Brainiacs: Motor Recovery After Stroke in Mice and Men

, , Leave a comment


I know it’s been a long day,
so I’m gonna say snappy, controversial, provocative
things.>>[LAUGH]
>>Just to quickly get back to the riding a bike question. I’m not gonna have time
to talk about this, but I will say that I’ve
written about it. We’ve even had
an article in the Times. This distinction
between the motor and the cognitive is gonna turn out
in my view to be largely false. To get to what was
said by David. He’s absolutely correct about
what may be the reason for this, how difficult it
is to ride a bike. It’s also very similar to going
back to a country where you haven’t spoken French,
for example, and you start to practice, and
it comes back in chunks. And there is something called,
you can look this up, called free-lunch learning, and as David, with he knew
about free-lunch learning. Was that it’s true that neural
networks can have this property of associative learning,
where you actually inherent to the architecture
of these networks, will get back learning for
free when you learn a piece. So in other words, you learn
what table is again in German and chair comes back for
free, for example. And it’s absolutely not true
that somebody who hasn’t ridden a bike, when they do
it again after a year, isn’t a bit shaky and bit off,
just that they retrieve fast, which is similar to vocabulary. So it’s called free-lunch
learning and you can look it up. And the second point is this
distinction between the motor and the cognitive is mainly
a cultural bias rather than actually something
that neuroscience supports. The cerebellum has
unique properties, but it’s only one piece of motor
learning, and you wouldn’t be able to learn how to ride a bike
if you only had a cerebellum. Okay, we’re gonna talk about, Stroke, like Amy, just to put
stroke into perspective, and if you guys want controversial
things to write about, start writing now, because
a stroke is, in fact, far and away the most devastating
neurological illness. I think Jeff Rothgene mentioned
that 1 in 500 people will die of ALS in the US, 1 in 18
will die of stroke in the US, 1 in 6 women over the age
of 65 will have a stroke, 1 in 5 men over the age
of 65 will have a stroke. About 800,000 people
a year have strokes. There’s somebody having a stroke
in this country every 40 seconds, there’s somebody dying
of a stroke every 4 minutes. So it blows the other diseases
out of the water in terms of its impact, and I’m saying that
not to have a competition, but because stroke is
woefully underfunded. And what Amy was talking about,
and what she’s doing, and what other
people are doing in rehab, it’s time for
a revolution in rehab. Because those of you who do
go on to have strokes, and I’m afraid this doesn’t look
good with a number in here.>>[LAUGH]
>>You’re gonna be somewhat shocked by how medieval
rehabilitation is and how desperately in need
of new treatments we are, like the treadmill treatment
that Amy showed you. And just so you don’t think
it’s a disease of old people, in 2009, 34% of the strokes that were hospitalized were
under the age of 60. So it’s not a disease of old
people unless you want to come up with a new definition
in our immature age of old. This is our website,
and you can go to get the papers I’m gonna be
talking about from there. The only reason I show this is because I want you to
follow us on Twitter. We’re having a competition
with other labs, and we’re desperately behind. Follow us.>>[LAUGH]
>>All right, so recovery. I think this is very important. It can mean two things. One is we want, and this is what
we’re gonna be talking about mostly today, is we’d like
people return to doing things the way they were
doing them before. Or you can compensate. Now, in the motor domain,
compensation and true recovery are usually
meant in the behavioral realm. What RG was showing, was the
sort of interesting possibility that it’s more difficult in the
language domain to distinguish neural compensation
from neural recovery. And it’s probably because
there’s degeneracy and redundancy in how we construct
language that is more complex than movements, at least in
terms of how we can measure it. So what I’m going to be talking
about here is not neural or compensation versus recovery. I’m gonna talk about, does
the behavior look phenotypically the same as before or
close to it, or are you just finding
a new way to do it? For example, you’re
right-handed, you have a stroke, and you learn how to use
your left arm instead. So I’m gonna be talking to you,
again, another talking point about
mouse experiments and human experiments and
I think Jeff brought this up. The idea that you do fundamental
science in animals, and then you just apply it in humans
has always has been nonsense, but it’s particularly
nonsense now. And the only way that we’re
gonna move forward is by having two-way traffic between people
working in non-human animals and animals. As Amy showed in that really
good slide at the beginning, the tools that you use to do
science in humans are different from the tools that you use
to some degree in animals. But the way of thinking,
the rigor, the concepts overlap
tremendously. And in fact, the experiments I’m
gonna show you today done under my supervision by
an assistant professor, the mouse experiments, I designed them with with
him from our human results. So in other words, what you’re
going to see happening is two-way traffic, and people
working on non-human systems and human systems together rather
than this totally passé distinction between clinical and
basic research. This is just to justify why
we’re allowed to look at the rodent and
the human just in terms of this, I’m sorry it’s not
the greatest slide, but it’s amazing how much dexterous
prehension rodents have. Those of you who’ve lived in New
York as long as I did will see just how remarkable rats
are when you walk into the house unexpected, and
they’re on your computer.>>[LAUGH]
>>All right, so mouse. So the take-home
point here is that if you have a stroke in the
United States in 2014, you’re better off if you’re a rodent
than if you’re a human being. And I’m being serious. If you give strokes to rodents,
you then put them in these really fun environment, so
they have friends to play with, they have wheels to run on,
it’s an enriched environment. They do really well. And it’s a very
interesting thing, and it’s a distinction that’s very
important to have between the notion of motor learning and
training which have been, in my view, overemphasized for
rehabilitation. And the reason for that is because learning
is the way to compensate versus repair and plasticity,
they’re not synonyms. Repair and learning, where there
are these endogenous processes that can actually get you
back your true behavior. Is it possible that
having enjoyment and being allowed to do whatever you
want may be better for repair, at least early on, versus
picking a particular task and pounding people at it? And what we know, and this is
what’s very interesting is in the rodent literature,
it seems as though for a brief period after a stroke,
you recapitulate the super plasticity you had
during development. So you have a little
recapitulation of developmental plasticity for
a period after a stroke. It’s very interesting. And there are overlapping
growth factors, genes expressed,
levels of inhibition, and excitation that seem to
recapitulate ontogeny. And people have
talked about this. Does ischemia give you a brief
recapitulation of ontogeny? Okay, so you get a switch on. And then unfortunately,
you get a switch off. And during this period
between the switch on and switch off exemplar
growth factors, you seem to be particularly
responsive to training effects. And we need to exploit
that in human beings. At the moment,
people spend 60% of their time alone in hospital
rooms after stroke. They make homeopathic
amounts of movements, which we know don’t induce
cortical plasticity. And so unfortunately, despite
what the science is showing in rodents at the current time,
again, get your pens out, that’s not being translated
to the human world. Rehabilitation in the United
States is driven entirely by economics, not by science. This is Steve Zeiler,
he has a K award, these are the experiments
I’m going to show you. This is just to
show you a mouse, look at those little digits,
how cute that is, that’s the first time
it’s ever done it. But you can see, I bet you’re surprised that they
have that degree of dexterity. Now the way that Steve
set this up was so that they can’t grab
with their mouth. Rodents don’t really look at
things, they’re olfactory, and they would rather nuzzle
up against things. So we have to prevent them
by giving them a slit, and things like that. This is early on, and basically, this is the behavior that
you can train them on. The reason we’re using this
behavior is because it’s more analogous to human prehension, people don’t usually
eat with their mouths. I mean,
grab things with their mouths. And this is what
we’re gonna show you. So we can give a mouse a stroke. This is just a cross section, a little bit like the images
that Artie showed you, but this is histological,
and this is the two sides. And you can give them a stroke. In the equivalent
of motor cortex, it doesn’t do the full
thickness of cortex. You can train the mouse before
the stroke on the task that I showed you. You then give it a stroke
like I just showed you, and there’s a plummeting in its
behavior, and you’d basically, in this week just
put it in its cage. Alone. If you then start
to rehabilitate it, nowmatter how long you train it, you can never get it back
to the prestroke level. This it to show it’s
not just forgetting. Because if you take a mouse and
you just don’t give it a stroke, you train it and then you just have it not
train on that task for a week, and then you have it come back,
it remembers what it had. So it’s not as though this is
just a forgetting phenomena. However, if you give it a stroke
and then you wait 24 hours, and then you start retraining it, you get it right back
to where it was before. So in other works, in the mouse,
it’s quite dramatic. That if you don’t dilly dally
and you start the rehabilitation early, then you can actually get
quite a dramatic restoration of behavior which is already on the
way and then you wait a week. So this window of opportunity,
this brief recapitulation of ontogeny is already decaying
after a week in the mouse. So rehabilitation really is
taking this sensitive period and wanting to start training early
to get you up rather than waiting and doing that. And what I’m saying is human
rehabilitation is a combination of delay and too low a dose of the treatment
even if you were to go early. All right?
So, it’s dose times time. Okay, so enough about the mouse. Let’s quickly go into the human. And this is a study that
has been going on now for the last three years and
is gonna start a trial component this summer, involving Columbia,
Hopkins, and Zurich. So a very large team. I’ll show you some
of the players. Just like everyone
else has said. And this cannot be done without
a lot of wonderful people. What I’m gonna show you
here is the Columbia wing data of this study. And just to show you that there
are analogs to what I just showed you in
the mouse in the human. Okay? So instead of making people
reach through slips of pellets, you can make people make
movements on what looks like a hockey sled. There are reasons for this task. There’s a task which if you
constrain the trunk and you have the elbow and shoulder at the
same level you can’t compensate. In other words, the only way you
can do it the way we want you to do it is if you get your motor
control back, not for example by moving your trunk or finding
some other way to do it, okay. So it’s sort of a compensation
proof test of motor control. And if all of you
would have a go at it. So, in other words you just have
targets arrayed right around like a clock. You just sit there with your
elbows and shoulders same level, as on sledge and
there’s no friction and you make these movements
to the targets, and you would all
look like this, okay. This is what
a patient with a mild to moderate stroke looks like. This is what a patient with a
severe stroke looks like, okay. In other words they cant
compensate to look good either way because we
make it impossible so their motor control
deficit is revealed. Now, you can take those
trajectories, and I’m not gonna go into detail,
but you can do something called hierarchical PCA to actually
quantify those trajectories. And you can divide them into
a mean and a variance and what this study
shows unfortunately rather like the mouse data is
that control, not compensation, improvement is over by
four weeks in humans. So this plasticity that is there is used up at four to
five weeks in humans. All right, so you have this
critical window that I’m arguing is frissured away by current
forms of revertation. It’s interestingly, I won’t show
you the data today for the hand, it seems like we have
more time to play with. For the arm,
it’s depressing actually. So we’ve got to do
something dramatic and it’s not a floor effect because these
are controls that technique has all the sensitivity it needs to
detect improvement from this and it’s not happening. Why is this happening at all, what is that improvement
that’s going on? In the first four weeks where I
would argue and we have argued that its not related to
rehabilitation itself. Its an endogenous process that
we need to find a way to jump on top of that we’re not. And I don’t have time
to go through this but we’ve devised a statistical rule
to show that this recovery is fairly predictable based on your
initial level of impairment. And there’s also a neural
network similar in derivation from what Archie was showing
where you can actually show the neural correlates of
the subsequent improvement of the statistical rule
that goes on and then stops after four to five
weeks regardless of what we do. So, what’s the model? In a healthy subject you’ve
got a normal brain and you have normal
levels of plasticity. In a chronic case, you’ve got a
damaged brain and your learning mechanisms are no different to
what a healthy subject has. Unless your lesion just happens
to be in a critical place, but that’s not the issue when
it comes to hemipareisis. And then, there’s this special
case of hypoplasticity and brain damage that we’re not sure
how long it lasts at least for the arm that’s closing more
quickly than we were hoping. So what to do? Well, we need to take
the sensitive period and we need to prolong it,
enhance it. We need to then exploit it. In other words,
even if there is a window, we have to do something
in that window, right? What do we do and one thing that lot of people do
is task specific training and other one is to give them
an enriched environment. So what’s equivalent
to making it fun and colorful and
non-boring, all right. So as Archie mentioned,
Fluoxetine is all the rage and I don’t have time and that’s
my way of saying cuz I don’t know exactly what fluoxetine
is doing, there are people in this room far better qualified
than I am to say, but people have seen changes in LTP,
increased expressions of BDNF. It certainly does seem as though
there are effects of fluoxetine and other SSRIs that are not
related to it, [INAUDIBLE] fact. Okay, and the important thing,
this was a paper important paper showing that it was restoring
plasticity in the visual cortex and then there was this trial
it was done showing improvement in patients if that was given to
them for the first three months and interestingly the average
time of onset of giving the 20 milligrams of Prozac in
this trial was nine days. So, going back to the mouse. This is what I’m talking about, this is his dance, right,
between the mouse and the human. Flame was done,
let’s have a look in the mouse. Steve then basically
did what I said before, you train two groups of
mice on this reaching task. You then start fluctuating
within a day of their stroke. They plummet and then you
start training them, and really very pleasingly
in the animals that get the fluoxetine you can actually
get them back to normal levels. Having delayed a week which
was not the case before. So other words perhaps we
can actually extend or preserve this heightened
plasticity for longer. Which will be good because
sometimes patients are very ill and they can’t start
doing intense treatment as quickly as we’d like. And this is the result of the
flame study also showing a sort of improvement piggybacking
on the spontaneous recovery. I’m talking about in
the humans who got fluoxetine. So what about after the window? I just wanna get to chronic
because I want you to understand that we’ve got to do
something early because there’s a certain futility
in once you go late. With the motor system. So this what again before I
moved from Columbia to Hopkins. And I’m just gonna quickly
tell you the punchlines. We did two studies,
the two most exciting forms of rehabilitation that you
probably heard about. A robotics and
constraint induced therapy. Constraint induced therapy is
when you sort of put the healthy arm in a sling or in a mitten so you’re forced to
use your unaffected side, okay? Punchline, in
the robotics study, we found feeble,
improvements in motor control in response to the robotic
therapy for two weeks. Okay, so in other words, unlike
the spontaneous recovery that’s quite dramatic early
which isn’t due to us. If you try and force it yourself
after the window is closed you get no mileage. Similarly no improvement
at all in motor control. After constraint
induced therapy. So the two most advertised and talked about forms of therapy
when it comes to motor control can’t do what your brain
is doing on its own early. You can compensate, I’m not
dissing physical therapist, occupational therapist, but what they’re doing is
teaching you this. Instead of doing the high jump
like this, you do it like this. This is a change in strategy. This is the Fosbury flop that
changed the world record in the high jump. Doing the same task in a totally
different way with what you have left is perfectly valid but
it’s not repairing the brain. Okay? So what to do? Well, enough criticism for
me to actually do something. So, SMARTS II which is funded,
again, by the group that did the natural history and this is,
some of the Hopkins people that I’m going to talk about is
a trial that we’re hoping to do. So you want to again, write
about something exciting I’d be happy for you to advertise it. We’re trying to do something
brand new which is to take this window, exploit it with
a combination of robotics and video gaming. So this is the co invested
Hopkins from this cell. The fellow and the research
assistant who will be doing it. And this is the gaming
team in my lab, who have created something which
I’ll show you in a minute. And what we’re trying
to do is enrich. Okay? So this is a standard environment. This is the resident
artist in my lab, Kat. I’ll just show you her, Kat. It’s good to have again,
Humanities. Very important. You need artists in labs. I’m serious. I don’t think there’s another
lab at Hopkins that has as full time artist in it. Terrible sad. All right. Standard environment. Alone most of the time watching
dreadful reality TV programs. This is what you want. You want a robot in the room. You want to be immersed
in virtual reality. And you may even make me want
to have your brain simulated which is where when
he designed the study with me he’s an expert
in brain simulation. We’ll be doing this plus imagine
a battery stuck on your head. Here’s the robot. All right. This the study. And I’ll finish by just showing
you what the patients have already started playing with. Can we turn the lights down
a minute just to see this, is it possible or
do I have control of that? Why isn’t this, oops I’m sorry. [MUSIC] So this is a person
actually controlling your development, okay? And you basically don’t even
get any task instructions, all right? Other than eat as
many fish as you can. [MUSIC] And then when you get fish
fueled you can them jump sort of better and better. This is with the robot attached. Okay? So the idea here. This is being done
with 3D sound. We are thinking of doing
it 3D and patients that have done this thus far, they
don’t want to get out of it. They want to stay in the ocean
with this bizarre feeling of immersion and just they don’t
realize it but we’ve set it up do that the dolphin is moving
in the places we want you to go. And you’re just
basically enriched. And the beautiful
thing about this, is that you can play
this with a loved one. So you can both dive into
the ocean and do this. And in the ocean,
you’re equals, right? Cuz we can play with
the parameters. Right? So we have to change the clinical, dreary,
depressing environment. And we have to give
what the rats have and this is the way we’re doing
it and we’re gonna try and do this in the window where this
plasticity is still on our side. And I’ll stop there,
thank you very much.>>[APPLAUSE]

 

Leave a Reply