Building a Circuit-Diagram for the Brain (Jennifer Raymond, Stanford University)

Jennifer Raymond (Stanford University) is building a "wiring diagram" for the brain. By bridging the gap between individual synapses and whole-brain learning & memory, Raymond's research offers new insights and strategies for medical rehabilitation and K-12 education.

Prof. Jennifer Raymond's website:
http://raymondlab.stanford.edu/

Stanford University:
http://www.stanford.edu/

Stanford University Channel on YouTube:
http://www.youtube.com/stanford

Isn't it amazing how the face of an old friend can seem so familiar even if you haven't seen

them in years or even decades? On the other hand, the names of some of your old classmates

may have been forgotten. Have you ever wondered whats going on in here to support these successes

or failures of learning and memory? Well this is the subject of the research in my laboratory.

We're trying to understand exactly what changes in your brain when you learn and how those

changes persist over time to support memory. And one thing that we know which helps explain

why some things are easier to remember than other is that learning is not a unitary process.

There is no single mechanism of learning in the brain but Instead there are distinct kinds

of learning that depend on distinct brain regions. A brain structure called the hippocampus

supports memory for facts and events in your life. This is what you rely on to remember

someones name or what you had for breakfast. Where as another structure called the amygdala

supports emotional memory. You can have a fear of dogs even if you've lost the explicit

hippocampus dependent memory of being bitten by one as a child. So these memory systems

are fairly independent. The basal ganglia supports habit memory. This is what you're

using when you brush your teeth or drive to work when your mind is elsewhere. The cerebral

cortex supports perceptual learning. Even basic functions like being able to see depend

on experience and learning. And this structure down there is called the cerebellum. It supports

motor learning. This is the process by which you acquire skilled movements. If we were

to zoom in on any one of these brain areas we'd find that they are made up of the same

basic building blocks. Neurons which are specialized cells of the nervous system and synapses which

are the connections between neurons that allow for one to signal to the next. But unlike

wires synapses are not static and can change with experience. The electrical and chemical

signals that flow through your synapses as they process information can induce long lasting

changes. So on the one hand we know a lot about how learning and memory are organized

at the functional whole brain level. On the other hand we know a lot about its implementation

at the cellular level with neurons and synapses. The next great challenge and the one that

my lab is tackling is to try to bridge the gap between these very different levels of

organization and understand how learning works at the level of the neural circuit. Because

it is the circuit level organization that causes changes in synapses in the hippocampus

to be able to encode the name of someone whereas changes in the synapses of the cerebellum

improve your tennis game. A lot of the magic occurs at this intermediate circuit level

of organization. And of course thats true not just for brains but for many things its

that intermediate level of organization that is really critical for how that thing works.

So if for example you wanted to understand how a car works so you can fix it you might

go down to your auto parts store and carefully examine spark plugs and belts and gaskets

and hoses and things like that. And you might also draw on your experience as a driver to

know that there is a power system that makes the car go and a steering system that makes

it turn and a braking system. But thats not enough right? If you want to fix your car

the critical thing is to understand how all those parts interact to give rise to the engine

that makes the car go and how all the parts fit together to make the steering system that

makes it turn. Its this critical intermediate level of organization that is necessary if

you want to fix your car. Of course for the car, we have things like engineering drawings

and auto mechanics repair manuals that give us that information about how the parts work

together but we have no such thing for the brain. And so thats what my lab is working

to produce because thats what we need if we want to be able to fix it. And of course we

do want to fix it. One in twenty kids has a learning disability. One in seven people

over the age of 70 and half of the people over 85 have alzheimer's disease or a related

dementia. And the treatments available at this point are not as effective as we'd like.

They are mainly pharmaceutical. And for most drugs we have some idea how they act at the

level of individual neurons or synapses. But we don't know much about how the effects at

that level then effect the next level up, the neural circuit and its ability to process

and store information. So sometimes the drugs work and sometimes they don't and often we

don't really understand why. Some really new and exciting technologies are being developed

that will allow us to manipulate the brain with a precision thats not possible with drugs.

Even if tomorrow someone were to hand doctors a magic new technology that would enable to

control neurons and synapses with whatever precision they want safely inexpensively.

Those doctors would still not be able to improve school performance of the kids with learning

disabilities and they would still not be able to prevent cognitive impairment and loss of

memory in older adults because at this point we really don't understand enough about learning

to know which neurons and synapses within a circuit would need to be tinkered with.

So at this point giving doctors this magic tool would be like giving me some fancy wrenches

but no repair manual and asking me to fix your car. I could go in there and make some

changes and I might get lucky but if it was my car or my brain I would like to have the

detailed repair manual available. So thats what my lab is working to produce. We're trying

to understand how neurons and synapses work together in circuits to support learning and

memory. So what do we know about neural circuits? The function of a circuit is to compute. To

take an input and generate an output. And this transformation of input into output is

accomplished and shaped by the very precise patterns of interconnections synaptic connections

between neurons and your neural circuits. And this is really how information is processed

in the brain. Information is processed and transformed and used to make decisions through

the many individual synaptic signaling events that occur in a neural circuit. So for example

your driving down the road and see a yellow light. That input will activate neurons in

the visual parts of your brain and when they're activated they'll send signals to the neurons

on which they make synapses. A typical neuron has connections with thousands of other neurons.

If some of those neurons get enough input they will then become activated and they will

signal to the next neurons which will signal to the next neurons until eventually an output

is generated, a movement of your foot to the

accelerator.

A lot of you are nodding but

a few of you are frowning disapprovingly. But to those of you frowning never fear because

those synapses that define our neural circuits as I said are not static but can change with

experience. So for example

If you get a ticket for running a red light

this is likely to induce changes in your brain. Some synapses might get stronger, others could

get weaker

and this will cause this circuit to process information differently the next time it is

activated. So the next time you see that yellow light the output of your circuits could be

quite different and you might move your foot to the brake. This is just one simple example

of the kind of computation that your brain is performing every day. And we think that

virtually all of the computations that the brain performs are powerfully influenced by

experience and learning. My lab focuses on the effects of learning on the computations

performed by this brain region, the cerebellum. And the cerebellum has some cognitive functions

and also as I mentioned earlier it plays a key role in motor learning, the process by

which motions become smooth and accurate with practice. And you might think first of musicians,

athletes, and dancers, but if you've ever observed a small child then you probably realize

that most movements are learned. Even mundane acts like walking or reaching accurately for

something without knocking it over are gradually learned through a lot of repetition and practice.

And even once we acquire those skillful movements the circuits that produce those movements

need to be recalibrated as the body changes, as it grows and then ages we need to make

those recalibrations or our movements will again become clumsy like those of a child.

And this is in fact what is seen with damages to this brain area. So the cerebellum has

some important functions but the main reason my lab focuses on this structure its the brain

region where we have the very best chance of understanding how learning works at the

level of the circuit. Why is that? Well, one of the very first things that you need if

you want to analyze the circuit is the wiring diagram. We need to know which neurons are

connected to which and how signals flow through the pathways, through the circuit. And for

most learning tasks we do not have that but we do have it for several learning tasks that

depend on the cerebellum and the heart of that is shown here. So now with this wiring

diagram in hand we are able to go on and ask the next level of questions about how the

circuit computes and how learning affects that computation. And my lab is asking three

very fundamental questions about this process. The first is where in the circuit do changes

occur when you learn? If you go out and practice your golf swing this weekend, which synapses

in your cerebellum will get strengthened and which will get weakened. Will it be the connections

from the green neuron to the red or from the green to the purple? Are particular types

of synapses more likely to go under changes than others? And are all the changes happening

at one stage in the signal processing pathway or are there multiple serial changes? These

are the kinds of issues that we are looking at. A second very fundamental question is

how are changes induced in the circuit. Which neurons in your cerebellum are monitoring

the accuracy of your swing and deciding when the circuit that produces that movement needs

to be updated? Which neurons know when you made a mistake? And the third fundamental

question is how are synaptic changes in the circuit read out? How do particular changes

in the circuit alter the way that it processes information the next time that it is activated?

And we don't have all the answers to these questions but what we've found so far by studying

the cerebellum parallels in many ways what has been seen at the whole brain level. We

know that there are distinct kinds of learning and memory that depend on distinct brain regions.

But until very recently it was thought that within the brain region there was one main

learning mechanism. So that every time that that brain region learned it did it in very

much the same way. In contrast what we found out is that the cerebellum contains within

it multiple learning mechanisms. So if your golf swing needs a particular type of adjustment

you may be able to accomplish that through different combinations of changes in the circuits

of your cerebellum. We find that individual training sessions can engage more than one

learning mechanism and fairly subtle changes in the way that we do the training or the

way we practice can determine which mechanism are recruited or not recruited, which synapses

change or do not change. And this has some really important implications because we think

that the recruitment of different learning mechanisms will affect factors like how long

the learning is retained, whether what you learn in one context will generalize to other

contexts, and we think that it will affect the ability for learning to be reversed if

circumstances change. And this is important not just for your golf swing but for other

things like developing rehabilitation strategies for patients that have had a stroke and developing

education strategies for our kids. But of course along the way if we do find something

that will improve your golf swing that will be ok too. Thank you for your time.