Photo credit: Geir Mogen/ Kavli
Institute for Systems Neuroscience
The most advanced surveillance
system you will ever find is built into your own brain and nurtured by
evolution. It comes equipped with a coding system that stockpiles and maps your
lifetime of events in high definition. Through new research tools and insights,
scientists are gradually coming to understand this coding system and its
intrinsic mathematical principles.
Researchers have long known that
different kinds of neurons play different roles in the brain, but only in the
past few decades have scientists had access to the imaging and measurement
tools they need to see how different neurons react when the brain is challenged
with different tasks. We have focused on understanding how the brain helps us
navigate in the environment, both because it is intrinsically interesting and
because it turns out that navigation—finding our way—is linked to the way we
We now know that this coding
system works like your own air traffic controller—monitoring every movement
you make, knowing every step you ever made, and creating links to every event
and experience you have had. Essentially, while your brain is making mental
maps to help you navigate, it is also overlaying memories—experiences, smells—onto
From Map to Memories
This ability of the brain to
overlay recollections creates a cognitive map—a multilayered collection of
memories—rather than a mere cartographic map. It also means that learning how
the brain computes navigation is a step toward understanding how networks are
built up in the cerebral cortex, the part of the brain that is responsible for
imagination, reasoning, and planning—thought processes that make us human.
Further insight into how the
brain builds networks in the cerebral cortex can potentially lead to
interventions that spare millions of people from the debilitating effects of
brain disorders and diseases. The economic consequences are already significant.
One study attributes an estimated 35 percent of all burden of disease in Europe
to brain disorders.1
study put the
cost of treatment for brain disorders in 2010 at roughly $1.09 trillion.2
Partly due to economic forecasts, the Obama
administration has recognized the importance of funding basic neuroscience
research by establishing the National Institutes of Health’s Brain Research
through Advancing Innovative Neurotechnologies (BRAIN) Initiative.
All of these factors underscore why developing insights
into the detailed workings of the brain is pivotal for both preventing and
treating disorders of memory, and why
focusing on the workings of the mammalian spatial-navigation system is so crucial. While what we do is basic research, our work nevertheless examines the very same system that collapses in the case of dementias such as Alzheimer’s disease. Every new piece of knowledge that researchers gather contributes to
understanding the puzzle posed by the brain. And we are only just beginning to
see the bigger picture.
Encoding Experience on a Map
For a long time psychologists
have studied how animals move in and relate to space as a way to understand the
larger rules governing how and why we do what we do. Initially, most scholars
thought that behavior was simply a matter of stimuli triggering responses. But
in 1948, cognitive physiologist Edward C. Tolman suggested a new way to view behavior.
The brains of humans and other animals, Tolman said, have a kind of map of their
spatial environment, and they encode experience on top of that map.3
This idea led to the introduction of the cognitive map.
Tolman’s idea was debated but not
fully accepted until 1971, when John O’Keefe and John Dostrovsky discovered
place cells.4 Place cells, which fire when an animal is in a
specific place, are located in the hippocampus, a paired structure deep inside
the brain underneath the cerebral cortex. In experiments, these cells “fired”
whenever a rat was in a certain place in its local environment—an indication that
there was suddenly something in the brain that actually looked like a map. This
finding also helped demonstrate that humans and other animals could make mental
maps rather than simply relying on landmarks.
In 1978, O’Keefe and Lynn Nadel
took this one step further, proposing that place cells provide animals with a
dynamic, continuously updated representation of space and the animal’s position
in space.5 Tolman, it seemed, was right after all.
Discovering Grid Cells
The discoveries of the 1970s gave
scientists more clues about what to look for and where to find it. So they
looked—and found. One key discovery was head direction cells, which fire when
animals face in a certain direction, regardless of the animal’s position.6,7
In 2005, our laboratory
discovered another kind of cell, called the grid cell, which is located in the
entorhinal cortex, right next to the hippocampus.8 As its name
suggests, grid cells create a regular, triangular grid by firing when an animal
passes over equally spaced locations. The grid looks very much like the pattern
on a Chinese checkers board.
Three years later, our lab and
another lab simultaneously reported the existence of yet another type of cell,
called a border cell, which fires when an animal is near its environment’s
border, such as a wall or an edge.9,10
The collective significance of
these findings is that the reactions of the neurons can be matched to what is
found in the external world. It is still too difficult to trace other types of
complex thinking to their sensory origins. Where information is combined across
sensory systems, the firing patterns of the neurons involved are too diffuse for
us to detect patterns and relationships to what is happening in the external world.
We know the hippocampus is
critical in forming memories, in part because of the unfortunate experiences of
an American patient known as HM, who had surgery in 1953 to remove most of his
hippocampus as a cure for his extreme epilepsy. The surgery succeeded on one
level by reducing his epileptic seizures, but it left him unable to make new
HM is not the only patient whose
experiences have illuminated the workings of memory. Many others who suffered
injuries to the hippocampus have helped underscore the important connection
between this part of the brain and memory formation. We know that one of the
first symptoms of Alzheimer’s disease is that patients get lost—and that the
first place where a patient’s brain cells begin to die is in the entorhinal
Essentially, if you have lesions
in these areas of the brain, you lose your ability to find your way—and your ability
to recall all other types of memories. Memory is deeply and physically
connected to our perception and encoding of space. Thus, a detailed understanding
of this region of the brain and the operations of its neurons may have the
additional benefit of illuminating the mechanisms behind Alzheimer’s disease
and related dementias.
As young researchers, what we
most wanted to understand was behavior, as well as the origins of complex
psychological functions. It’s a question that will take many lifetimes to
answer. So by focusing on something more accessible, such as the way space is represented
in the brain, we can begin to understand how the brain computes itself, and how
external inputs from the senses get into the primary sensory cortex.
Finding these cells required us
to use microelectrodes, tiny wires that are thinner than a human hair. They
must be correctly placed close to individual neurons in the brains of rats to
allow the firing of the neurons to be recorded.
A rat’s brain is the size of a grape. Inside there are about 200 million
neurons, each of which has direct contact with approximately 10,000 other
Inside each side of the rat’s grape-size brain are areas
that are smaller than a grape seed—collectively, the hippocampus—where memory
and the sense of location reside. This is also where we find the place cells—neurons
that respond to specific places. But from which cells do these place cells get
The answer is to look “upstream”
of the hippocampus, to the entorhinal cortex, which feeds information to the
“Listening In” to Neurons
Microelectrodes allow us to listen in on the electrical activity of the
cells inside the entorhinal cortex. We have advanced this technique to a level at
which we can listen to several hundred cells inside a single rat’s entorhinal
cortex. Listening to many hundreds of cells has allowed us to discover that the
brain has a number of modules dedicated to self-location.11 Each
module contains its own internal, GPS-like mapping system that keeps track of
movement, as well as other characteristics that distinguish it from other
Different modules react differently to changes in the environment. For
instance, some scale the brain’s inner map to the animal’s surroundings, while others
do not. And the modules operate independently in several ways. The brain can
use this independence to create new and varied combinations—a very useful tool
for memory formation.
This finding suggests that the ability to make a mental map of the
environment arose very early in evolution. All species need to navigate, so
that some types of memory may have arisen from brain systems that were initially
developed for the brain’s sense of location.
The grid cells in each of the brain’s modules send signals to the place cells
in the hippocampus. The combined effect of this grid cell activity creates an
activity field in the hippocampus, the place field. This signaling, in a way, is
the next step in the progression of signals in the brain. When the environment
changes, the different grid modules react differently to the change—firing at
new positions in the environment, and the linear summation activates different
place cells in the hippocampus.
In practice, this means that the grid cells send a different
combinatorial code into the hippocampus in response to the slightest change in
the environment. So every tiny change results in a new combination of active
cells—cell ensembles that can be used to encode a new memory, and that, with
input from the environment, become what we call memories.
Recent advances in technology
have given us opportunities that we could barely dream of only a few years ago.
One is the ability to create detailed functional maps that show which neurons
talk to each other. We are particularly interested in how grid cells and place cells communicate. The
answer to this question will allow us to understand how the deepest parts of
the brain are wired together.
When neurons send signals to each
other, they share many similarities with electric cables. They send an electric
current in one direction—from the “body” of the neuron and down a long arm,
called the axon, which extends to the branched arms, or dendrites, of the nerve
cell next in line. Brain cells thus get their small electric signals from a
whole series of such connections.
A recent technique in our lab involves
using a highly modified adeno-associated virus
(AAV) as a biological transport system
within neurons to better understand which neurons talk to place cells in the
hippocampus. The virus is modified so that it can enter specific neurons and
travel upstream through the axon and into the dendrites. We attach a
light-sensitive gene to this viral transportation system. This gene integrates
itself into the neuron’s DNA and makes the neuron light sensitive. Normally, of
course, the neuron is tucked away in the deepest recesses of the brain, in the
dark. So this process allows us to install the equivalent of a light switch in
a neuronal network.
We used this technique to insert
light switches into place cells. Then we inserted optical fibers into a rat’s
brain, which enabled us to transmit light to the place cells that had light
switches in them. We also implanted thin microelectrodes between the cells to
detect the signals sent through the axons every time the light from the optical
fiber was turned on. This allowed us to see exactly how the cell-to-cell communication
was wired and to map small and large networks within the navigational
computation system of the brain.
When we put together all the information, we saw that there is a whole
range of differently specialized cells that together provide place cells with their
The brain’s GPS—its sense of place—is created by signals from place
cells to head direction cells, border cells, grid cells, and cells that have no
known function in creating location points. Place cells not only receive
information about a rat’s surroundings and landmarks, but also continuously
update their own movement—an activity that is actually independent of sensory
We were surprised to find that cells that have no role in our sense of
location actually send signals to place cells, because until now, the specific
kinds of brain cells found to be involved in navigation—place cells, head
direction cells, and grid cells—all have specific jobs. What is the role of the
cells that are not actually part of the sense of direction? They send signals
to place cells, but what do they actually do? This remains a mystery.
We also wonder how the cells in the hippocampus are able to sort out the
various signals they receive. Do they “listen” to all of the cells equally
effectively all the time, or are there some cells that get more time than
others to “talk” to place cells?
Speed Cells and Decision-Making
It is easy to forget, as we move effortlessly from home to job, or from job to
supermarket to home, the enormous number of processes and steps that make up
our ability to navigate. We are now working our way through different aspects of
the brain’s navigational system to better understand how all these pieces fit
At the moment we are studying what we have dubbed speed cells—cells that
react exclusively to the speed of an animal’s movement—and how these types of
cells factor in to the navigational equation.
We’re also looking at decision-making. As an animal moves through a
labyrinth, it must choose which way to go or what turn to make next. The neurons
involved in this decision-making can be found in the prefrontal cortex, which
connects to the hippocampus via a small nucleus in the thalamus.
Slowly but surely, we and other researchers are
expanding our understanding of other parts of the brain to figure out how
everything is connected. And because everything is connected, we are hopeful that as we and others make ever more
detailed maps of neural networks, we become more and more likely to find clues
that will help prevent and cure brain diseases in the future.