I'm Pamela Bhatti and I'm a professor
in the school of Electrical and

Computer Engineering.

And my roots are actually
in violence in the area and

I had an undergraduate degree
of bio engineering and

I worked in the pharmaceutical
industry for a while in the ninety's.

So late mid eighty's early ninety's so
I guess I'm dating myself a bit and.

After being in the biomedical engineering
domain I really felt that I wanted to

build skill intellectual engineering so
I worked in industry for some time

before getting my Ph D. at Motorola
as well as a software startup and.

I don't know about you but for

me there's nothing like a software startup
to motivate you to go get a Ph D. and

things that are really active in gauging
where you get to call the shots.

So my Ph D. is in micro lecture and
mechanical systems and

basically it's a field where you
leverage the same technologies

the same methods that you use
to make integrated circuits.

You can develop novel neural
interface technologies.

So that's the memes piece in my talk and

I'm very in gauged in
how we can take the men.

And marry that to the neural interface as
well as taking it one step further and

trying to translate memes based
technologies into the clinical setting.

So the two main applications that I'll
talk about today are in the auditory and

vestibular systems and I think I'm going
to turn off the lights because I think

the slides will show up
a little bit better.

Thanks for my help are here.

And feel free to stop me as I go because
this is an interesting crowd because I'll

probably give you maybe
too much biology and

maybe less of the engineering of
the electrical engineering piece.

So.

I'm also involved over at
Emory with the clinical and

translational Sciences Institute and
it's really a nice way for

those who are interested at Georgia Tech
in looking at translating their

work in the clinical setting and building
skills so there is a program where you can

obtain a certificate certified
certificate in translational science.

So I like to give a broad overview
of what our lab does so in

the center here our target is biomedical
Microsystems and Neuromodulation devices.

But that involves a whole host of
more I would say engineering type

technologies with the signal
processing circuit design and

the implantable Di any implantable device
we have the wireless communication aspect

power battery all very very
critical components and

then we move over to the neural
engineering side where we directly

interface with neural tissue and
that involves validating it in human

models as well as testing it out as well
as we can and given my research focus.

I do a lot of work with folks and
a little are involved speech and

hearing as well as.

Rehab medicine.

So I'll start with the Coakley or
prostheses work and

that I like to begin with because
I think it's much easier to

understand the interface and you probably
all know that this is a technology

that provides a sensation of sound to
individuals who have severe to profound

sensory neural hearing loss by directly
stimulated stimulating auditory nerves.

And there.

This is a very prevalent device.

Over two hundred thousand individuals
are you have an implant world.

Wide and the age range is
very very broad in the U.S.

we implant children as young as
nine months and in other parts were

primarily Europe they go though
implant as young as six months and.

Does anyone have an idea why we want might
want to implant very young children.

Exactly so.

So they can learn to speak and establish
neural pathways neural plasticity So

we have a formative period
in our first nine months.

And if we don't engage
the auditory system.

We will lose or

severely impact kind of
the the learning curve for language.

And I work with pediatric totaller and
college of silver at Emory and

he's very very pro
implanting young children.

On the altar side there's folks in
the community who feel that we're forever

changing them and they don't believe that
young children should be implanted so

it's a somewhat controversial issue.

So.

People with implants
can do very very well.

Scores of nearly a hundred percent speech
recognition in quiet environments and

the challenges in noisy settings
where you have multiple

talkers coming at you from
different positions in space.

So there's a lot of effort on the signal
processing side on how to improve implants

in terms of capturing that auditory
information becoming more and

more common our bilateral implants
where they're implanting both sides for

an individual and I have a student who
worked in my lab and during the course of

time in my lab she went from one to two
implants and it was very interesting

because she said that when she got
the second implant they turned off the.

First to see how she
could come up to speed.

So I'll talk a little bit more
about how these devices work.

So stepping through the purple out
atory system we capture sound and

we excuse me conduct it to
the middle ear where we have loops

where we have these three are these three
bones that form the a secular chain and

it's very fascinating because going
from the ear drum to the inner ear.

We've got a gain of twenty D.B.
so the ears of beautiful filter.

It's a beautiful mechanical
filter the middle ear.

Once we go from the middle ear we have
an interesting challenge of matching

impedances because the middle
ear is we have air and

the inner ear where the Coakley or
asides is filled with fluid.

So it's a fluid filled chamber and

traveling waves are set up in that
chamber to depolarize nerve fibers.

So I'm looking more closely.

So if you start here at the top.

I've just taken a wedge
out of the Coakley area.

And it's a multi chamber device.

So now let's take that wedge and
look directly at the sensors

that transduced mechanical energy to
electrical energy the hair cells here.

And so this membrane called the basil or
membrane is a very very beautifully

designed mechanical structure so as the
traveling waves move through the Coakley.

This membrane moves up and down.

And this motion then deflects these

outer hair cells which then get
connected to the inner hair cells.

So a series of electro chemical
reactions occur that essentially

translate these mechanical deflections

into neural impulses carry through
these per four processes of all.

Dettori nerve fibers.

So the beauty in this from an engineering
standpoint is if we take this membrane.

And we look down on it
kind of looking down on a.

Cinnamon bun or something we see that it's
spiralling structure makes it sensitive

in a frequency the placemats being so
we're more sensitive to high frequencies

here at the onset of the Coakley as at
the base of the Coakley as you come in.

It's a stiffer membrane.

And as you spiral towards the top
of the cold clear we are more

sensitive to low frequencies and so
we can resolve to a first order.

To first order we can resolve
frequency to place and

this is pivotal and how will Coakley or
prosthesis functions.

So.

Sensory neural hearing loss of.

What I started with in terms of
the candidates for Coakley are implants.

They have lost hair cell function and so

what I'm showing here is the top of the
hair so the stereo cilia on the hair cell

which deflect and
on the right hand side we have.

Damage hair cells on the left hand
side we have a healthy hair cells.

So there's a variety of reasons
that can lead to hair cell loss or

damage there can be a congenital
defect bacterial or

viral infections as well
as acoustic trauma.

So a lot of folks in the military have
sensory neural hearing loss because

they've been on aircraft carriers
been exposed to these and

so what we lose this ability to translate
electrical to mechanical energy.

We have her cell.

We have hearing last.

So the nice.

I guess quote unquote nice thing in the.

Story is that while we have damage
to hair cells there is the ability

to conduct electrical energy in the neural
pathways and so that's exactly what a cold

clear implant does the hair cells have
been lost but some also prefer processes.

But however the Coakley or nerve response.

And stepping through a Coakley or

prosthesis if we start
on the external portion.

Job of the signal processor so we capture
sound through a microphone one or more.

So the more advanced devices
have multiple microphones.

So they can try to capture
directional information.

And those are translated that information
is translated into a series of

voltages essentially and
the external speech processor here.

Makes it really makes what it does
is it decides what information

is important for speech or music and

that's a bit more advanced but
it's very sophisticated the first.

Well currently the external speech
processor is have four D.S.P.

cores in them.

So a lot of engineering goes in
the external speech processor and

these algorithms has evolved over time and
what I've shown here is the most

commonly used speech processing algorithm
and essentially it distills out or

it breaks frequency into sixteen
to twenty two spectral bands.

It filters the energy.

So it filters those bands so it can pick
out the energy because you have frequency.

So you've taken our wealth of sound and
picked out some frequencies.

Now for each frequency.

How do we simulate the auditory nerve.

So we pick out the energy
of that signal and that.

Energy level is used to
control the current level.

That stimulates auditory nerve fibers.

So if we you know all that information or
all those decisions are made externally

and transmitted wirelessly to
an implanted receiver stimulator here.

Shown as number five.

So from that implanted receiver stimulator
pulses are delivered to our Dettori

excuse me to the Coakley listed
as number six here where

we've now taken frequency and
were hitting specific frequencies

in the Coakley via the position of
the electrodes on the electrode array.

So looking a little more
closely at the electrode a ray.

This is the one method to

target the vestibular nerve with this
much spectral information as we can.

So I'm a little cheat a little
bit because I've done so

much work on the periphery that
ultimately it's a perception.

So you need sound from as many

sources as possible and how do you form
that image centrally of the sound so

the goal here is well if we can improve
the connection through the periphery.

We can provide as much richness
to the auditory system and

the auditory system can then provide
the person perception of sound so

this is an older electrode
array the black are the.

Electrodes so the banded design.

And it's inserted into the scale attempt
any of the bottom chamber here of

the Coakley.

Most electrodes or rays do not go
all the way up to that to the tip

to the apex we have two and
a half turns in the human Coakley.

And while I've discussed
the frequency to place math the.

That's how the membrane is designed but
in terms of resolving those

neurons that we can simulate once we get
up to the apex we lose some of that and

that's one reason why cochlear implants
the electrode ray does not go all

the way to the top the other reason is
it's actually really hard to insert

an electrode array and imagine that
you need to insert this device.

However you want to less trauma
as you can introduce into

the Coakley A because you're really you're
after maintaining this neural population

because that's your conduit
to the auditory system.

So let me play a few audio clips.

I hope this works well to give you
a sense of how a cochlear implant works.

So what we've done is we've taken
in the speech processing algorithm.

We've distilled out
the auditory signal and

then we've then remakes it as an auditory
sound signal for us to appreciate.

The stimulation aspect.

Such.

Try.

Can hear.

So these sentences.

Should Show Up to two more.

She.

So.

How many of the sentences
do you think you got.

None.

So I was asking how many of those
sentences was at the volume Well

look I can reply I am no.

So these are these
are sentences that are provided

during the speech recognition test and
I was hoping.

Often when individuals hear these those of
us who hear these after hearing a few we

can recalibrate to this poorly
spectral represented information and

listen to the sentences but
clearly it's shifted

in frequency from what it's it's more
robotic a little bit uncomfortable but

what's impressive is a cochlear implants
cochlear implant users can make

this is an analogy but
they can make an electrical input

provide this and
use an image to them an auditory image.

So I'm going to move to music on this.

Let me.

This one I can play a little bit longer.

So hopefully it will
have more of an impact.

So this is just music.

So you

heard instrumental in voice right.

OK So

let me play and
get my miles here let me play the song.

So courtesy of John Lennon.

We've got a long ways to go.

We really have a long ways to go and
that's why I continue to do

a lot of this research while really
good performers can gain a lot.

We have a huge range of levels
of perception Coakley or

imply users would like to hear much
more of the world around them and

another piece that I
didn't mention is that.

Tonal languages so Mandarin Chinese
jobby there's a whole host of languages

that need a different way of expressing or
they need.

Basically they need
a different selection method.

For providing that spectral input to them.

And so the work that I've been doing
in my lab is going for some time is

enhanced seeing that spectral content in
that's through electrode or a development.

So to first order and those who work.

There's always this big debate in Coakley
or in the Coakley or implant world.

So I get in a lot of trouble when I go
on and on and on about the periphery and

how if we improve the periphery.

You know that's all we need.

That's not true.

It's not true.

We really need.

There's a range of users but for
poor people who have a poor.

And you want to use the word poor but
they're more challenge to

with a Coakley or a plant they can
benefit from improving the periphery.

There's individuals who do well was just

eight channels meaning that you don't need
as many electrodes you don't need as much

they don't need as much spectral
content somehow and they'll do fine.

So when you have more electrodes what you
can get and what you can benefit from and

what this represents is how we engage So

here are you know hair cells
of the cell bodies and

we're trying to engage we're
trying to activate these neurons.

So if we can refine that activation
through more like your own.

Space more closely together.

We can make the claim that we can respond.

We can further restrict that
spectral information and

this becomes important in music as
well as hopefully looking towards

timing because when I talked
about bandpass filter ing and

doing the rectification a lot
of timing information is lost.

So that's another critical piece.

So if we can activate a more distinct
neural segment of the neural population

for those who have a spine
kind of a more sparse

neural pattern we have a better
chance of activating them.

Furthermore as we age we will we
will experience hair salons so

if you put imagine you're
implanting a one year old

you want that electrode
array to be there for

the duration and
if you can improve the signal processing.

But then the electrode
are really can't match or

cannot adapt to that signal processing
you're not going to help the individual.

So that's another reason why we feel and
clinicians are really really pro.

High channel count.

I've never had a clinician tell
me I want fewer electrodes.

So and the other piece is that
if you have more imagine.

Some of us have had so much haven't.

If you have more point sources.

If you think about it that
way you have more ways to

engage the neural Pollak population
through more sophisticated stimulation

paradigms it's not just a single lecture
It's multi electro so you can do.

Wave shaping as well.

And current steering and

what you're after is engaging the neural
population that's all that this is about.

So this is some of
the earlier work that I did.

At the University of Michigan
when I pursued my Ph D.

degree so
I worked on the electrode array itself.

And we built this is sized for guinea pigs
so we built a fully implantable system and

back here is a microcontroller
with the D.S.P. core here.

And so this is a nice
demonstration of a technology and

it's for an animal model but
it's a it's a Palm Eric cable here.

It's a seven channel cable that went
to the electrode array and we moved.

Instead of having an implanted receiver
stimulator remote from the electrode away.

We moved all the stimulation
on the back end of the array.

This is a silicon based device so

we leverage the technology of
making integrated circuits.

So that we could more precisely
place the electrodes so

the current way of the current technology
all these devices are made by hand.

Three main in three main
manufacturers in the world and

every every manufacturer makes
electrode area by hand and

the electrodes are placed in a mold and
soldered.

So this is a way to use integrated
circuit based method so

that you can make multiple
many many devices honest.

Single wafer You can also dramatically
scaled down the electrode sites.

What happens is you can
scale down the sides.

However you have you have now you're

battling a little bit with
a thin thin thin film.

So it's we have a secure area here at
the back end because we need to do signal

transfer and you need something stiffer
when you're bonding to the back and

to the leads here but
the thin portion that can go and

spiral in the Coakley are is
only about you know this one is

a fourteen Micron base but it's probably
about a four a four microns weakness for

the insertion portion and what we did is
we scaled down the stimulating sites.

Dramatically in place them two to three
times more closely than existing arrays.

So this two hundred fifty microns
space is at least twice as close.

And so I won't go through all the details
here but the nice thing about using

the silicon based technology is
that we can make a shallow and

deep portion and when I say Boron
it's just Boron silicon and then

the beauty in that technology in Michigan
is that the boron serves us in at stop.

So essentially it tells you where to
stop eating through the silicon and

you can make some really
interesting structures.

So this devices are layered so
on top of the silicon we have

metal traces dielectrics for
insulation and.

Using the.

The stimulating site here.

Most all commercial devices are made of
platinum and we move to radium So imagine

if you scale down the the electrode
Let's move to the electrode here.

If you.

Scaled down the electrode size.

That's all the space you have to
do these nice reversible electro

chemical reactions of the electrode
electrolyte interface in the tissue.

So if you scale it down but you still
need to depolarize nerve fibers then

you're you're challenged with improving
the electrode electrolyte interface so

we moved to radium oxide it's it's more
challenging to work with because you

need to activate Iridium oxide which means
you need to build this and hydrous layer.

So you can have more valence transitions
and and basically do more charge transfer.

So it's a little different.

Now this is the electrode or
a that we tested in the guinea pig and

on the bottom here I put
a representative commercial array.

So in terms of the area we've gone
down quite a bit from the radio.

Mark side.

Excuse me from the platinum
to the radio mock site.

So once you do that we wanted
to see well are we are we.

Stimulating nerve fibers and

we did this in a guinea pig model and
we first did monopole or

simulation which just said he would know
monopole or stimulation is a couple.

Yes.

So one electrode site internally so
in the Coakley and

then a return electrode So you need to
have over a term past in the animal model.

So we found that we were on par
with commercial technologies so

we wanted to benchmark our
new method with existing

methods Bi-Polar make sense right.

Two sites two colinear sites in the.

Coakley.

And what we did is we
stepped through closer and

we went further and
further apart and as you go.

Further apart.

It takes less courage so
bipolar while it restricts the field.

It's a more energy hungry strategy.

So.

I guess I've got a back up for a second
before I go into my newer technology so

the silicon was great in
terms of micro fabrication.

But in terms of putting any torsion
on a silicon array you break it.

And it does it Has anyone worked with
silicon arrays in their own Nexus

probes here.

You'll feel that you know you'll know
my pain if you've worked with them so

you know if you think of a plane or
you know it's a silicon wafer and

now you want to twist it and turn it.

That's trouble.

You can.

You know do a lot in plain stuff so
still great from memes not so

great for electrode arrays
that need to twist and torque.

So once I came here I was on this this.

Path to improve the electrode or a so
we went to polymers because polymers

are much more flexible probably more
challenged in terms of long in tissue.

So we went to.

We started working with folks said
Georgia Regents Medical College of Georgia

because they wanted to start doing
more research with their residence.

So I said well let's try
to make a poem Eric or a.

And that's not so hard but getting it
into the cold glue it is so we look for

technologies or ways to introduce the
electrode array into the Coakley and so

on the top here what I was showing is
that the film or a made of quality of it.

And our first step was well you know
why don't we try to mate our way

with existing methods with commercial
arrays so mad L. Coakley or

implant company in Austria was able to
take these are nonfunctional arrays.

These are two types and
insertion test device and.

An insertion electrode So what they've
done for us is what they did for

us is they gave us the silicone in the top
one this is kind of a weird device and

no surgeon uses it even
though they make it.

It has a stopper.

It's to assert into the cold clear before
you put in the real electorate or a.

Surgeons don't use those because that can
introduce trauma and also via imaging

vill know what the path is in inside
the Coakley or for placing the array.

Then the less we use that to try
to we use that as a vehicle so

we essentially glued our thin
film arrays on to these two.

Silicone silicone based devices.

So.

We had mixed results.

And we introduce these
into human cadavers.

So through either the round window or
by drilling a hole in the cold clear and

so I probably should point out what you're
seeing here because that may distract you.

This is the silicone from the the carrier
those two devices a gave us.

And what we did is we glued our array
Here's the gold Those are you can see

the interconnect there and
this is an example of the array

Coming Apart from the insertion device and

penetrating the vascular membrane so
that none of this is good.

This is not how you want to go.

So we found mixed results also this is
a bit strange is that we had an implanted

controls and twenty nine percent of them
exhibited in their insertion trauma.

So suggesting that probably
the method of assessing trauma is

a bit confounded if something that had
no electrode or a exhibited trauma.

So we said this is probably
not the way to go but

we need to bolster we need to strengthen
these electrode arrays and so we thought.

Thought about it he said Well.

You know.
OK let's just try to put some

Let's just try to paint it with
silicone so you can buy silicone so

rather than taking two pieces and putting
them together where you know they're

coming apart just take
the electrode ray and

code it nicely called it was silicone and
see if that does the trick.

And so this is an example
of a really really nice.

Holder I guess to make to

apply the silicone it he
said to the thin film array.

And I like to put this here because
sometimes as engineers we get all excited

and think well this is what the clinician
needs they need an array that spent

because the Coakley has spent you know the
first turn coming out of that first turn

and so.

It's fairly comical because here
we are bending our arrays and

thinking that will help.

And making this nice pretty you know
here's a silicone on the ridge and

I wanted to show this just to give
you a sense of the interconnect

how we do the signal transfer which is
not really elegant but gets the job done.

So when the surgeons inserted them into
the cadaver the human cadavers they said

we can put these in because they're bent
and we're used to dealing with straight

arrays or having a stylus which
helps the Rays pretty curved but

as they insert it they pull on
the stylus that releases e-re So

they said Don't don't don't prick her
love for us we don't like this and

we had a Caesar resident projects
we had a new resident on it.

So we saw a lot of insertion trauma.

We also implanted into a completely
different structure so.

We said fine.

We won't curl them will just
paint them with the silicone and

we saw a very very good.

Good results with this with this approach.

We did this in the cat so
we're looking at a life prep

to see how well we could do polarize
auditory nerve fibers we only saw one D.

lamination meaning that the silicone
separated from the array.

And to move along a little bit
more quickly the research or

the experiment that we did is we first.

So this depicts a system level or
a block diagram level

how we do these experiments so
we start with an intact auditory system.

And we apply acoustic input to validate
that we are seeing a response.

So we were measuring in

the auditory brainstem response which is
something you can measure out externally.

And then we come in and
we damage the her cells with

an antibiotic the same
antibiotic that can cause her so

loss in humans and then we came
in with an electrical impulse.

So we applied charge balanced by Faiza
current pulses so the same amount of.

Charge you put in the tissue
you need to take back out and

then we looked at the response.

So number one you want to get a response.

So this is a representative
auditory brainstem response.

The first you know you see this
big jump here if anyone's done

electrical stimulation you see
the stimulation artifact and

then you start really listening so this is
a triple peaked evoked auditory response.

So number one we saw a response great.

And number two.

Well how much current is
it taking are we on par and

again we we did pretty well one
hundred seventy migrants commercial

stimulators only go up to two million
apps which is really really high so

about you know three three hundred
micro have says where you want

to be maximum in terms of what it
takes to stimulate auditory tissue.

So those results were promising.

Enough that we were continue.

Going to do this work.

With the thin film arrays.

So let me give you a sense

of what a thin film array looks
like inside the cat Coakley here.

So I like this.

So this because this is a micro C.T.
image with fifty microns resolution and

you can only put you know animal or
extracted human.

You can do this with a human being
just because of the resolution and

the radiation exposure but
I like to show this because it's

extremely rare that you're going to
be able to resolve electrodes sites

place two hundred microns two hundred
fifty microns apart in a C.T. image.

So this is then this is available to those
of you doing animal studies it's at Emory

so they will allow us to access that if
you want to image some fine structures but

highly recommended.

So what that enabled us to do was take

measurements from let me point
out a few structures here.

So this is the base will turn and
you know if you have an electrode or

a your target elements are in
the middle of the central structure so

we did pretty well during the implantation
where we came close because if you back

away at the at the lateral
portion you lose the benefit of

high density because you have to go so far
in a way to stimulate the auditory nerve

fibers that you're going to be spraying
a lot of current into the Qalqilya.

And just a representative two
hundred fifty micron spacing so

we could image the electrode a race.

So moving on from the cold

clear implant work where we are continuing
to look at improving the electrode

arrays is the vestibular bio
system work and I think the.

This is where we get
the most translational And

I think this is some of the most
challenging work to understand.

So what we're after is looking at the
vestibular system so here's a coke Leah.

And here's a vestibular system.

So what is our vestibular system do so
you know.

Balance folks said Vallance and
I I mean I don't

really think about what my vestibular
system is doing until I'm dizzy.

One day.

So it's really which
this low key system and

the moment you have a deficit is
when you really really feel it and

says it's different than
the auditory system in that respect.

And so we're looking at that both
from an implantable way as well as

externals So you know this is a very
prevalent so dizziness is extremely

prevalent forty percent of Americans and
also in the aging population.

This is a significant problem and balance
you know if you if you have trouble

with balance you're going to walk slower
you're going to not want to do as much and

it really impacts your quality of life and
it's you know false or.

You know a big problem in
the elderly community as well.

So Kim of the the good news
on vestibular function and

this is a lot of the work that I do
with Emory is that you can train

the vestibular system to compensate to
overcome to engage other senses because

it's very much a multi multi modal
integration for the vestibular system.

So that's often if you have
the unilateral function

dysfunction only one side you can kind
of get the other side smarter but

the bad news is those who have
bilateral vestibular dysfunction.

They have no therapeutic options at.

So that's of a similar processes
work that we do in our lab.

So moving to the good news first is that

this is an example of a series
of exercises that a group.

Susan Hermann around to said at Emory

pioneered in that's where
you can engage your.

Visual system.

To overcome a poor of a serial or
signal and

the nice thing about this research is that
you can train individuals to overcome for

this law to compensate for

this loss the challenges you know
although we say here seventy five

percent improve with gay stabilisation
exercises twenty five percent don't.

And their theory is that while
twenty five percent don't.

Well we don't know if they're
just not doing their exercises.

We don't know if they're
doing their exercises wrong.

And so we've developed a really simple.

Head motion monitoring device so
I come back to my memes roots and

get to use you know accelerometers and
gyroscopes and

measure head motion in
a really simple low cost way.

And we can get a nice signature a nice
profile of what it individual is doing

outside of the clinic.

So there's a huge explosion really

of cell phone based you know
of a similar rehab devices and

all these different tools but you know if
you take an individual who's sixty five or

seventy five and you say well what do you
wear kind of the same path that you would

when you're gardening and you know
just turn it on and do your exercises.

It's very low cost.

They're much more agreeable than if you
hand them a cell phone or tablet or

something that is just too too much for
them.

So we're after low cost and simple.

And so we do have a clinical
study on putting these

this pad on patients the other
thing is that it helps.

Normalize.

P.T.S. because they've routinely
they don't get this information

unless a uses big very expensive
system that they can't send home.

So moving to the case where
there are individuals who need

a bilateral who have bilateral loss.

We can develop a vestibular
prosthesis the vestibular system

is different in the sense
that hair saw a loss.

So the hair cells are the transducers But
rather than modulating you're not.

Stimulating place you're modulating rate.

So if we can capture external head motion
translate that into a neural signal and

apply that to the vestibular nerve in
a position specific way position meaning

how we are moving in space then we can
use a similar paradigm to Coakley or

implants to offset vestibular loss.

It's all speed up a little bit
here in the interest of time.

And so this is an example of
the peripheral vestibular system.

These are the hair cells and

we have this amazing structure here
called a coup people so our semi-circular

canals that detect angular head
motion in are in three points.

It's deflection base.

So we have this nice right from
here occluding So we have a canal.

We have this diaphragm and as we move
the fluid in the canal is an inertial

element so we move in the fluid lags and
we bend these hair cells.

We bend the steerer silly
on these terror cells and

we marginally the firing rate
of this to Bill or neurons.

So we just have a tonic rate at about
one hundred hertz and if we can kind of

modulation down marginally rate and the
beauty of our system is we have two ears.

So we modulator we can go up and down so
we get that we can amplify the signal.

And so what we've done in terms of looking
at a vestibular prosthesis is to me.

Challenges and one is that when
you have inertial sensors and

you implants them they're
very power hungry.

So what we did is we thought well how
can we model this vestibular sensor

as a member structure and
I'll skip over that a little bit and so

here's a representative semicircular canal
so this is a first order hydrodynamic

model and if we can get the same
frequency response as well as

deflection we can use that to
code the angular head motion.

Instead of using a gyroscope So
our device is purely passive.

And this is a little bit
of modeling of it but

I'm just going to skip to the memes piece
and so I can close up pretty quickly.

So you can make a vinyl emetic
semicircular canal using

a diaphragm and you can make that
out of P.M.A. we did this with.

Maxine McLean and we can make it as you
weight mold so essentially you're making

a similar structure to the to
the semicircular canal and

you have the kind of the you have
the bendable element here and now you

have to think about well how do I sense
that so it's a capacitive based sensor.

So one side the flex so the capacitance
changes and you have a reference electrode

So you're changing very similar demands
a lot of memes devices you sense

a change in capacitance I
make it sound really easy but

it's actually not very easy
to get the sensitivity you

need because men's devices are linear
We linearize our world and memes.

But you and I are not linear beings.

So the cuckoo actually has
a huge dynamic range and it's.

One this is very much a work in progress.

I was hoping my.

Soon to graduate student would be here and

he could talk a little bit about
the signal processing aspect of that work.

So we do circuit design in our lab.

So we pretty much do it all we
do memes signal processing and

a lot of circuit design here on
building the vestibular prosthesis.

And I'm going to skip through this
just a little bit and talk about.

Well I think I just skip through all
the circuit work here but I wanted to

ignore all IJ the broad range of trainees
that we have in our research group.

Greg graduate students from Georgia Tech
as well as Little earned dollars

You residents and medical students which I
think really in Rich's the work that we do

so we can better translate that from
the engineering domain to the medical

domain as well as I've always really
proud of the number of undergraduate

students that have gone through our
lab because they've really made some

impressive contributions and two of
those students are you hearing impaired.

One is a Coakley or implant user and
another uses hearing aids and

we have a sign language interpreter
during our group meetings.

So that's all was pretty interesting for

us to get that insight
from these individuals.

So I'd like to acknowledge the funding
sources and assess and I ate as what is

as well as a net L. Coakley or implant
company lot of donations for us and

close with the overview slide of how
our group tries to integrate as much

engineering as we can to meet the need
in sensory neural and vestibular loss.

Thank you.

Any questions.

Chris.

And.

So there's a lot of work.

House here Institute was
really Pari pioneering that.

And so they're using thin film technology
in implanting the auditory nerve.

So while it's a less
constrained environment

the surgical approach is more difficult.

The surgical approach getting to the
auditory nerve is a bit more difficult.

So you tell you about in the periphery or

you're talking about further up
in the auditory chain because.

So in the theory of cool Occulus
that we actually have the frequency

to place mapping in terms of these layers.

So they're doing in Faircloth Kilis
they're implanting in the inferior click

yes they were hoping to
see a better resolution

of frequency as well as lower
the threshold because in the Coke.

You brought it.

I mean you hit the nail in the headers
to your constrained years.

You're not targeting the neurons as well
as you could they have not seen that.

They have not been able to drive gel
you want to drive down threshold and

you want to improve resolution and
on both counts.

They haven't seen it.

I'm not sure why they're
not resolving that.

As well as they would like because they're
implanting directly into the nerve or

into the in for cool occulus.

They're not getting
the resolution that they need.

Did you have a comment.

Go ahead and.

So I think.

Ideally the further you go up the pathway.

The less you have to do in a sense because
there's levels of integration as you go

up the pathway the challenge there
is the surgical approach and

the current paradigm is
the closest place that you know

you can activate the neurons the auditory
nerve fibers is where you go.

So the individuals who don't have so

you may have a acoustic neuroma on
the vestibular Coakley or nerve.

So you can't conduct that information
though so they try to implant further up.

I think that as well as you
try to improve the periphery.

If you can't improve it in higher auditory
centers you do have a loss but I think

people are much more afraid because you
know say two hundred thousand individuals

are implanted with a Coakley or
implant surgeons understand them better.

There's a lot more data on them and
so I think that's why there's it's so

expensive for a company to take on the
liability and I think that's why we tired.

Try to mine the heck out of the periphery
because it's less invasive and it's safe.

But that's an important question
because it's perception ultimately.

Yes.

So that's a really important question.

There's a lot of analysts and
I age finding in her cell regeneration.

There's stem cell approaches.

It's mixed.

There's a few groups in the country
doing some really novel work.

They have been able to
regenerate hair cells and birds.

Some work in mice but

that's a very very active area
because if you can revive it.

If you can revitalize that population
that's really what you need.

I don't think.

Someone recently told me that they
recently told me that they heard it.

Talk about hair so regeneration.

But it was not it was in a mammal but
not in a human.

So we are losing them for
sure all the time.

We are.

You know that we're we

promote the neural population but in terms
of hair cells the sensors and cells.

I don't think that we region.

Yes.

Yeah

yeah.

So if you look at the.

The Coakley or

electrode array we deliver eight
hundred thirty three pulses per second.

So you know all that fine
timing information is lost and

met Al has the most sophisticated.

Most sophisticated signal
processing paradigm where they do.

Think it's a camera.

The name of the filter design but that's
why people are also going after bilateral

implants because we hear through two ears
and in we need that timing information.

So a lot is lost a lot is lost but
we can talk more.

I mean it's easier for
I show you a picture so.

Yeah

so

those recordings are.

How do I say this in a good way.

So the best way to map
a different stimulation paradigm.

Those recordings are monopole or like.

So the I can envision adding more

spectral information to those
recordings as you improve

the activation meaning that if
you can do bi polar stimulation

Ideally you should be able to
more like having a tighter filter

as well as more then
more bandpass filters.

Now this is not what you see in patients.

Because of the integration.

Because if you do research in animal
models and compare a model polar with

bipolar information bipolar stimulation
and if you stimulate in the Coakley and

record in the click Kilis which
has a frequency to place math.

And you will see if you look at the
spatial tuning curves they will be tighter

for bipolar Tri-Polar they will.

But if you take a patient and
you apply model polar and

bipolar most patients
prefer a monopole or.

Although if you look at the profile you
are being more restrictive one reason

is that the implanted electrode arrays
today are so far from the neural elements

that the patient just really needs to
get a broader activation for perception.

So that's a really important
question with out a good answer yet.

So I have a counterpart who's
who's doing a lot of research

on Tri-Polar simulation and
in patients who are poor performers.

It helps.

But not in the good performer
Tri-Polar just means using three colon

electrodes and

you have a really a much tighter spectral
much tighter peak in the stimulation.

We form.

This is.

So it doesn't because first speech.

I mean the sweet spot is about three

three hundred hertz to
about a killer Hertz.

Yeah yeah because the array is sort of but
they play these games.

There's so much work that's been done
on how you can and how we can adapt.

So what they'll do is they'll move.

Although the sound we hear has a specific.

They know that that's not
what they can stimulate so

they play these games of shifting it down.

And that and it works.

Yeah but I wish I had an audiologist
here because it is an art.

I mean audiologists are amazing
individuals surgeons put it in their kind

of you know I'm done.

But the audiologist is
a person who fits the patient.

OK last question Chris.

We're trying very hard.

Yes yes.

So what it is is it mimics our system but

we would never penetrate a semicircular
canal canal and put it in there.

So it's a model it's
inspired by biology but

it's an inertial sensor and
it would be implanted.

But not replacing It's not like
a banana coup pill or something.

So if you could insert it and

couple that if you could insert it without
compromising the semicircular canal

it would make sense but there you can't
really like open them up and put in this.

It's not like a retinal prosthesis.

Where you can put it in and

close it back up but
if you penetrate the semi-circular canal.

You can't surgically.

You know fix it up if that makes sense.

So that's I talked about that.

At the beginning my work and the
clinicians looked at me like I was not so

I guess so.

So no auditoria is my first love
vestibular is a strong need.

So I try to kind of who make it to you
can see how I ran out of time when

I got to the vestibular but that's
because of a similar so much harder but

much more exciting.

I mean my clinical work
is invested biller and

my auditory work is more basic
science universal interface.

But I'd be happy to stay an answer
some questions for the for

those who want to stay but thank you.

I enjoyed this appreciate.