The winners of the 2014 Kavli Prize in Neuroscience – Brenda Milner, John O’Keefe and Marcus E. Raichle – discuss what led them to study memory and cognition and the challenges they faced in getting their discoveries about the brain accepted.
UNDERSTANDING HOW THE BRAIN PROCESSES INFORMATION, forms and retrieves memories, or establishes a sense of space in a new environment requires the overlap of several disciplines from psychology to molecular biology to neuroscience. The merging of research in these various fields coalesced into a new discipline that we now call cognitive neuroscience. This year’s Kavli Prize in Neuroscience laureates have made fundamental contributions to this vibrant field of research:
- BRENDA MILNER– Dorothy J. Killam Professor at McGill University's Montreal Neurological Institute and professor in the department of neurology and neurosurgery at McGill University. Dr. Milner is best known for work with an epileptic patient known as H.M., who had become unable to form new memories after having parts of both temporal lobes of his brain surgically removed. Her more than three decades of work with H.M. established that people have multiple memory systems, governing different activities like language or motor skills, opening the way for a greater understanding of how the brain works.
- JOHN O’KEEFE– Professor of Cognitive Neuroscience in the Department of Cell and Developmental Biology and Inaugural Director of the Sainsbury Wellcome Centre for Neural Circuits and Behaviour at University College London. Dr. O’Keefe’s research identified neurons in the hippocampus, called place cells, which respond selectively to an animal’s location. The discovery of place cells suggests that the hippocampus functions as a cognitive map, enabling animals and humans to establish their space and navigate in the environment.
- MARCUS E. RAICHIE– Professor of Radiology, Neurology, Anatomy and Neurobiology at the Washington University School of Medicine, St Louis. Dr. Raichle’s research led to the development and use of positron emission tomography and functional magnetic resonance imaging, tools that have enabled scientists to safely and non-invasively study the living human brain and track and record its function in health and disease.
The Kavli Foundation recently held a conversation with the new laureates to learn what led them to study memory and cognition, the challenges they faced in getting the neuroscience community to accept findings that often went against the conventional wisdom of the time, and where they see cognitive neuroscience as a field headed.
MARCUS RAICHLE: I didn't start my career with an interest in cognition. I trained as a neurologist with Fred Plum at Cornell Medical Center in New York. In Fred’s lab I learned how to measure metabolism and circulation in the brain, and this was driven by the notion that metabolic disorders of the brain are among the most treatable diseases. The issue was to understand that. At the same time the idea of using these kinds of measurements to understand how the brain is organized was emerging, largely driven following Seymour Kety's work at the University of Pennsylvania and by David Ingvar at the University of Lund in Sweden and Niels Lassen at Bispebjerg Hospital in Denmark. They arranged probes about the brain and looked at how the brain did things in terms of its blood flow in all sorts of cognitive tasks.
Marcus E. Raichle, Washington University School of Medicine in St. Louis, US
Following a stint in the Air Force, I was approached by a physicist at Washington University in St. Louis named Michel Ter-Pogossian, who pioneered the use of short-lived radioactive cyclotron-produced isotopes in biology and medicine, which to most people was completely novel. In fact, most people said, why would you be interested in that? But I thought as a lark, I would go to St. Louis and have a look at this. I hadn't been there more than 18 months when computed tomography, or CT, was invented at EMI Laboratories in England.
In Ter-Pogossian’s lab I was the lone neuroscientist in a group of physicists and chemists, and the first thing they wanted to do was build a better CT scanner. In a matter of months that gave way to the notion that maybe we could use these cyclotron isotopes to create an imaging device that could essentially do autoradiography of the human brain non-invasively. This became the PET scanner. The first machines were built outside my office, and it became my job to figure out how to use these machines to study the brain. So in the 1970s we began to figure this out. By the time we got to the '80s it occurred to us that the idea of measuring blood circulation and metabolism to understand the brain that people like Kety, Lassen and Ingvar had been putting forward could now be done in humans rather neatly. So we got started on this and figured out how to use the tool and all the little tricks that are now considered commonplace.
One of the great benefactors of our work was James McDonnell of McDonnell Douglas Aircraft Company. His requirement in all of this was we had to have a decent psychologist involved. As luck would have it, that ended up to be Mike Posner, who was and is one of the leading cognitive psychologists in the world. That just fit like a glove with the imaging project because the whole issue of dissecting behaviors was at the root of what Mike and his colleagues were doing. We proceeded to up the ante in terms of sophistication because now, of course, you are talking about language and memory and we could actually begin to take those things apart.
TKF: John, you also began your career in a completely different field, correct?
John O’Keefe, University College London, UK
JOHN O’KEEFE: Yes, I was first interested in engineering. I studied aeronautical engineering at New York University in the evening and worked at Grumman Aircraft on Long Island in the daytime making airplanes. While I was taking the engineering degree at NYU, I used to moonlight a bit and take philosophy and psychology courses, and I began to be much more interested in the whole area of the philosophy of mind and the way in which psychology tried to explain some of our cognitions and related behaviors.
I then went to City College in New York City to study full-time, where I met two very influential teachers, Phil Ziegler and Danny Lehrman, who were both interested in bird behavior. They really got me excited about the possibilities of trying to understand the mind and behavior in terms of brain function. I was then very lucky to get a post to do graduate work at McGill University, which in those days was one of two or three Meccas of brain and behavior research. For my Ph.D. work I studied a part of the brain called the amygdala, which controls many innate and learned behaviors including the fight or flight response in humans and animals, and I developed many of the techniques for recording single neurons in behaving animals that I subsequently refined and worked on when I went to University College London.
A few years after arriving at UCL, I decided to shift my focus of attention to the hippocampus and its relationship to animal behavior. We took an ethological approach and monitored the activity of single cells during many different natural and learned behaviors. After many months of looking and listening for correlations between lots of different behaviors and cell activity, I began to realize that the major correlate was not what the animal was doing, whether it was eating or exploring an object or carrying out a simple tasks such as pressing a lever to get food, but something about where it was doing these things in the environment. And I remember one day it dawned on me in a flash that the important correlate was the animal's place or location and that must be what the cells were coding for. And I must say that was a rather thrilling moment.
TKF: Brenda, you also came to study memory and cognition after trying a different field.
Brenda Milner, Montreal Neurological Institute, McGill University, Canada
BRENDA MILNER: In England before World War II, I went to Cambridge to study mathematics. After a year I decided that I wasn't going to be a great mathematician, so I switched to psychology. And I found I just liked psychology from the very beginning. I liked the empirical side of it.
However, I was not fascinated by memory when I was an undergraduate. I was fascinated by visual perception. My interest in visual perception continued after the war, when I went to Montreal and had the opportunity to study with Donald Hebb at McGill University. Wilder Penfield, founding director of the Montreal Neurological Institute, had allowed Hebb to send one graduate student to the Neuro, as it is known, to study his patients. At that time I was teaching at the University of Montreal, and I was older than the average graduate student. One day, he asked me if I would like to go to the Neuro and I immediately took to it like a duck to water.
At this point I was still interested in visual perception, not memory. But when I went to the Neuro I met these young epilepsy patients who complained of memory problems after undergoing operations that removed part of the brain’s left temporal lobe. So, this is a piece of obvious advice to give to students: If the patient complains of poor memory, don't say, “Well I'm not interested in memory, I've come to study visual perception.” You're not going to get anywhere if you do that. The patient may be wrong, but you have a scientific obligation as well as a clinical one to investigate that complaint. So, I started working on memory because of these patients with memory problems.
The first real memory loss I observed associated with bilateral medial temporal-lobe damage occurred in two of Dr. Penfield's patients whom I studied in Montreal, and who had this memory loss after a unilateral temporal removal that included the bulk of the hippocampus. And this was so different from our regular series of patients with such problems that Dr. Penfield was really very concerned, because this was elective surgery and you don't do elective surgery if there's a risk of serious memory loss – it’s obviously much better to have epilepsy than to lose your memory. The memory loss was unexpected, and Dr. Penfield and I suspected that the surgeon had operated on the wrong side of the brain, and that there was some damage in the corresponding region in the opposite hemisphere, so that we were seeing the effect of a bilateral lesion. We presented our findings in 1954 at the American Neurological Association meetings in Chicago. And it was after that that Dr. William Scoville, the neurosurgeon from Hartford, CT, called Dr. Penfield and said, “I am so interested in this abstract because I think the memory loss you are describing is what I have seen in my patient” – and it was H.M. – “in whom I did my operation.” Dr. Penfield sent me to Connecticut, and I began my studies of H.M., which lasted more than 30 years.
TKF: John and Brenda, you both trained at McGill University with Donald Hebb, and John, you have credited Dr. Hebb with fostering an environment that gave students “the freedom and the opportunity to test their ideas” about memory, cognition and perception. I’d like to ask each of you to describe Dr. Hebb’s influence on your careers.
MILNER: As I said earlier, Hebb sent me up to the Neuro, which was a very nice thing to do. I think the impressive thing for me about Don Hebb was he didn't give me any guidance once I went up to the Neuro. All Hebb said to me was, “Make yourself as useful as you can, and don't get in anybody's way.” That was very good advice, and I tell my students to be as helpful as they can.
And that was really all the guidance he gave me, except when I wrote my thesis, which I was quite proud of. I gave it to Hebb, and he gave it back to me and said, “I can't read it!” I was so disconcerted, but I managed to change it to something that pleased him. Hebb cared about how you wrote. I would say what I learned from Hebb was how to write and that I try to teach my students that, more than how to do research because he gave me a free hand with that.
O’KEEFE: As I said in that piece you just quoted, Hebb really believed in giving students the freedom to explore, to try out new ideas, new techniques. And we were given the freedom to make mistakes and not be punished for the mistakes. When I got to McGill, it was a hotbed of excitement, with faculty and students generating lots of new ideas. And many of them derived from Don Hebb’s own thinking. I give him credit for having laid the foundations for physiological psychology.
What Don Hebb did was present a way in which we could think about representations in the mind – perceptions, thoughts – in terms of neural circuitry. That now seems like something that we pretty much all accept, but in those days it was not at all accepted. It was really a startlingly novel idea that you could make a connection between circuits in the brain and the way in which the mind operated. I was very taken with that and it certainly inspired a lot of ideas in me and led me to many of the experiments that I subsequently carried out.
TKF: The history of science is filled with discoveries that were met initially with skepticism before they gained acceptance. I’d like to ask each of you about how you reacted to the skepticism of your findings – Brenda, your findings on memory, John, your discovery of place cells, and Marcus, your discoveries regarding blood flow and metabolism in the brain – and whether any skepticism persists.
MILNER: There was skepticism at different stages. The first thing I was finding was that some patients with visual-perception deficits had lesions of the right temporal lobe in the right hemisphere, which is the nondominant hemisphere for speech. I can tell you I had a lot of trouble getting this accepted because it was always, “Well, probably the surgeon makes a bigger removal on the right side of the brain” or “Are you sure it's not due to a visual-field defect?”
I do realize why there was skepticism. We were observing only a few patients with these problems, and there's always the argument, “Well, how can you be sure there isn't something else wrong in the brain?” And we couldn’t be sure because this was before the newer imaging techniques were invented and we couldn't see the patient's brain outside the operation site.
The skeptics also said, “Give us an animal model, and if we could see the same thing in a monkey with those brain lesions, we'd begin to believe.” And of course people started looking for an animal model, and that was a failure for ages. Lots of us tried to think what were the right kinds of tests to use with monkeys and rats to show the equivalent deficit. And it was not obvious at first. It's amazing to look back. It took about 17 years until Mort Mishkin and his colleagues at the National Institutes of Health were able to show something equivalent in monkeys with equivalent brain lesions to those in our patients. But, until there was an animal model, people were skeptical because the patients were rare and we couldn't prove exactly what was wrong in their brains, just a description of what we'd done.
In response to the skeptics, I tried to convince them, but I wasn't offended by the skepticism because it wasn't a personal attack. I think it's a good scientific attitude to say, “You have to convince me.” It's fair enough.
TKF: John, what can you tell us about the skepticism to your findings?
O'KEEFE: I'd have to say there was quite a bit of skepticism on several grounds when we first started reporting that we could show very specific correlates to the activity of cells in the hippocampus, and specifically, that they were related to where the animal was in the environment. The hippocampus is deep in the brain and far from brain areas responsible for receiving sensory signals or producing motor movements. I think many people, including leading physiologists, thought that it was just not going to be possible to show correlates of activity at the single cell level in brain areas that were so far away from the periphery. I think they accepted you could do that in the early visual areas of the brain, but to go so far away from the early stages of sensory processing and look at neural correlates in the hippocampus raised a lot of skepticism.
More recently, there has been skepticism related to our claim that the hippocampus in animals is exclusively a spatial system. That came partly from some of Brenda Milner’s observations that the memory deficits in patients like H.M. belong to a much broader class of memories than just the purely spatial. We've worked hard to see how we could bridge the gap between the idea that the hippocampus is a purely spatial memory in the rat to its broader function as an episodic memory system in humans, which is the memory you have for something you did at a particular time and place. One idea here is that humans have a sense of linear time – the notion that there is a past, present, and future – and that this becomes incorporated into the spatial mapping system to produce a spatiotemporal mapping system, a way of localizing oneself in time as well as place. I think it's only very slowly – and I don't even want to claim we've won the argument completely yet – that neuroscientists have begun to accept that there is, at least in animals, a purely spatial function for the hippocampus and that in addition to its preserved spatial function, this could form the basis for an episodic memory system in humans.
TKF: Marcus, was your work met with skepticism?
RAICHLE: It sure was, and on more than one occasion. The first episode really caught me by surprise, both the finding and then the reaction to it. The conventional wisdom at the time – which dated more than a century – was that blood flow in the brain seemed to follow function. In other words, if blood flow increased in an organ that's very expensive and very dependent on oxygen, for example the brain, it must be related to delivering more oxygen, in order for that organ to carry out whatever function it is planning to do. The evidence for that was very, very limited, and in fact, was based on a study my lab published on a single patient with Alzheimer's.
I had the sense this was pretty flimsy evidence, so we conducted a routine study to simply confirm what everybody suspected about oxygen consumption increasing with blood flow. And to our utter dismay that is not the result we found. The blood flow went up, but the oxygen consumption didn't. We published this first in Proceedings of the National Academy of Sciences and later in Science, and the pushback was really quite remarkable. Other scientists just did not believe our results, because they went against conventional wisdom of blood circulation and metabolism in the brain. The accusations went so far as to suggest we were doing bad science. Tentatively a few other labs ventured into this area, and lo and behold found the same result.
Ultimately, this discovery provided the physiological underpinning of functional magnetic resonance imaging or fMRI by explaining the origin of the blood-oxygen level dependent, or BOLD, signal used for this work.
More recently, we showed that when we place people in brain scanners and ask them to do nothing, their brains are spontaneously activated in a highly organized manner during this resting state, which we later called a default mode of brain function. A prominent feature of this activity became known as the brain's default mode network, a central component to the brain's organization, or default network of the brain. Our first report of this immediately brought concern in the scientific community that we hadn't really controlled for what people are doing when they are not doing anything. (It would have been a bit unusual for, in this case, 134 people to all do the same thing, but that argument never really surfaced.) At one point we submitted a paper that again brought this forward, and it was summarily rejected. We spent the next two and a half years nailing down the evidence for what we had observed. We received a ton of criticism along the way to the acceptance of this concept, but now I would have to say the default mode almost dominates the field of functional brain imaging.
TKF: All of you have made seminal contributions to the field of cognitive neuroscience. I’d like to ask each of you to tell me what you see as the most intriguing developments on the horizon for this field. Marcus, can you give us your thoughts?
RAICHLE: I think one of the most important things is what cognitive neuroscience has done --and this goes back to people like Brenda. Cognitive neuroscience brings human neuroscience into the much broader arena of neuroscience. The human brain is being looked at in an expanded manner thanks to the tools, especially imaging, we have for its study. What we are learning from these studies adds important knowledge about the human brain in both health and disease. I'm really intrigued by the application of cognitive neuroscience, defined broadly, to the understanding of disease and how we think about it and treat it.
Another thing I will add, though, is cognitive neuroscience doesn't do this alone. As Brenda and John pointed out, we are deeply, at least in my work, dependent on areas of inquiry such as neurophysiology, developmental neurobiology, basic cell biology, and genetics to help us understand the interesting patterns that we are uniquely in a position to observe. Integrating across levels of analysis should be a guiding principal as we move forward in neuroscience to understand ultimately the human brain. I am very excited about where all this is going. I just feel lucky I'm around to be a part of it.
O'KEEFE: I would concur with Marcus. We're getting to the early stages of understanding how the brain works at the neuronal network level. And this applies to other areas of the brain as well as the hippocampus, where we are beginning to ask questions about how these areas become dysfunctional in mouse models of Alzheimer's disease. We already know, for example, that place cells are less able to identify the animal’s current location in some of these mice. So we're just beginning to be able to use our basic understanding of these brain areas to study aspects of important and distressing human diseases.
In terms of basic neuroscience, I think we're entering a really exciting era where recent technical developments will enable us to look not just at the activity of single neurons, but also at the way networks of neurons work together. We are arriving at the cusp of an era where we will have the ability to look at thousands of cells and monitor their activity in animals that are doing things in real environments or, more powerfully, in virtual reality environments. This is going to enable us to look at the way networks represent not only the animal's current location, but what it's thinking, how it's feeling, what it’s intending to do.
MILNER: If I stay with my own level of interest in the hemispheres of the brain, I'm very interested in what we’re doing now, working with people with normal healthy brains. And I am particularly interested in exploring the relationship between the left and the right hemispheres of the brain. We're such a language-dominated culture, yet there are so very few things that are purely language tasks. I’m interested in how and in what way do those nonverbal processes communicate across the brain’s hemispheres.
Cognitive neuroscience is a wonderful field to get into today, and I feel grateful to be here and still able to carry on. Marcus talked about that, but I'm really very, very lucky to be healthy.
RAICHLE: You're an incredible model, Brenda!