today very pleased to have Bob Gross
here with us Bob is an MD PhD trained at

the Albert Einstein College of Medicine
and is now a clinician scientist at

Emory University where is the MBNA
Bauman professor in neurosurgery and

neurology he's the president of the
American stereotactic and functional

neurosurgery society and at Emory is the
co-founder and director of the Emory

neuromodulation & Technology Innovation
Center in Tights which is very unique

combinations of clinician scientists and
engineers seeking new treatments new

therapies for diseases I encourage you
to check out their website to a Google

and in Tison Emory and find out what
they're all about some really

interesting events we bring clinicians
and engineers and scientists together

today we'll be hearing about the work in
his lab the translational neural

engineering lab at every one University
and very pleased to have him here if

you're joining us remotely I encourage
you to please mute your microphones so I

can hear someone fiddling around in the
background thank you very much Bob

thanks for joining us welcome thanks my
pleasure thanks for inviting me Chris

all right well I will I'm not an
engineer even though I'm program faculty

in engineering and and PhD advisor to a
number of bona fide engineers but I was

majoring in engineering for the first
semester of college so you eventually

returned to your roots I found my way
from there to computer science and then

ultimately to neuroscience which is
where I ended up but it's interesting

how things go around and come around so
I'll talk to you about neuro technology

really at the interface of science and
medicine so that's really where I focus

anything that I do in a research project
has to have an application as you will

see some of the applications are
immediate now some of the applications

are quite far off in the future if at
all but we always have that angle of

approach it's fine for other people to
do discovery science

and and even to do it of course
engineering that doesn't have a clinical

application but that's the unique focus
of my lab I do have some disclosures in

respect to some of the things I'm going
to talk about really clinically and this

reflects the fact that that if you are
going to deal with technology at the

interface of medicine you will be
dealing with commercial entities

commercial entities in our capitalist
society and economy are important

because they drive the application
there's the saying many of you might

have heard no margin no mission and so
we have to work very closely with them

and so so I work as a consultant for
some of these I don't have any actual

ownership in any of the companies but in
particular I'll talk about some

technology from Medtronic MRI
interventions and some of these others

and neuro pace so the focus of my
applications today are epilepsy epilepsy

is characterized by recurrent seizures
where you know you are seized ultimately

named after being seized by the devil
essentially so we still use that arcane

terminology and when you have recurrent
seizures you have epilepsy and it

affects a lot of people about one in a
hundred people have epilepsy per se it

involves synchronous discharges of
neurons throughout the brain and as and

we see that on an electroencephalogram
which is recording electrical activity

at the surface and usually it sort of
marches off in this fashion where it

looks pretty flat it's not completely
flat or you wouldn't be doing very well

and but it has these slow oscillations
when all of a sudden large populations

of neurons begin to fire together and
and of course it's not every neuron in

your brain that's firing together if
it's just a very small subset but if you

imagine in this room if only three
people started whispering at the same

time the same thing than it actually
would be quite distracting to the rest

of the rest of the people in the room
and if 15 people under that of the

hundred in the room started whispering
the same

I wouldn't be able to continue to talk
over over that cacophony so it doesn't

take much synchronous behavior to to
have an epileptic seizure and and that

may be one of the problems in treating
it and that is that we you have to

identify these small populations of
neurons that are firing together the

epidemiology is such that the actually
the cumulative incidence or the risk per

person is 2 to 4 percent so if you think
of if there's a hundred people in this

room four percent or four people in this
room will either have had or and I won't

ask who's had a seizure but either have
had or will have a seizure during their

life or have epilepsy during their
lifetime and the risk of having at least

one seizure is 10 percent so 10 people
in this room will have a seizure at some

point during their life we call it an
emergent property of the nervous system

so in doing the things that you're
trying to do perhaps right now which is

learning and remembering what we're
doing that sets the system up to have

this abnormal types of overly
synchronous connection between parts of

the brain now two-thirds of the patients
are actually treated with medicines and

that have epilepsy and don't continue to
have seizures upon taking medication but

one-third of the patients that we see
have medically refractory seizures so

one so thirty percent thirty to thirty
three percent of the people that we that

have epilepsy is treated with anti
epileptic drugs will continue to have

seizures despite that and that has not
been changed since the very first anti

epileptic drug potassium bromide was
discovered you know close to a hundred

years ago so despite decades of drug
targeted drug drug discovery we have not

moved the marker on that so what do we
do in people that have drug-resistant

epilepsy well that's where I come in
okay so we have surgical treatments and

this has been going on really since the
dawn of the field of neurosurgery in the

late late 19th century really in the
1890s

and we've been removing large chunks of
the brain

this shows removing part of the temporal
lobe and and then the medial temporal

structures or the hippocampus over here
and this shows a more selective approach

to that and you can see we remove large
chunks of brain in order to treat this

problem and you can imagine that
something like this might have some

adverse effects on patients functioning
okay strikingly sometimes it doesn't but

that's really where we're trying to what
we're trying to deal with is this

problem and sometimes we do things even
more radical this is called a

hemispherectomy so back in the 1930s it
was discovered that if you remove a

whole hemisphere of the brain and the
epilepsy is located in that that you can

stop seizures okay that's pretty
dramatic approach nowadays we do things

that where we leave the hemisphere in
and we just disconnect it and sometimes

even more subtle where we just this
every place is a black blue this is a

ventricle but every place is a black
mark is cutting through the pathways

that connect the epileptic area to the
rest of the brain so called functional

hemisphere Atum II but that's pretty
radical and we do things like cutting

through the corpus callosum that highway
that as you know connects one side to

the other so a corpus callosotomy can
divide the brain and disrupt the

synchrony of bilateral discharges and be
a reasonably effective operation

although with adverse effects like the
disconnection syndrome which is actually

what drew me into neuroscience when I
was an undergraduate as many of you are

was being very interested in hemispheric
asymmetry that's what drew me into this

field to begin with so this is very
effective and this shows you a

randomized double randomized single
blind control trial where patients were

randomized to just continuing medical
treatment and looking at the percentage

without seizures and if you take this
group only 8% of the patients became

seizure free with continued medical
treatment whereas 56 really 64 percent

but the details are not important became
seizure free with a temporal lobectomy

so that's pretty effective that's good
news for patients I can tell you

two-thirds of the patients will be see
you're free and that's just the temporal

lobe and that pertains to all operations
we do basically if you look at sort of

meta analyses or systematic reviews 65%
becomes seizure free and 25% becomes

seizure free and off their anti
epileptic drugs which is a cure so

that's really good news on the one hand
on the other hand it creates problems as

I as I said removing large chunks of
brain can have adverse effects so this

shows the picture of after we remove
these this medial temporal structure

that there can be white matter changes
in the in the areas that are in the

vicinity of the places we're operating
on and if you actually look at the

amount of what we call collateral damage
off target effects and the more

collateral damage you have the more loss
of function you have and the less

collateral damage you have the more
gain-of-function you have actually from

from rendering patient seizure free so
this is at a cost and these are what we

are taught what we really are trying to
do is trying to maximize the benefits in

anything that we do and I told you we're
at about 65% that's over here well we

want to get up over here and every time
we take out more brain tissue as we take

more brain tissue out we get two more
deficits so this is a patient that has

benefits but that is deficits this is a
patient that has no benefits and no

deficits and somewhere in between is the
sweet spot where you get benefits and

minimize the deficits and that's really
what we're trying to do by engineering

the system a little bit more
so minimize off target collateral damage

this shows someone going after a fly and
ruining the room so that's a depiction

of the collateral damage and we want to
optimize the on target therapeutic

effects and we also want to decrease the
pain and suffering that goes along with

surgery because as you can imagine when
I take out a large chunk of the brain

there's a lot of pain from the surgery
and and you know people take a long time

to get back to work they stay in the
hospital and because of that pain in

offering a lot of patience never come
forward for surgery so we are very very

under penetrated in the market of
patients that could be helped with

surgery because we might say well I know
you're having seizures despite your

medicine if I do an operation they say
no no no I want nothing nothing to do

with that not going to brain surgery or
even perhaps more importantly the

referring doctors are very scared of
surgery so patient might come in to see

their doctor and say hey I saw a report
on CNN about how this could help brain

surgery in this area no pay people end
up as vegetables next question I have a

new drug for you so so this pain and
suffering in the collateral damage

decreases the accessibility of patients
to getting treated and people die from

epilepsy every day so the first
engineering advance I'll talk about is

one that was not my own but because I
was really at the forefront of dealing

with new types of technology I was one
of the earliest adopters of this therapy

called Laser interstitial thermal
therapy so there's a very long history

of using lasers in in neurosurgery it
dates back to the invention of the laser

back in the 1950s and these very large
and cumbersome devices much of this work

was done at Bell Labs and the Nobel
Prize was given to the first functioning

laser and then there was as you'll see
becomes important the invention of the

light emitting diode actually and the
LED laser source is very important in

making these available for use and you
can see these early lasers were not

particularly elegant devices and and
they also were not very minimally

invasive this is hardly what we would
call a minimally invasive neurosurgery

and you can see there's a large amount
of damage and the original lasers were

actually developed as cutting
instruments okay they caused immediate

vaporization of tissue and because of
the cumbersome Ness of this and the fact

that they were only used as
cutting tools and we have other cutting

tools without having to deal with
something like this

there was a very slow adoption we had
lasers when I was a resident back in the

1990's but I kid I know we had them I
can't say I ever used one during a case

because they were cumbersome and not
particularly useful but they have the

potential for precise delivery of
thermal energy and and with the

advancement in certain types of
visualization technology they ultimately

became more useful so the idea of using
a laser instead of as a cutting tool but

as an interstitial tool was developed in
the 1980s some thirty years after their

invention okay the idea here was to put
laser light into the tissue and release

photons at a much lower energy level
where those photons would cause

localized heating and that heating can
cause a thermo coagulative necrosis or

damage so this idea was born in the
1980s and as compared to the photo

mechanical damage that goes with these
short pulse high-energy laser exposures

that cause vaporization and instant
perforation

this causes photo thermal damage which
is a medium exposure level

continuous-wave laser and tissue
coagulation thermal denaturation

and this looks quite a bit different
than the other type of picture I showed

you okay where there's an area which has
lost its blood supply and where the

tissues have been heated to the point of
enough destruction that they will

ultimately die so rejoining the timeline
here this was used in the 1980s to see

to treat certain deep brain gliomas but
it wasn't particularly useful because

you couldn't tell how hot you were
heating the tissue okay if you're doing

what's called a radiofrequency ablation
you can which uses a metal probe you can

put a thermocouple in there and monitor
the temperature and temperature is

critical but until we could do that with
with lit it didn't become useful but

there were certain other technical
advances engineering advances that

occurred in the 1990s which is the
initial use of MRI and you could observe

the effect of the laser but not
so afterwards you could see the

destruction but you couldn't see it
immediately while it was going on but in

1995 now Toshiba released the paper
demonstrating the ability to observe

temperature based on the phase change of
water molecules in the tissue so you

could actually use the phase change in
the MRI as a thermometer and you can use

the MRI as a thermometer and this
married up very nicely with these with

this laser interstitial thermal therapy
and this shows what that looks like so

the system can use fast used use MRI
sequences which have high susceptibility

so not your typical one just for looking
at structure which requires a lot of

fine-grained detail but the type of
sequence you would use for an fMRI with

high temporal resolution and you can you
can in a linear way exactly relate the

change change in temperature to the
change in the phase of the protons

during heating so this very nice and
tight linear relationship allows you to

look at change of phase and meet a
temperature change and this is what the

imaging looks like so it's obviously not
very good for structural imaging

appearances you're looking at a phase
map here so it encodes all phases

changes by using warm and cold colors so
this is a heat map of a heat map that's

a joke I always telling you I saw one
guy chuckle over there at least in some

ends in a crowd of engineers some people
chuckle with that so and so this is what

it looks like as put together ultimately
by the company that sells this and this

is a thermal map right over here and
this is now I just looking at a thermal

map I can't tell exactly what's
happening but if I overlay this on top

of a structural image and you have in
there Co registered in MRI space so

there's no no movement allowed from this
thermal map to the archival or

historical imaging set then I can see
that this damage is happening to this

area of the brain and now I'm using the
temperature change but if I as a surgeon

I'm sitting there and reading the
temperature of the brain I can't

what's happening I can just tell that an
area has gotten to 60 degrees or 40

degrees and what I want to know is what
has been damaged well it turns out that

the chance of being damaged is a
function of two things one is

temperature and time the longer you have
something at a particular temperature

the more chance it will be damaged and
so we're looking at the area under the

time temperature curve and that is
quantified by this integral now which

I'm sure some of you know what that
means

and and this basically looks at the area
under the time temperature curve and

what we do is this is the actual live
thermal map and I can see that this area

is being heated up but by calculating
the Arrhenius model we can depict the

areas that have been faded to be damaged
because they've reached a critical mass

a critical amount of time temperature
and that grows in front of us so I can

see the area that has been faded to be
damaged here by this damage zone

estimate so this takes this advance and
then Couples it together with laser

interstitial thermal therapy first done
in the late 1990s and allows us to use

develop this mr thermal imaging and
couple it together and this is what

really drove the whole field forward so
these are the critical advances that

have moved this from a therapy which was
in the 1980s not something usable and

just try it on a couple of patients with
gliomas to something that we're using

every day now then now the diode laser
was an told you about the LED that was

really critical because you have to be
able to reel these things around I'm

bringing this thing down to the MRI
suite

I can't wheel that used laser down to
the MRI suite so that was an important

advance having cooled catheters certain
automation and these real-time damage

zone estimates are critical to this
therapy so in 1997 in the late 1990s

Asha Gowda and Roger McNichols who are
two biomedical engineers working down in

Houston they both trained at Rice they
put their heads together and they found

that a company called biotechs based on
this application

they eventually spun this off as visual
the

you'll lace clever name laser visual
visualizer so visual lace and a visual

lace was eventually purchased by
Medtronic and the first procedure using

this turnkey approach where the laser
this is the diode laser over here this

is the imaging city or the computer
display and the computer down over there

and there's a peristaltic pump to cool
and this was eventually used in 2006 it

was FDA cleared by a 510 K mechanism
someone wants to ask me about later on I

could talk about that and the first
nerve surgical procedure was done in

2008 in the United States and the first
time it was used for epilepsy was by a

friend of mine Dan curry down in Houston
in 2010 and we did our first procedure

in 2010 as well this is what it looks
like so the laser is introduced through

a fiber optic and this is done by
stereotactic technique and you can see

the laser at the end of this the the
fiber-optic comes down the catheter and

then there's a diffusing tip at the end
which diffuses the laser light into the

tissue around and where the photons
bounce around and cause localized

heating by their interaction with
macromolecules that absorb at that

particular wavelength of light so though
it's a very particular wavelength of

light that's used and this is how the
whole thing works in that I told you

about vaporization of tissue at greater
than a hundred degrees at 43 degrees

nothing happens but somewhere in between
you get to this time-dependent thermal

damage in this area between 45 and 60
degrees so this again was what the

thermal map looks like the damage zone
estimate this is a already showed you

what that looks like this shows a very
nice screen for for tracking these and

this shows what it looks like in actual
practice okay this is a coronal view and

this is an axial view and you can see
the thermal map you see the fiber optic

being passed into the hippocampus and
then you see the irreversible damage

zone beginning to accrue in the
hippocampus and I can follow that and

when I've said enough is enough and I
can also make

sure that there's no damage to
surrounding structures like the visual

pathways over here or the brainstem over
here because I can directly visualize

them I then pull the catheter back and
then and then you may want to yeah yeah

someone on the remote side if you can
turn your microphones off please you are

echoing over here thank you
I think we made some progress great okay

now now this is introduced by various
technologies some of these these are all

stereotactic tools that we use to put
these things in and this is the way the

system was designed and this is how we
use it so this is another nicely

engineered platform so this is called an
MRI targeting platform and this allows

us to finally adjust this in the MRI
scanners way the way I do this procedure

is I bring the patient directly to the
MRI scanner and we're able to do the

surgery in the MRI and we attach this
frame to the patient's head and this is

basically a gun sight so this has a
gadolinium containing tube over here and

I can see based on that gadolinium tube
where it is virtually projecting and

then I can adjust that using these dials
to adjust the azimuth and declination

and the translation until I get it just
on target and then I can introduce this

fiber-optic right into the into the
brain through that technique and it

works out quite nicely this shows what
the damage zone looks like this shows

what the what the actual damage is and
these overlap very very nicely so the

damage done very nicely predicts the
actual damage area and so we have looked

at the outcome of this and this is a
paper we published earlier this year

with the help of my graduate student
Matt Stern who's sitting in this room

Matt raise your hand so it worked very
hard on this and what we showed was that

about 60% of patients have become
seizure free

for at least a year using this technique
so it's about the same level as as we

were seeing with open temporal surgery
well have we moved the marker as far as

that collateral damage that I told you
about so this shows the work of Dan

drain and our group looking at the
deficits and basically object

recognition and naming and if you go
down over here you've had a decline in

object recognition and this population
or all patients that have had open

temporal lobe surgery the kind that I
showed you earlier where we actually

removed the temporal lobe so all of
those patients experience declines in

object recognition and naming whereas
the patients that had laser ablation

experienced no decline and in fact
experienced a recovery of some of those

functions so we are able to get a
reasonable degree of effectiveness and

decrease the collateral damage and this
has substantially increased the

acceptance of this operation whereas
previously I would see patients in my

office and I would propose doing a
temporal lobectomy and they would say

many of them would say no thank you I'm
not interested and they would continue

to experience seizures whereas now
nobody is refusing that surgery anymore

and that's actually from our
neurologists we don't have any patients

that are not interested in surgery
because of the laser so we have a lot of

different things that cause epilepsy
aside from temporal lobe epilepsy

there's tumors ganglia gliomas there's
this cavernous malformations hamartomas

malformations of cortical development
and this is a useful tool for all of

these actually so this shows a very
deep-seated hypothalamic hamartoma which

you can imagine if i were to do this in
an open surgery would be quite difficult

to get to and associated with a lot of
potential for injury and this shows the

damage zone the thermal map and the
damage zone associated with that and

this shows other types of deep-seated
lesions so once you have a hammer a lot

of things look like nails and we are in
the process of spending a lot of time on

on exploring other usefulness of this
device I showed you a picture earlier of

the callosotomy of dividing the corpus
we can even use this for dividing the

corpus callosum so this shows a patient
with five different bolts in their head

for introducing the laser over five
different places and you can see that

we're able to do very nicely oblate and
destroy the whole corpus callosum and

we've done this on about 13 patients now
and it also is benefitting them as well

well what do you do though however when
a destructive procedure is just not the

right operation for a patient so there
are patients for example that have an

onset zone right in their motor cortex
region of their dominant hand all right

well there there's no chance of
developing a therapeutic window damage

that gets rid of the epilepsy also gets
rid of the function so no therapeutic

window what are we doing those instances
so this is on to the second engineering

advance so new approaches to neuro
stimulation well this is a beautifully

engineered device that was developed in
part by intellectual property developed

here by blah a guy named Brian litt
who's now a professor of biomedical

engineering and a neurologist up at
University of Pennsylvania and while he

was here he developed helped to develop
the the idea that if you run an EEG and

you detect the seizure onset you can
stimulate that area electrically through

through a pulse generator generating an
electrical signal in the brain and that

will actually suppress the seizure so
there are certain seizures that if you

hit them hard just at the right moment
you can actually break the seizure and

so this was encapsulated into a device
called a responsive neurostimulator that

runs a continuous EEG and with it and
this is what Brian's contribution was

the closed-loop feedback activation of
the unit in response to detecting a

seizure and actually what's nice is that
the patient can actually this will store

that information and the patient uploads
the information to the internet where I

can then go online and see if the
patients had any seizures of over the

last 24 hours and how effective this
device is so this is a skill

mounted device it has to be skull
mounted because it's recording neural

signals and you need to have the head
stage pretty close to the electrode to

minimize the noise floor and this shows
electrodes in this particular instance

in the hippocampus one on each side one
on each side and this shows a patient of

mine that has a electrodes that are
implanted into the motor cortex and this

is just such a patient as I just
described

she had onset in her in her okay
did you mute me or something oh thanks

Chris
I'll go back and redo all that so so

this shows a patient with electrodes
implanted into them into her motor

cortex and the device implanted in the
skull and she has had no complex barceló

generalized seizures since the surgery
and a reduced frequency of focal these

focal motor and sensory events so she's
had a nice response but she's not

seizure free completely and patients
want to be seizure free so this is a

randomized double-blind controlled trial
of sham stimulation not turned on and

stimulation that was turned on and this
showed about a 35 to 40 percent decrease

in seizures in this population of
patients and this was sufficient to

garner approval by the FDA for this
device this shows the percent change in

seizures from their pre implant period
and you can see that some patients have

no response or very little response few
patients actually got worse most

patients got better this is a hundred
percent response over here in a handful

of patients so about 10% of patients
became seizure free with the majority of

patients 65 percent having a reasonable
greater than 50 percent benefit so if

you have an onset in an area that cannot
be removed and you are not someone that

I can cure of their seizures this is a
great device except that when I tell

people that the happy news is about 60
you get about a 65 percent reduction in

their seizures except for a handful of
patients they say that's great I'd like

a hundred percent please
so we have to try to continue to move

the marker this is great but we have to
keep going forward so how can we

increase the effectiveness of
neuromodulation for epilepsy so that's

been the focus of my group so this is a
multi electrode array on which is

cultured neurons cortical neurons and
this was work done by Steve Potter who

was a faculty here at Georgia Tech up
until just a couple of years ago and

this shows what happens when you put
neurons on these micro electrode arrays

so this is stacked 64 channels okay and
you can see that after a time this shows

the after a time period that they start
to fire very synchronous bursts of

activity so that when they start to fire
they start they entrain each other and

they fire together so if these were
firing all separately you wouldn't

actually see anything here really but
the fact that you see these big black

mark shows that these are synchronous
populations of neurons firing together

and what Steve showed was that if you
stimulate them all synchronously that at

a low frequency that you can actually in
train the whole culture together so it

actually makes things worse but as you
increase the the signal size and as you

increase the number of electrodes firing
you can begin to back this off so that

you get basically close to this
situation and this is with single site

stimulation when you just single
stimulate one site and then he looked at

what happens if you stimulate across via
all the arrays all the electrodes across

all 64 so in a distributed multi-site
fashion to basically try to break up the

activity and he did this in a feedback
controlled way where he actually used as

a biomarker the spikes per second or a
wide so we're trying to get rid of these

synchronous spikes and we can look and
see when the synchronous spikes go away

but you can also use this surrogate how
many spikes per second are happening are

a wide and as you pass the electricity
into it more and more electricity you're

getting more and more spikes elicited
and that

is a surrogate for actually decreasing
the amount of synchronous behavior as

well so as you turn it up higher to 500
spikes per second or a wide you're able

to get rid of the synchronous activity
so you can't see these action potentials

here but the fact that there are no big
black bands tells me that there's no

synchronous behavior so Steve was able
to eliminate synchronized bursting in

these cultures by using a surrogate
spikes per second or a wide as he

delivered multi electrode multi-site
distributed stimulation so when we

started working together we said well
how can we get this to the bedside

alright how can we move from the bench
to the bedside well it'd be kind of hard

to go into a bunch of venture
capitalists and tell them that hey I've

got this thing and it works great now I
want to stick this thing into someone's

head okay so you need a translational
pathway and so we started out with the

rat and this is the work of John Ralston
and saranya Desai did this in in the

group John Ralston is now a neurosurgeon
who's an MD PhD got his PhD here and

he's now a neurosurgeon at University of
Utah and tirana Desai is a PhD and she's

not working for that company neuro pace
that made the response of neuro

stimulator I just showed you before that
was implanted in the head alright and

during their thesis work they made these
multi electrode arrays and they we put

them into the hippocampus in a rat and
we passed patterned fader frequency

stimulation and we hypothesized this
would suppress the seizures we used a

rat model called the tetanus toxin model
where we infused tetanus toxin which

changes the ratio of excitation to
inhibition in the favor of the former

and leads to seizures in a close to a
hundred percent of the rats that we do

this to so that's our animal model and
what we did was we looked at pre

stimulation and then stimulation and
then post stimulation and look to see

what the effects of stimulation were and
we did different types of approaches to

this where this was continuously done
versus being two minutes on two minutes

off during that period of time and we
also looked at a mic macro electrode a

big electrode like the electrode I
showed you before when putting in that

responsive neurostimulator so does this
perform better than a macro electrode

using these micro electrode arrays so
when we pass synchronous pulses in we

can in train the whole the or rather
this is what the synchronous pulse

pulses look like look at different types
of pulses and this is the data and it

shows that when we stimulate using a
synchronous like Steve did okay not

synchronized across the whole population
a continuous stimulation or intermittent

stimulation in the theta range we were
able to suppress seizures during the

stimulation period as compared to pre
before and after stimulation and if we

look at various other synchronous
populations and different approaches we

were unable to to get that same result
so that supports the idea that that this

distributed multi electrode asynchronous
stimulation can actually suppress

seizures I didn't show you one of the
control conditions before was with a

macro electrode we couldn't get the same
effects with the macro lecture all right

so now we're not going from the bench to
the bedside we're going from the rat to

the bedside and can we go right from a
rat to a human that's also a very

difficult jump to do especially because
we're not only trying to show in a much

larger species that it works but that
type of array that we use in a rat using

a neuro nexus probe with with 15 spikes
putting that into the hippocampus deep

in the brain is not something you can do
so we really need to translate both the

finding as well as the technology and do
it in a fashion that is amenable to a

large primate and so the the pathway for
us is through the non-human primate and

I know that last week dr. diver nias was
here and she talked about some of the

non-human primate work that we're doing
and she's one of the principal

investigators on this with Babak
Mammootty a biomedical engineer working

with us and Narol
just and the PhD in my lab and this is

the scheme behind what's called Aug 3 u
h3 grant from the NIH from the brain

initiative that was given to us a couple
of years ago to translate this work so

what is ug 3 u h3 stand for well the
first question that I asked when I got

on the phone with this with the program
officer after we were told of our score

and we wanted to talk about the
milestones for this grant I said to them

the first question I have for you is
what is ug 3 u h3 stand for and the

answer was we have no idea so and what's
really you know I was especially curious

about it because we applied for this
twice the first time we applied for it

we it was called a u h2 u h3 and then
they felt the need to change it to ug 3

u h3 without even knowing what it stood
for so the NIH works in mysterious ways

but the ug 3 refers to what's called the
preclinical part part and the u h3 is

the clinical part so maybe the H stands
for humans and then the G stands for the

nonhumans I'm not you know so good
animals something like that so so this

is our scheme for getting this to human
so the second part of the grant is to go

into humans with a device and the first
part of the grant is to translate that

into non-human primate model so we start
here's the rat these are the studies

we've done already and first we have to
have a data acquisition system that's

appropriate and instead of challenging
ourselves with having in the

translatable data acquisition system to
begin with we're going to use just a

typical Blackrock and we started this
work about a year or so ago and develop

the type of electrodes that we need for
the humans and then ultimately go into a

system called an RC plus s which is a
type of deep brain stimulator system

component of this is to develop our own
surrogate so I told you that spikes per

second array wide was a surrogate that
allowed dr. Potter to tune his device so

that it could achieve the clinical
result of getting rid of the burst

firing
so we have the same problem call this

the burst frying over here and we're
looking for a biomarker of response like

spikes per second or a wide it takes
weeks and years to evaluate the clinical

response of these types of things
whereas using a biomarker allows us to

figure it out in anywhere from
microseconds to at least months and so

one of the components of the grant is to
look for biomarkers of the response this

is just some of the the the
parameterization of how we're going to

do it this shows the development of the
electrodes the type of electrodes we're

developing these are our equivalent of
of these neuro Nexus arrays for putting

into the non-human primate and
ultimately to the to the human this

shows the complexity of doing these type
of things on an animal skull this shows

the seizures that Ana L showed you last
last week and then ultimately this gets

boiled into them into human electrodes
that look something like this and this

is our RC plus s so this is not that
dissimilar from that our responsive

neurostimulator in that it records
neural signal and stimulates but is

capable of higher degrees of analysis of
of spectral doing spectral analysis and

and being able to look at biomarkers and
so forth anyway that's for the future

but another side of the lab we've been
asking the question now can we have a

more mechanism and neuron specific
approach so electrical stimulation we

haven't talked about the mechanism at
all we don't really know how electrical

stimulation really helps epilepsy or
even Parkinson's disease in fact we as a

field have have been like a sine wave
going back and forth between thinking

this is inhibiting neurons and it's
activating neurons

maybe it's activating and inhibiting
neurons so we don't really know the

mechanism of action of these things so
so we're really interested in trying to

figure out a more mechanistic approach
and instead of just sticking electrical

wires into the brain and hoping that the
right neural population has the right

thing done to it we are working towards
more specificity so so we're using life

for this and I know that many of you are
familiar with the approaches of

optogenetics and Garrett Stanley is here
and he of course his lab as you as you

know where I works very hard in this
area and so so these optogenetic

channels just for those of you that
don't know there's they come in both

activating and inhibitory forms where we
can actually introduce sodium ions or

chloride ions or protons and thereby
specifically activate neurons using

channelrhodopsin or inhibit neurons
using a halorhodopsin

or Mac so there's a lot of different
flavors of these things I don't want to

spend too much time going into the
details for Translational approaches as

we are always focused on we're not very
interested in transgenic animals so so

you know the likelihood of making a
transgenic human is is very low

especially since epilepsy usually
presents later in life the ship has

already sailed on that so so we use
viral vectors to introduce optogenetics

into animals and then you can put light
light guides in light fibers and and

achieve behavioral responses so we
started actually Steve Potter and and I

started a number of years ago to to
bring this into the epilepsy model

system this shows you one example of one
person who beat us to the punch by a

long margin so Esther cook Magnuson
showed a nice responsive system this

on-demand optogenetic system so just
like that responsive neurostimulator you

can look at the EEG menhir and decide
when there's a seizure and then turn on

an optogenetic LED in animals that have
been transfected she used halorhodopsin

to inhibit excitatory neurons and
channelrhodopsins drive inhibitory

neurons and you can see that when the
light is on

after a detection the animal doesn't
have a seizure and when the light is not

turned on the
seizure light on no seizure light off

seizure and this just shows
amplification of this and a clinical

response so so this actually does work
and in fact there have been a large

number of papers now published on using
light in optogenetic form in various

instantiations for for accomplishing
this in our lab we're concerned with

putting large amounts of light and light
fibers into the hippocampus so for

patients whose hippocampus is doing
something putting in a single fiber is

insufficient and putting in multiple
fibers is potentially damaging and so

we've been looking at targets that are
much smaller and that are easier to

control
so the medial septal nucleus is one such

area that projects its fibers to the
hippocampus and is actually the

pacemaker of the hippocampus and it has
various different populations of neurons

gaba cholinergic glutamatergic and is
the type of area that if you put an

electrical micro electrode in or macro
electrode in here you have no control

over what you're doing to these sub
population of neurons so it lends itself

very nicely to to being able to do cell
type-specific control of jabba

glutamatergic and cholinergic neurons
and so this shows an a viral infection

of the medial septum and then nice
expression along the septal hippocampal

pathway all the way into the hippocampus
and this just shows our approach to this

of turning them on turning them off and
this is work that's ongoing in the lab

sang Park who's a who's a bioengineering
student in the lab so hopefully we'll

have results to show you on that in the
coming months and years now the

translational problem ultimately is a
big one for me so that dr. genetics

works in mice in particular and even
rats is fine but how do you get opto

genetics into the human are we going to
be in the in the near future putting

light guides into patients eyes
you putting a fiberoptic into a

patient's brain for destroying the
hippocampus okay but are we going to be

able to distribute enough light without
causing damage into the poke APIs to be

able to get optogenetic control the
problem is really depicted here by the

fact that at this is the amount of
spread that you get in a rat brain and

the size difference is is substantial
between that and a non-human primate not

even to mention into a human so so this
is the problem at the light of the

spread of light and these are some of
the solutions that are being promised

and being developed and and but we
decided to take a different approach to

get around the physical limitations of
external light sources in the brain so

this is our answer to that that problem
so light is produced by many other

things aside from light emitting diodes
for example this sea of bioluminescent

organisms provides copious amounts of
light and they provided at various

different wavelengths so with Ken
Berglund who's an assistant professor in

the lab we've been developing
bioluminescence coupled to optogenetic

channels so in the presence of substrate
called seal and Terezin or basically

luciferase light is generated by an
enzymatic reaction and then can be

coupled ultimately to these two
halorhodopsin channel adopts and what

have you so this is what it looks like
in a depiction this is channel adopts

in' and this is a it's yfp and this is
Galu sephorus which is introduced as a

fusion protein to the channel adoption
channel and then in the presence of Sela

terezin makes blue light and can
self-regulate its own channel you can

also switch the channel adopts into just
to point out that we can also as a check

use standard opto genetic approaches
because this does not invalidate the the

approach of introducing external light
so we can compare the two and then we

can switch the channel adoption to
halorhodopsin and make a separate

different type of light and you connect
do this in the same cells if you want so

this is Manila luciferase which
generates with its unique substrate

yellow light to activate halorhodopsin
so what this looks like is this is a

fluid static fluorescence image and then
in in cultures let me submit can get

this mouse to come over here you can see
that in the presence of substrate we get

nice pile in essence over that whole
network so again these are the two

different forms we've called the
inhibitory form of this inhibitory lumen

opsin and the excitatory excitatory
lumen opsin and this just shows the

experimental implementation of this and
again we can use these these opcodes

micro electrodes coupled to opcodes as a
as a means of using the conventional

approach and we can put in multi
electrode arrays like we were doing

before
with the electrical approach but here we

simply use it for recording within the
CA 3 and the CA one area we've been able

to show this works in patch-clamp
experiments so you see the fluorescence

and the luminescence and you can see
Sealand Terezin and Janet generates a

nice spike and a nice inward current
this we ultimately are able to show that

this can actually control rotational
behavior these videos are inactive but

you could see them rotating if you
wanted to and we've quantified this and

this shows that with Sealand Terezin we
can actually cause EPSA lateral rotation

when put into the basal ganglia into the
substantial nigra pars reticulata and

that nice thing about this is that you
can actually see the luminescence and

the animals themselves so you can see
the control signals being generated as

we're doing these experiments we are
applying this in various different

approaches we've recently applied this
in Parkinson's disease to show chronic

activation of dopaminergic neurons by
lumen opsin and we're actually looking

this in this in a neural restorative way
where by by increasing the activity

axons we can actually get BDNF
neurotrophic factor secretion and have

neuroprotective effects on these animals
I'll show that in a second with Ling way

and chopping you I'm sorry and uh

chanting chanting we're doing this in a
stroke model looking at transplantation

of engineered stem cells combined with
chronic excitation and James Zhang is a

graduate student in the lab that
recently defended showing this and and

miles McCreary is also doing that
presently in in Champaign in Ling ways

lab and we're doing this in my lab with
epilepsy so this shows the the rat

experiments for Parkinson's disease
protection and we can see that with

Sealand Terezin were actually able to
get a nice clinical response in these in

these Parkinson model animals this is
the work of Jack Tong who got his PhD

and in biomedical engineering here a
couple of years ago and is at Stanford

right now and what he was able to show
in cultures these are these multi

electrode arrays that I showed you that
Steve Potter that really started the

whole thing these are the very same
cultures on micro electrodes and you can

see in the ilm oh the inhibitory lumen
opsin expressing cultures the firing

rate is decreased as compared to control
sister cultures and quantified over here

we then translated this into the awake
behaving rat and again you can see the

same thing that in the presence of
Sealand Terezin we get decreased firing

in in the hippocampus this is a bike you
killing model I'm not that dissimilar to

the tenth to the tetanus toxin model and
you can see that we're able to get

decreased activity decreased seizure
activity again you see here is with a

little bit of delay this isn't the
vehicle and this is in the zoo and

Terezin infused animals you see a
decrease in the number of discharges in

the hippocampus this shows then decrease
them seizure activity this they're

animals that
that where the lumen opsin inhibitory

lumen opsin is infused into the
hippocampus or the dentate gyrus and we

can see decreases in in seizure act
seizure duration what's interesting here

is that the seizure duration decreased
but the actual intensity of the seizures

didn't decrease when we just infuse this
into the dentate gyrus the nice thing

about the aluminum s'en approach is that
we can infuse this in multiple brain

regions we're not limited by all the
areas that we can put a light guide or

fiber optic into any place that we can
get gene expression by transduction we

can we can we can target for epilepsy so
we've been exploring now the idea of

multi-site neural modulation and so here
he put this into the dentate gyrus but

also into the anterior nucleus of the
thalamus which is part of the circuit

that involves the hippocampus in the so
called circuit of PES and in that case

we were able to see decrease in duration
as well as in can't even see that on

here but the well in this particular
this is the anterior nucleus alone but

took out the slide sorry about that this
is the anterior nucleus alone in the

anterior nucleus alone we got decrease
in duration and we didn't get a change

in the Racine score in the intensity but
if I left the slide in you could see

when we coupled this together we finally
got decrease in seizure duration and in

the Racine score as well so these things
seem to be additive so lumen options for

epilepsy are nice because they allow
more widespread neuromodulation than

conventional optogenetics there's a gain
in the distribution but one of the

things we lose is temporal control so
instead of being able to flip the light

on and off like cook Magnuson was able
to do we are to rely on drug delivery

but what we're looking to develop now is
there's been some really neat

microfluidic devices for for very tight
drug delivery and control drug delivery

so we're thinking about beginning to try
to apply this work where we can have

great control of the drug delivery and
then we've solved both problems we've

got both temporal control as well as a
better distribution so

we need to solve the substrate delivery
hurdles there are certainly pros and

cons of the systems and of course some
of you may have heard of designer drugs

the designer receptors exclusively
activated by by designer drugs it's

another chemo genetic approach like the
luminescent approach which has certain

advantages and disadvantage it's
compared to lumen options and we're

working on autonomous responsive lumen
options whereas where we could provide

the substrate continuously but not have
the lumen option be activated but

building in a responsiveness into the
bioluminescent part of the molecule for

example putting calcium calmodulin and
conferring response to calcium levels

within within the neurons to make an
autonomous response of lumen opsin and

this hopefully will lead eventually to a
completely non evasive approach using

perhaps vascular delivery of the of the
viral vectors vascular delivery put in a

drug pump at least there's something for
me to do as a neurosurgeon and and and

hopefully get more widespread adoption
so to summarize I'm going to summarize

this by saying what all of this is
basically exemplifies is unmet clinical

needs which I and my in my physician
colleagues can identify treating

epilepsy is good but not great we can
get to 65% we don't often get to a

hundred percent seizure reduction
patients want to be seizure free off of

taking drugs so we identify unmet
clinical needs such as these and we

marry this to clinician scientists and
engineers working together to make new

therapeutic options these are our
targets I've told you about some of the

tools that we have and all of this is
basically what entices mission is on the

emory side and then we married together
with the neural engineering center Chris

and Garrett and Lina ting and Rob Butera
and others here on the Georgia Tech side

we really are one center we formed as
two separate centers because of the

administrative aspects of working in
under two different universities but we

one Center really working together and I
think somewhat seamlessly towards all of

these advancements and that really is
the is the focus of the group together

so just to acknowledge the people that
are involved I have to involve

acknowledge John Willie who's a junior
partner of mine who's been doing this

with me
especially the laser stuff and in

particular for the last us
I guess it's seven years Matt who I

already acknowledged some of the other
people work in our group the Emily Emory

neurologists and our and Babak and Anna
Elle working on the monkey experiments

and some of the previous students have
already already named and of course our

funding sources have to be gratefully
acknowledged so with that I will

conclude and if anybody has time to
stick around happy to take your

questions

questions yes Annabelle yeah I mean well
first of all this the substrate now the

substrate sealant Terezin is a foreign
substrate so it actually as opposed to

dreads does not actually have any other
targets to drag the metabolize or NPOs

metabolize said they're wrong as CNO
it's metabolized to clozapine so sealant

Terezin might have some toxic effects
actually in japan they sell sealant

tears they sell bioluminescent lollipops
so so we don't we're not worried too

worried about that but yeah the idea is
not to chronically activate these

systems I mean we right now we're using
the substrate and it has to be substrate

dependent we don't know what the effects
of long term activation are with

electrical stimulation though there is
some data on this because while the

neuro pace approach uses responsive
stimulation there are other approaches

using continuous stimulation of the
hippocampus and remarkably it has pretty

minimal long-term effects we've been
doing electrical stimulation for you

know decades and similar patients so I
think there's the possibility of

habituation I found that a company
called neuro Vista with some colleagues

and the idea was that was that we don't
want to have these things be active

long-term because the nervous system
would habituate and so we have to make

these responsive but we don't really
have any evidence to actually support

the adverse effect of long-term
continuous stimulation in most settings

so we'll have to see

right right so yeah I mean this has been
looked at looking at various markers of

microglial activation and so forth and
in fact you know anytime we put

something in the nervous system there is
an inflammatory reaction around it it it

and and gliosis that occurs as well in a
more long-term fashion so these are

important as far as the inflammatory
effects of neural modulation per se we

and others look at this and we we don't
see very much with respect to that in

fact there are certain scenarios where
you can actually see an

anti-inflammatory effect of neural
modulation

I think perhaps the reason for it is
because we are working in the context of

how the nervous system works ordinarily
you know we're talking its language

certainly if you do things that are non
physiological that you know we certainly

get into damage zones and inflammatory
responses there but as long as we stay

within the physiological range and just
feed those physiological signals in in

you know a non physiological way I think
we're we're usually pretty safe with

respect to inflammatory effects