An Intricate Network

Most of us can remember where we were during major events in our lives. Neuroscientists have determined this isn’t just an accident of memory; the imprinting of place in our minds plays a fundamental role in remembering events that take place in our lives. They’ve also discovered memories aren’t formed and permanently lodged in just one location in the brain, but rely on an extensive network of memory highways that reaches several regions. Among these regions is the entorhinal cortex, which is getting more attention by neuroscientists trying to decipher how memories are made and stored.

These are significant findings that may in turn may help neuroscientists develop ways to enhance learning, aid memory disorders such as Alzheimer’s disease, or guard against memory loss from aging. To discuss the new findings more in depth. The Kavli Foundation brought together three researchers who have been integral to advancing our understanding of memory:

• Bradford Dickerson – Associate Professor of Neurology, Harvard Medical School and Director of the Frontotemporal Disorders Unit and Dickerson Neuroimaging Lab, Massachusetts General Hospital in Boston. Studying brain structure and function, Dickerson's lab uses magnetic resonance imaging to try to understand the roles of various brain regions in normal human memory, and to investigate the locations and degrees to which brain regions are affected by disease, as well as how these changes relate to clinical symptoms and difficulties with the performance of cognitive tasks.
• Mayank Mehta –Professor of Neurophysics at the Brain Research Institute at the University of California, Los Angeles. Mehta's research has focused on computational and experimental investigations of learning and memory, seeking to understand how the brain learns and remembers how to navigate unfamiliar environments -- research aimed at paving the way to better understanding mechanisms of learning and memory in neural networks.
• Edvard Moser – Director of the Kavli Institute for Systems Neuroscience and the Centre for the Biology of Memory, both at the Norwegian University of Science and Technology. Moser's research has provided key insights into how spatial location and spatial memory are computed in the brain. One of these insights has led to an immediate revision of well-established views of how the brain calculates position and how the results of these computations are used by memory networks in the hippocampus.

The following is an edited and revised transcript.

THE KAVLI FOUNDATI ON (TKF): Dr. Moser, your work has focused on the memory of place and direction—or to put it another way, understanding how we know where we are, where we have been and where we are going. To begin, could you briefly explain how our location is tied to memory of events?

EDVARD MOSER: Location in space is fundamental to all kinds of memories of events and facts. Try to think about any event you have experienced, and you always associate it with a place. Space is really part of the memories that you store, and remembering a location reminds you of the events that took place there.

TKF: Your lab focuses on the entorhinal cortex, a major hub in a widespread network for memory and navigation. What do your findings suggest about how exactly location in space is encoded and contributes to memory?

EDVARD MOSER: We’ve learned that at least three different functional cell types, including the grid cells that we discovered, are mingled together in the entorhinal cortex, and that their pattern of firing maps an animal’s position in space. The output from the entorhinal cells is a kind of matrix or coordinate system on top of which you encode all kinds of other things. The synthesis of space and these other things is then stored in the hippocampus.

TKF: So cells in the entorhinal cortex locate where you are and appear to provide an underlying framework for memories. But how do these cells work with the nearby hippocampus, a brain region crucial for memory making but not for storage of long-term memories?

EDVARD MOSER: The entorhinal cortex generates a map with distances and directions based on an animal’s movement around in an environment, which can be anywhere. But in order to find your way in the environment, you also have to know about your surroundings--how your office is different from your bathroom, for example. You need to know about specific things that are in the room. To know how environments differ, the generic map of the entorhinal cortex is not sufficient. You also need what are called “place” cells in the hippocampus. Both the entorhinal cortex and the hippocampus work together to make memories of events.

MAYANK MEHTA: Also keep in mind that the entorhinal cortex has two parts—one that has the grid and other cells that code the spatial or “where” information, and another part that seems to code for the objects or “what” information in an environment. Both streams of information converge in the hippocampus.

TKF: Dr. Dickerson, you use innovative brain imaging techniques in humans to decipher what parts of the brain play a critical role in memory and memory disorders. Have you also found this to be true for humans?

BRADFORD DICKERSON: There are a hierarchical series of anatomical regions thought to be doing similar things in humans, but a lot of the details are less clear. When you walk into a room for a meeting, you first perceive the items and people that are in the room. That information travels from several parts of the brain but eventually goes into the portion of the entorhinal cortex responsible for processing objects and then into the hippocampus. At the same time, information about where you are as you walk into the room and where everybody else and the items are in the room comes through a different pathway in the brain, but also ultimately probably feeds into the entorhinal cortex and hippocampus. Both major streams of information are somehow bound together in a way that allows us to mentally time travel back to that meeting and recollect a lot of details the next time we see a colleague that was present at the meeting, for example. This model of memory is largely informed by rodent work and to some degree by research in the monkey.

TKF: What about your human findings?

BRADFORD DICKERSON: Recently, we’ve been using functional MRI to detect the strength of brain activity of different key nodes in the neurocircuitry thought to underlie memory of events. We found that the stronger the correlation of activity between these key areas of the brain, which are spread apart in space, the better a person does on a memory task. The efficiency of that large-scale neuronal network may help explain differences in memory ability. The entorhinal cortex is part of the network, and there are many thinkers in the field that view this portion of the brain as critical for memory in humans. But many other parts of the brain are also part of that memory circuit.

EDVARD MOSER: I agree that what we recognize as a memory is something that takes place all over the brain and involves a huge network. But the hippocampus is special in many ways because it is where the “what” information converges with the “where” information. It has these linking abilities that are very useful for memory. Although, in the last few years we’re recognizing that the hippocampus doesn’t do anything of this in isolation.

BRADFORD DICKERSON: Don’t get me wrong; I do love the hippocampus. But when it comes to human memory disorders, people have really neglected the entorhinal cortex and an adjacent region called the perirhinal cortex, which plays a key role in the memory of objects. Some research is now being done on patients that have lesions restricted to these areas to find out what they do in human memory of events, or “episodic” memory. The patients with the most profound amnesia or loss of episodic memory, such as those with Alzheimer’s disease, have damage to a central region of the brain that includes the hippocampus, the entorhinal cortex, and the perirhinal cortex. So we know those are critical regions in the human brain necessary for episodic memory. But there are other memory disorders that are due to lesions in other parts of the brain. They don’t tend to be as profound and can affect memory in different ways, but they still can have major impact on the person’s ability to live an independent life.

TKF: Dr. Mehta, your lab recently showed the key role the entorhinal cortex plays in encoding permanent memories in mice. Can you tell us about your recent findings on the entorhinal cortex and hippocampus, and how they relate to memory storage in the brain during sleep?

MAYANK MEHTA: By recording the electrical activity in the brain cells of rodents, we detected in the hippocampus a signature for a memory of a new route the animal took. We then wanted to find out what happens to that memory trace during sleep because it was thought that it was during sleep that memory becomes permanent and gets stored in some part of the brain outside the hippocampus. One possibility is that the memory trace formed during behavior gets stronger during sleep so it’s easier to retrieve. The other possibility is that the memory trace is selectively pruned so that useless or non-salient memories are thrown away. We found evidence for the latter.

Surprisingly, we found that during sleep – when the portion of the entorhinal cortex that codes the “where” information is active, as if remembering – it made the hippocampus become the most active. Other studies have shown that the neocortex “talks” to the hippocampus during sleep and the making of permanent memories. But we found a new player—the entorhinal cortex—is involved in this process and it’s having an enormous impact.

TKF: Once the memory is pruned down to its essentials, do you think the entorhinal cortex is stimulating the hippocampus to deposit it permanently in the neocortex?

MAYANK MEHTA: We don’t know that yet. Our tentative hypothesis is that this entorhinal-hippocampal activity going on during sleep may be playing a crucial role in erasing some unwanted memories in the hippocampus, which is needed for formation of new memories, or it may be linking up new to old memories. That way you can have a coherent perception of an entire space over time, rather than I was here and then I was at the neighboring place.

TKF: So the neural network for memory proves to be quite extensive. Are any of you surprised?

EDVARD MOSER: With hindsight I would say no—it makes sense. But if you consider what we knew 20 years back, it was easy to think that everything happens in the little part of the brain that you knew a lot about.

MAYANK MEHTA: There are some memories that don’t have anything to do with space. There are memories for procedures, for example, like riding a bicycle, that don’t require the hippocampus. Clearly the hippocampal-entorhinal circuit is crucial, but I can imagine certain types of learning that occurs in different circuits. Does the memory circuit have to be so complex? It depends on the nature of the memory. The way the brain is organized says something about how complex the circuit may be. But why the brain is organized in a way that sometimes is not intuitive is a good question.

TKF: Part of that could be the way the brain evolved in mammals. Isn’t it thought that one portion of the brain was built on top of another or used for a different purpose in mammals than it was intended for in lower animals like insects?

MAYANK MEHTA: Yes, and the hippocampus is one of the older brain structures, so it’s possible that crucial functions like memory were assigned to this circuit and then other upper brain structures—such as the “neocortex” or new brain—arrived that took on more specialized functions by building onto this rudimentary but important memory capability.

TKF: Dr. Moser, you hypothesized in a recent journal article that originally the navigational component of the brain was purely for that purpose in lower animals, but then in mammals it perhaps evolved to give us the complex memories we need for learning and other tasks.

EDVARD MOSER: Yes; it’s just an idea, but we realized that in the entorhinal-hippocampus circuit you get a lot of storage capability that you don’t need for simple navigation, like finding your way to a place a few meters away. A simple insect brain can navigate very well with very few neurons, which makes me wonder why we have these several hundred thousands of grid and other cells in the entorhinal cortex. It provides a mapping system that is very good at finding an animal’s position in space, but with a lot of redundancy. What’s been used in the development of mammals may be the capacity it has for storing many other things too. We hypothesize the mammalian brain uses the mapping properties of the entorhinal cortex and the hippocampus to store, for example, not only your path in space, but actual sequences of events for memories.

TKF: What are the practical implications of understanding the neurocircuitry behind memory formation?

BRADFORD DICKERSON: In regards to what Mayank was just saying, when we give patients with Alzheimer’s disease one of the standard medications, many seem to remember more of their dreams, some of which relate to what they did during the day. I’ve thought of that not so much as a side effect but as potentially part of the fundamental nature of what this drug is doing to their brains while they’re sleeping and trying to make permanent memories of information they aren’t able to learn very well because they have damage to this system of the brain. Understanding more about the role of sleep and related quiet wakeful states have in influencing brain activity is also important because a lot of common sleep aids disrupt the normal brain activity that occurs in these states and may be affecting memory.

But more generally, patients and families both are very distressed when their memory starts failing. People’s ability to live independent, meaningful lives is highly dependent on the integrity of their episodic memory. We hope the insights that come from neuroscience will not only help us treat patients, but help us think about our normal memory abilities and what we can do to optimize this aspect of our brain’s health in our routine lives.

MAYANK MEHTA: I agree that some of these drugs for sleep are probably influencing memory in unintended ways. I also wanted to point out that if we focus on how these memory circuits link up to each other during behavior and sleep and what affects them, we may find things that could improve memory. We have a finding, for example, that suggests exercise may improve your memory. This finding builds on the findings of Edvard’s group of a type of brain activity in the entorhinal cortex-hippocampus circuit that is called gamma rhythm. Studies indicate this gamma rhythm plays a role in learning, memory, and attention while an animal is awake and active. We found last year that we could enhance this gamma rhythm by making rodents run. So if that gamma rhythm is also good for memory in people, you may able to improve your memory by running.

Another practical side might be that as we understand how neurocircuits are laid out and function, we may be able to design computers or machines which mimic them and what they can do. We’re seeing some of those devices already and how we can’t live without them.

EDVARD MOSER: In order to get better at preventing and treating the one-third of all diseases that are brain based and to help the millions of people affected by them, you need to have a general understanding of the brain. Memory is so central to anything the brain does. So if you want to understand the brain, you need to understand memory.

MAYANK MEHTA: Another reason to do this research is just to satisfy our simple curiosity. So many things in the past were done just for curiosity sake—crazy things like flying a kite in a thunderstorm. But that experiment was the origin of the electricity we so depend on now. The applications of science always happen in unforeseeable ways.