is called consolidation. It’s really a continuation of learning, you know, when you take some information on your computer and you drop a file onto your computer’s hard drive, it’s there and you can go get it later, but our brain so quite works. That way. If we want to add information, we can’t just drop it in somehow and expect it to be there. But we have to be sure that information is going to stick. And this consolidation processor really depends on gradual changes in memory storage in the networks of neurons and their connectivity that hold that information that allow us to use it later. And so how does that happen?
- Well, one of the probably biggest steps is rehearsal. We have to continue to practice what we want. If we want to remember the names of new people we meet, we need to use that and sort of practice recalling those names so that we can then do that when we need to. Personal isn’t. It can happen intentionally when we try to rehearse when we’re trying to practice, it also happens unintentionally at other times, and one of those times is probably sleep. Again, what happens in this consolidation process? Is not just taking new information and making it solidified. But that’s not how memories are stored in the brain they’re stored more interactively so, new information has to come in related to other things we know there’s this integration of new information with existing information that happens as part of consolidation, as is generalization where we might lose some details but maintain some general principles we’ve gained from many different events. We also transform the memories and we have a basis for creativity as relate. We relate new things to different things we’ve learned earlier and put things in new combinations. So that’s all part of consolidation. But an interesting idea that hasn’t always been discussed by memory searches is that maybe this memory consolidation happens in part during sleep. And so that’s the idea that one of the benefits of the many benefits of sleep could be that we take new ideas with new facts, and we’ve acquired during the recent times, maybe that day, and integrate that information with other things we know so we can hopefully have use of the things we’re going to need in the future.
- Now, this might be called Sleep Learning, in a sense, I guess I’ve called it that in a paper I published in Scientific American, mainly because the editors really wanted that in the title. But sleep learning is a little misleading because it could mean learning new things in your sleep, which is interesting too, but here we’re really talking about learning happening and then continuing on through this consolidation process, during sleep and that’s that’s what I was going to talk to you about today. I wrote another paper in 2018 for a really cool journal called Frontiers For Young Minds. It’s targeted at kids from, you know, eight to 14, science education and so forth. And so I was able to come up with a more conventional title, rather than the one the editors wanted in Scientific American. So I picked this other title. Do house elves clean your brain while you sleep and so this is a fun little idea that, you know, they’re little creatures, essentially, your brain, organizing your memories while you’re sleeping. Maybe that’s a nice way of thinking about what happens that new memories have to come in. And if you want good use of your memory system you have to really keep things organized and get rid of the things that aren’t needed and have, you know, a nice order of connections between things so that you can access things later. So that's how I think about memory consolidation during sleep. Now, how do we study that?
- Well, in my lab, we can have people come in and learn something and then have a little nap. And so here’s the procedure for what we do to record their brain activity while they’re sleeping. People wear this cap with a set of electrodes on it and then they can go and lay down in a comfortable bed and sleep while we’re recording their brain activity and we’ll look at the brain activity and then be asking, Well, what does that mean? What’s the brain doing during sleep this whole time continuing to be operative? And how is that brain activity during sleep useful for our waking activity? That’s a big question. And here’s one of the ways we get that. This is a procedure we use called targeted memory reactivation. In these studies, we have people come into the lab. The first thing they do is they learn some things, in this case, some spatial information, so they learn where objects are located on a screen. All 50 objects are in a random location. They have to memorize exactly where each object goes. And each object comes with a sound like this cap. And we’ll test how well they did and how well they can place each object where it goes. And then the next step is some sleep and during sleep, we do this TMR method where we’re presenting a subset of the sounds they heard during learning again, when they’re in slow wave sleep, with the idea that these sounds can then reactivate the information they learned in this case spatial information about where objects are located.
- Then, in the last step, people wake up, they get tested again, they have to say where was each object located? We measure how accurately they do that. And what we have found repeatedly in many experiments is that, well, there’s forgetting but there’s less forgetting for the objects are related to the sounds we present during sleep compared to the other objects if we don’t present the sounds and so there’s some benefit, which we think is due to these sound presentations during sleep, that they’re reactivating those memories so that those memories don’t decline. They aren’t forgotten to the same extent as other memories that don’t get reactivated. And we’ve seen this phenomenon many times with different types of memory tests. So we think it’s a way to try to understand what’s happening during normal sleep, when the recent memories from the day get reactivated and then integrated and go through this consolidation process. So this study by my students has been then repeated in fMRI with a lot of different methods to try to understand how the phenomena works. And again, the basic idea is that we’re strengthening individual memories by reactivating them during sleep by preventing these sounds and we’re not trying to do new learning, but building on the learning that people did while they’re awake were also to do this work, we had to escape the dogma that sounds get blocked during sleep. In fact, sounds can be processed by the brain even while we’re asleep and influence storage of memories in this case.
- And we’ve in these experiments, mainly focused on slow wave sleep, that’s a deep sleep stage that seems to be very relevant for memory processing. But we’ll also be looking at REM sleep in a moment. But now that from this point on we can sort of say sleep seems to help with memory storage, even without sounds presented. We can see some of the same neurophysiology is unfolding to allow this memory consolidation process to happen and we look at that to try to understand what those brain waves mean? And sleep physiology shows us some really interesting hints. There are things called slow waves that give us the name slow wave sleep, their sleep spindles, which are another phenomena and their set of different phenomena that we think are sort of hints to the neural mechanisms of this memory consolidation.
- Other types of memory also influence and let’s look at a type of memory that’s like a skill memory. And this is a type of memory that’s like, learning to play a melody. So here’s how it works. There’s video games. They learn to play and they press these four buttons with their left hand and they have to if you’re familiar with Guitar Hero, when a little circle crosses that white target that they have to press the corresponding button at the right time. When you do get to hear this wonderful little melody and produce it and as you practice get better and better. Learn that melody, one melody with the red dots, then another different melody with the blue ones. And so at the end of the training, they’ve learned how to play these two melodies. But then we randomly take one of those melodies and while they’re asleep, and slow wave sleep, we present that melody, and that reactivates the memory and the memory in this case, isn’t remembering what the melody sounds like. It’s remembering to press the right button at the right time when they see the visual circles. And that’s in fact, what we found is shown here that they’re there. They’re really better at both melodies after they wake up, but they’re more more there’s more of an improvement for the one they heard while they were asleep. Even though they don’t know they heard that they wake up and they’re unaware that we were presenting sounds during sleep and yet their memory storage is better as we can see by their better performance on this task. Another example I have for you is habit memory. In this case, that habit is how we’re influenced by our implicit social bias.
- And in fact, we can train people to reduce their bias a little bit and there’s a temporary benefit of that we can reduce bias based on gender and bias based on race with a little training that’s called counter stereotype training is basically being exposed to people that exemplify the opposite of the usual stereotype. So we reduce bias in that manner. And this study by Xiaoqing Hu and colleagues we showed that if you then remind people of that bias training during sleep, you can get an even better improvement or reduction, that bias for the cute bias, whereas the other bias is going back to the pre training level and people having the same bias stay locked in with and so that’s a way of reinforcing that type of learning. Again, this is more of an automatic, habitual type of training that you might train on during the day. And then your brain is still going to work on that during sleep to allow you to continue to to keep that habit strong in the ideal circumstance. And so taking together we have sort of evidence for lots of different types of memory and this is sort of a chart of the different types of memory that we think about in memory research. There are different brain mechanisms for these different types of memory, but perhaps many of them, maybe all of them depend on more processing during sleep to have those memories be strong and maintained and be enduring memories. So we’re using that to continue to try to study the neural mechanisms, but we’re also interested in applications. How can this be used? In the case of general learning maybe can be helpful for people acquiring a language or second language for expert skills that they want to get for changing particular habits you’d like to adjust, addictions and so forth, and perhaps in contemplative practice for continuing the practice you’re working on during the day in your sleep? That’s an interesting one to talk about in this context at our workshop.
- There’s some clinical applications we’re thinking about. We’re studying people that are recovering from stroke and trying to recover their motor function and asking, in addition to their rehab training, can reinforce that learning during sleep to have better outcomes from their rehab. We were thinking about other therapies that also depend on learning as well as some sleep disorders. And one more example I want to give you is with problem solving, because that connects to this issue of creativity. And let me show show you one experiment from that because it’s really you know, we remember we need our memory system to working not just so we can reminisce but also of course, so we can use our memories in the things we need to do, perhaps making decisions that are informed by what happens in the past but also being creative and solving problems that we have to solve. So here’s an experiment that was done by Kristin Sanders, a grad student in Mark Beeman’s lab at Northwestern. And they know about creativity and they study that a lot. And they’ve used a very large set of problems to try to ask people to come up with solutions and study how people get to the solution. And so here’s a problem. You see this triangle of circles and you have to move three of those circles to make the triangle point up instead of down and people get like a couple minutes to try to solve that. Maybe you can see the solution. I’m just going to give it to you. Here’s the solution. So some of these problems are very difficult. They don’t always solve them. And here’s the design of our experiments like our other experiments with these different components. Except, in this case, they’re not getting tested on their memory Exactly. But what happens in the pre-sleep period is they get a bunch of problems until they’re six problems that they haven’t solved.
- And now, each of those six problems is connected with a different kind of musical or sound background. So they have to learn which sound goes with which puzzle so that when they hear the sound, they know which puzzle it went with. And then they’ve also in two minutes failed to come up with a solution to those problems. So they leave the lab and they haven’t solved any of those six problems. And then overnight we use actually a wireless EEG system they wear around their forehead, monitors sleep, and is connected to a computer that presents three of those music sounds. So they hear three of the sounds associated with the three of the problems they were trying to solve. And the next day they come back to the lab, they get all six problems. They try to solve them. And what we found, as predicted, is that with those cues during sleep, we’re able to increase the probability of them coming up with a solution so they go from what like 21% to like 32% Correct solving the problem when they try again within four minutes. The next day to try to solve each of those problems. And so this shows that memories were activated during sleep, this time in the service of working through that problem and coming up with a solution that they were then more likely to achieve that next day when they came to the lab to try and solve the problem. So an interesting connection between memory processing, consolidation and creative problem solving.
- So let me move now my final few minutes to REM sleep. So another use of our TMR procedure is to provoke a lucid dream and so if people do a learning session where they’re learning is about trying to do a good reality check and carefully, mindfully think about the present experience and decide, am I awake right now? Or is this a dream experience actually, that I’m having? And they practice that in the context of sound, that then we can play again, when they’re in REM sleep to provoke that same type of processing and questioning, to get them to decide, “Well, actually, I am in a dream right now.” And then they’re in a lucid dream.
- And that seems to actually happen, so we can present a signal that reminds them of the reality checking they need to do carefully, and then they can signal this back. It was already known that signals could be made by lucid dreamers. So they signal back with eye movements going left, right, left, right, and when they make that eye movement, we see the signal in our recording from the electrodes quite clearly, it's a very distinctive signal. It’s not like what we would normally see in REM sleep. So that’s a pretty interesting use of the TMR method to provoke lucid dreaming. One more thing we’ve done and we reported in a recent paper was to go one step beyond that. So instead of just having a signal from the lucid dreamer, left, right, left, right, when they achieve a particular thing like realizing their dream, we asked whether we could communicate with the lucid dreamer and have a dialogue. Can we ask them questions and they could answer with their eyes moving signals and that’s impactful, we send what we found real time two way communication. My student Karen Konkoly is the first author and this is something that is collaborative with four other labs.
- Well, three other labs at least in Europe, and so a long list of authors and we all found something kind of like the following somewhat different ways so you can record the brain activity before sleeping, and show what it looks like when they’re awake and when they can make eye signals volitionally, and then they’re asleep. And with the instruction to make this triple left right signal when they are having a lucid dream. And so this was a student who had only had one or two lucid dreams prior in his life, and those were all in my lab too. And so then when we presented the sound, he could say, Yes, I’m having a lucid dream. And so here he is. This is two times he’s signaling I’m having a lucid dream. There were actually six signals. There were another four prior to this. So he was signaling I’m having a lucid dream. And then at this point, Karen was ready with the computer to press a button to present a random question to him. All the questions were math problems, so she presented this random math problem, what’s eight minus six? And it was her voice softly presenting it just like in our other TMR studies, except this was a soft question that demanded an answer and the sleeper that named Chris he was there sleeping, realizing he was in a dream, realizing Karen was asking that question, and he admitted the answer to answering the math problem correctly. And then again, at this point, Karen asked another math problem and got another answer.
- So we have multiple observations of this correctly answering questions. It’s a proof of concept, because we know what the correct answer is. And we can say yes, communication works repeatedly even though not every time we try, but we’re working on you know better procedures so we can get it to happen more often, but it can happen and that means that the door is open for now. More way more two way communication to probe exactly what people are dreaming about in real time, rather than having to wait until they wake up when they’re forgetting lots of their dream and they may have distortions in what they remember. We can ask them at the time they experienced to get more information. And as I said, with our collaborators, we use different methods, sometimes spoken word, sometimes beeps, flashes of light touching the arm, and we could have them express with eye signals or with muscle twitches, answers to questions we asked. And so a number of things were done. To just verify that these were correctly answered questions in this experiment. The next thing is, well, what else do we want to ask Dreamers? And what can they tell us about dreaming as we’re asking, you know, why are people dreaming and how is dreaming useful for waking behavior and how does it relate to memory processing, problem solving creativity, the kinds of questions we’ve been asking in our studies of slowly sleeping well, as well. So that’s where we are with that.
- And an interesting thing to bring up in the present context is something Alan Wallace alerted me to in these teachings. On the six Bardo’s there’s a little part of the book where the suggestion is made well to practitioner a little while after you’ve gone to sleep, our companion slowly arouses you and asks have you apprehended the clear light? So this is part of the instruction now, the usual instructions might be, have an intention before you go to sleep about what you’re going to do in your lucid dream. But the idea of actually presenting more instructions while they’re asleep is interesting because you can perhaps guide people through a lucid dream, and this opens the door to a lot of other scientific experiments that could be done as well as practical things. And so that’s some of the things I hope we can discuss. Is there some usefulness of having signals during dream yoga practices where there could be questions and answers, maybe a teacher and the student would want to communicate during lucid dream to move ahead with some of the instructions. During the actual time of the lucid dream rather than just afterwards and sort of planning for future ones. It could be part of a guided instruction. That’s an interesting question, I think. And again, are there connections between this type of processing of a dream and creativity and memory transformation and all the other sorts of questions we might want to ask that try to help us understand sleep and dreaming and consciousness.
- So let me stop there. And thank my students. My collaborators and mentors and funding agencies, and thanks for your attention and listening to the talk and I’m really looking forward to some further discussion later on in the workshop. Thank you.