Brain Machine Interface BMI

Loading...

En provenance du cours de Hebrew University of Jerusalem

Synapses, Neurons and Brains

700 notes

Explore Related Videos

At Coursera, you will find the best lectures in the world. Here are some of our personalized recommendations for you

Hebrew University of Jerusalem

Synapses, Neurons and Brains

700 notes

These are very unique times for brain research. The aperitif for the course will thus highlight the present “brain-excitements” worldwide. You will then become intimately acquainted with the operational principles of neuronal “life-ware” (synapses, neurons and the networks that they form) and consequently, on how neurons behave as computational microchips and how they plastically and constantly change - a process that underlies learning and memory. Recent heroic attempts to realistically simulate large cortical networks in the computer will be highlighted (e.g., “the Blue Brain Project”) and processes related to perception, cognition and emotions in the brain will be discussed. For dessert we will deliberate on the future of brain research, including the questions of “brain and art”, consciousness and free will. For more information see the course promo below and read “About the course.”

À partir de la leçon

Brain Excitements for the 21st Century

Welcome to synapses, neurons and brains!

Welcome Words and Great Thinkers9:31

The Blossoming of the Brain in the World7:53

The Connectomics10:31

Brain Machine Interface BMI16:48

Optogenetics7:24

Simulation of the Brain - Blue Brain Project8:57

Rencontrer les enseignants

Idan Segev
Professor
Computational Neuroscience

So the third frontier I want to discuss today is really exciting and very, very

important, also clinically, is what we call brain machine or brain computer

interface. The interface between the brain, which is

a machine but a very special machine and the machines we build for doing that, for

being able to interact with the brain, certainly the anatomy is not enough.

Okay, anatomy is a static thing. The anatomy may change in terms of

structure, and we'll show you later on when I talk about plasticity in the

brain. I'll show you that when you learn, there

are changes, anatomical changes in your brain.

But in order to interact with the brain online, I need to read the signals within

the brain. Not only the anatomy, so I need to read

the electrical signals in the brain. So, on top of the anatomy, on the

foundation which is anatomical, there are all these signals.

We'll talk about, about the electrical signals in the brain, what type of

signals, all the signals are running, so to speak, on the circuit, and these

signals stand for something that the circuit wants to convey.

So a particular circuit, when there are particular signals, it means that I need

to move my hand to the right and so on. So in order to really read out what is

happening in the circuit in terms of information, I need to read the

electrical signals. And as I said, we'll focus a lot about

the electrical signals in lesson three and four.

So how do they look, these electrical signals that are being used to process

information to send commands to my hand, or to my mouth, or to my thought, or to

my feeling. So the idea is the following, you go to a

particular brain region, let's say here. You know, that this particular brain

region from previous studies is responsible, let's say for movement.

If you would zoom into this brain region, of course, you would see networks,

networks and anatomical networks of neurons.

If you zoom out, you see a big network, many, many, many, many, many cells,

billions of cells within this network. And you may record using electrode, very,

very fine electrodes that implant or in pain into this region into this

particular one cell. And you may listen, so to speak, to the

electrical activity of this one cell. So here is how it looks.

[NOISE]. So this is the electrical activity of one

cell in the brain of a mammalian, in this case in the cortex, when the animal is

doing something. So you see that the brain fires, we call

it fires, some electrical activity. We can put into a, a loud speaker and you

can hear. Of course, there is no sound within the

brain. This signals do not make sounds just

change electrical activity. But if I connect it to a loud speaker,

you would hear what you heard[NOISE]. This is so to speak, the bar code of the

brain, if you want to call it this way. We call them spikes.

There will be a whole lesson about spikes, but you just saw the spikes that

a single cell generates. Each cell generates its own spiking

activity. So, if you would recall, for many many of

them, you will hear a complete music, a complete noisy, noisiness of the network.

But each cell has its own particular kind of spiking activity.

And overall, overall the message to move my hand up or down is running in these

spikes. So all these spikes stand for a

particular movement. If this is in the mortal cortex of the

particular feeling. If this your outer regions or particular

picture. If it is in my visual cortex, so, but

the, the common language that we should talk a lot about the signal that insert

the common language are the spikes. And so the brains using spikes to

represent the word. So there is no music in my brain, there

are no movies in my brain, there is no movement in my brain.

My brain generates this electrical activity, and when it generates this

electrical activity, I believe that there is a movie there.

And when it generates this electrical activity, I'm moving my hand, and so

forth. So spikes are very important.

And I need to be able to read them somehow in order to interact with the

machine that I want to, let's say activate a robotic arm directly from the

brain. So this is a kind of a sketch summary of

what I said. So there is a particular region in the

brain here. Let's say the motor region, the region

that activates movement, so that's it, this is the region.

Typically, when this region is active, now it's active.

It goes through my spinal cord, so the fibers that goes from this region go

through the spinal cord and activate and movement.

That's the normal case. But I can also use the same signals, if I

can read them correctly, analyze what they mean to say, the signals, the

network, and eventually, activate a robotic arm.

This is extremely useful when there is, let's say spinal cord injury.

Because, when there is a spinal cord injury, the signal are being generated in

the brain but nobody listens, so to speak, because there is this cut, or when

I'm amputee. So the, the signals are going into my

limbs, but I don't have limbs, so it cannot move.

So we can, today, use the signals online in real time using many, many, many

recording electrodes from a particular network, analyse them in real time, and

activate the robotic arm. So I want to show you an example of doing

exactly that, recording from the brain of a monkey in this case, the hands of the

monkeys are tied. He wants to move his hand, because there

is something attractive there, and so, we can do that.

So, here is the monkey. Here is a robotic arm.

Separate it from the monkey, so the robotic arm is on the side.

And there will be food coming here. The monkey can see the food and he wants

to eat. His hands, as they said are tied, so he

generates the sequence which we can read, Andrew Schwartz from University of

Pittsburgh can read online hundred and hundred of sense together, and then

activate the robotic arm. Let's see how it goes.

So here is the food, the robotic arm is activated directly from the brain of the

monkey. The monkey thinks about the movement,

generates this electrical activity. And he moves the robotic arms directly

from the brain to pick up this little piece of food.

That's very impressive. This is brain machine interface.

There is a brain, there is a machine, and the brain activates the machine.

And this goes in a most beautiful, smooth way.

You can see that he uses the fingers, so he activates the fingers from, the

robotic fingers, he activates their limbs and so on, eventually, he gets the food

very smoothly. That's, that's amazing.

That's brain machine interface self-feeding with a robot and useful for

cases of amputees or spinal cord injury. And actually, there are already people

going with this technology with electrode implanted in the brain and eating, and,

and moving a robotic arm. The opposite direction, the other

direction of brain machine interface is from the machine to the brain.

Before I show you how the brain activate a machine, now, I want to show you how a

machine, in this case a battery, which is a machine activates the particular brain

region. And this came, basically, or the idea

came actually already from a machine activating our heart.

So there is a pacemaker, some people who have pace, pacemaking problem with the,

with the heart. The heart is sputtering, inject

electrical activity, but I'm going to show you the same technique and walk

through the brain. And, and I show you now how, how was it

developed. So in this case, there is a recording

deep, not in the cortex, but deep inside the deep region of the brain, the basal

ganglia, you can record electrically from there.

This is Hagai Bergman from the Hebrew University.

And when you record from this area with several electrodes, this is one

electrode, another electrode, a third electrode, you can see the normal

electrical activity in this region of several cells.

So, one cell is doing[NOISE]. Another cell is firing more

vigorously,[NOISE] very strongly, and the third cell is doing it less.

Each one with its own electrical music. So this is the normal brain in this

region. In Parkinson, apparently, when you are

sick with Parkinson, this same region, the cells in the region are generating

different electrical activity. You can say a wrong electrical code.

Because, suddenly, the cells are starting to fire these spikes, but in a very

different pattern. You can see[NOISE], stop,[NOISE], stop,

like a machine gun,[NOISE], stop. Not only that, but the cells seems to

fire similarly alike here, though they were different here, they're sent to

synchronize together. And so, the whole network is doing like

this fire, stop, fire, stop,[NOISE], stop, all of them.

And this wrong code, this oscillatory activity in the brain, the manifestation

of feed, the output of feed, bound of the manifestation is this tremor in Parkinson

patients. So the idea was, and it's very, very

effective today in many, many hospitals, is to take this pacemaker, this battery,

that is implanted here below the skin to generate signals in the battery.

[NOISE] with a particular frequency and strength, which you can manipulate.

And inject this current into this region that I just show you, where you have a

wrong code, wrong electrical activity if you are Parkinsonian patient.

And so, you want to intervene, you want to somehow affect, and we don't

completely understand to the end, why does it work so well.

I'll show you in a second. Why does this signal that, you from the

outside, this machine from the outside, inject into this region, suddenly, so to

speak, symptomatically, almost completely repair the Parkinsonian symptoms in some

patients. So, here is a patient, Parkinsonian

patient, actually very young person, already about after 10 years of

Parkinson's. You can see that he has all these very,

very, very difficult and very, very sad symptoms that it's very difficult for him

to generate movement. He cannot coordinate movements like this.

He's a very, very sick patient. Of course, he cannot drive like this.

It's very hard for him to go to the grocery.

So this is a very sick type patient. And so, the question is, after 10 years

of medication, what to do? And as I said, one new technology that we

call, the deep brain stimulation, is to stimulate the brain deeply into this

region and try to repair symptomatically these symptoms.

You can see how difficult it is for the patient to do something like that, that

we do so simply, because Parkinsonian people cannot do that.

So let's see what happens to him after the operation.

So lets stop this for a second and see this same person, now receiving this

pacemaking from this battery. So it's activated now, and you can see

how well this person, look, it's the same person some months after the operation.

You can see that he's doing things that are very, untypical to Parkinsonian

patients. So, this is the intervention, this is the

interaction between the machine and the brain.

The, the machine fixed, so to speak, at least symptomatically.

The Parkinson and the, and the person really can walk, and really can drive

and, and I mean the quality of life is now really improved dramatically.

Of course, when you turn off the machine, the Parkinson's symptoms will come back

immediately. So, it's a symptomatic repair.

So, this is beautiful. And this is really a direction that we

go, because we want to repair brains and we want to end pain to the brain.

And I just want to show you, just to complete this part, the future challenges

of brain machine interface. So one challenge is to interact with the

brain In a telemetric way. First of all, you want to put some kind

of a chronic recorder or chronic probe that will listen, that will record the

electrical activity of many, many cells. So you don't want to put wires in

electrodes, you want to use nanotechnology.

In order to put chronically a little piece of machine into the brain that will

telemetrically send the signal outside of the skull.

So this telemetric communication with a brain both from the brain outside, but

also maybe to stimulate the brain telemetrically, if we can do it to a

particular region. This is the future it does not exist yet.

Some of it exists, but not all. The next challenge is to really analyze,

you saw before, you need to do it in real time, the analysis of the signal has to

do in the real time. So the monkey or the person thinks about

movement, you need to now move the hand, you do not need to wait an hour, so it's

a real time signal analysis, real time. And if you have million of signals, is,

Obama's project is to have million or billion of signals simultaneously.

This means that you'll be able to analyze all the signals in the real time and

which is a, it's a challenge for signal processing community.

So this is another challenge, to develop these methods.

Then of course, the challenge is to develop a robotic arm that is very

sophisticated, so our arm is very sophisticated with many degrees of

freedom. I can do that, and that.

We don't yet have robots as sophisticated in terms of arm movement, but there are

developments towards this direction. And finally, and it's very, very

important when I'm moving my hand and my, touching something, so I'm touching this

bottle now. I get also feedback from the fact that

they touch. So I, not only I move, I not only move,

but I also get information about that it's a plastic, that it's soft, that it's

cold. All this information comes from sensors

in my hand, in my fingers, sensors that are going to another region in the brain,

the sensing cyst region, somatosensing. So we need to put in this robot, sensors.

When the robot will, so to speak, touch a bottle or a glass or somebody else, there

should be an information closing the loop.

So movement and closing the loop in terms of touching.

That's another challenge. So this is the future challenges from

brain machine interface.

Explorer notre catalogue

Rejoignez-nous gratuitement et obtenez des recommendations, des mises à jour et des offres personnalisées.

Coursera propose un accès universel à la meilleure formation au monde, en partenariat avec des universités et des organisations du plus haut niveau, pour proposer des cours en ligne.

© 2019 Coursera Inc. Tous droits réservés.

Télécharger dans l'App Store

Disponible sur Google Play