July 4, 2024 • 41 mins
On the next few episodes of What's Your Problem, Jacob Goldstein is talking with people working at the frontiers of technology to help elite athletes perform better.
Today’s guest is Silvia Blemker, a professor of biomedical engineering at the University of Virginia and the co-founder of Springbok Analytics.
Silvia's problem is this: How do you combine MRI scans and artificial intelligence to generate new insights that can help both elite athletes and people suffering from diseases that affect the muscles.
Springbok's sports clients include medical researchers, Olympic athletes, Major League Baseball and several professional basketball and soccer teams.
This summer, a bunch of Pushkin podcasts are coming out with Olympics-inspired shows. Revisionist History has a series about America's decision to participate in Hitler's Berlin Olympics in 1936. The Happiness Lab has an interview with a coach who coaches coaches. And Cautionary Tales tells the story of the family feud that gave us both Puma and Adidas.
See omnystudio.com/listener for privacy information.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:08):
Hushkin, what's one surprising thing your work has taught you
about elite athletes?
Speaker 2 (00:22):
I never thought I would see muscles that were so
developed they broke our scale. Wow.
Speaker 1 (00:31):
Yeah, like it was just too big the machine, the
AI couldn't figure out what it is.
Speaker 2 (00:35):
Well, no, the AI found it, but we are like
our kind of rating system.
Speaker 1 (00:40):
Wow. Was there a particular athlete or a particular sport
or particular muscle?
Speaker 2 (00:44):
What?
Speaker 1 (00:44):
What? What muscle broke the scale?
Speaker 2 (00:47):
Uh? The gluteus maximus breaks it A.
Speaker 1 (00:49):
Fair kidding fantastic.
Speaker 2 (00:53):
Yes, it's a pain in my butt.
Speaker 1 (00:56):
Like because it's too big.
Speaker 2 (00:58):
Yeah, it's just so big.
Speaker 1 (01:06):
I'm Jacob Goldstein and this is what's your Problem? This
month a bunch of Pushkin podcasts are coming out with
Olympics inspired shows. Revisionist History has a series about America's
decision to participate in Hitler's Berlin Olympics in nineteen thirty six.
The Happiness Lab has an interview with a coach who
coaches coaches and here on What's Your Problem, We're going
(01:30):
to be talking with people who are working at the
frontiers of technology to help elite athletes perform better. For example,
today my guest is Sylvia Blemker. She's a professor of
biomechanical engineering at the University of Virginia, and she's the
co founder of a company called Springbok Analytics. Sylvia's problem
(01:52):
is this, how do you combine MRI scans and artificial
intelligence to generate new insights that can help both elite
athletes and people suffering from diseases that affect the muscles.
Springbox clients include medical researchers, Olympic athletes, major League baseball,
and several professional basketball and soccer teams. You'll hear about
(02:13):
all that on the show, but first we're going to
pick up where we left off in the conversation. We
were discussing the extraordinarily large muscles of elite athletes, and
then Sylvia told me something even more surprising.
Speaker 2 (02:30):
The other thing is that they have some tiny muscles too, Huh.
Speaker 1 (02:33):
Like they have like smaller than a normal person's muscle.
Speaker 2 (02:36):
Much smaller.
Speaker 1 (02:37):
Huh.
Speaker 2 (02:38):
They put their muscle where they need it.
Speaker 1 (02:40):
What's an example, Like what muscle is tiny and what
kind of athletes?
Speaker 2 (02:44):
Calf muscles are small in most fast athletes huh? And
ether you look at a sprinter or like a running back.
Speaker 1 (02:53):
It's just all quad no calf, all.
Speaker 2 (02:57):
Like thigh no calf, yeah, THI And it kind of
makes sense because you know, if you're trying to run fast,
you wouldn't want to put a lot of mass like
at the end of your leg. It's like as a
lot of inertia to like move your leg.
Speaker 1 (03:09):
Huh.
Speaker 2 (03:09):
Because you know, the muscles are important for sprinting, that's
the interesting thing, but they just don't they're small, very.
Speaker 1 (03:16):
H huh uh huh. So I'm particularly interested at this
moment in the sports piece of what you do. I'm curious,
by the way. Do you work with any Olympic teams
or Olympic athletes? Yeah?
Speaker 2 (03:33):
Yeah, We've actually been working with several different Olympic athletes.
The ones that probably that come to mind most are
multiple players on the US women's national soccer team.
Speaker 1 (03:46):
Oh cool, tell me, like, tell me the story of
that of that work. So they came to you, what
did what did they what do they want when they
came to you? Like, how did that? How did that begin?
Speaker 2 (03:57):
They came to us along with their team. So the
technology we provide, you know, an athlete could understand it,
but really with their team to help them figure out
how to keep athletes healthy.
Speaker 1 (04:10):
So what did they what did they say? What did
they say when they came to.
Speaker 2 (04:14):
So, for example, one athlete that's coming to mind had
a known imbalance side to side that based on a
history of injury, and they really wanted to know where
that imbalance was coming from.
Speaker 1 (04:30):
So the woman had had hurt one of her legs,
and that leg was even after she came back, that
leg was weaker essentially than the other. I mean, is
that the sort of gross you know, macro.
Speaker 2 (04:41):
Way, Yeah, exactly, that's a that's a nice way to
put it.
Speaker 1 (04:43):
Yeah, And and they wanted a sort of finder like, okay,
but we can see that, but what's going on on
the inside, like muscle by muscle tell us that, yes.
Speaker 2 (04:53):
Exactly, that's precisely what we do. We go on the
inside because on the outside you see perhaps that her
knee extents or quads seem weaker on one side than
the other. But there's four quads, quadre steps, four muscles,
and so it's not clear which of those muscles are
actually the culprit for that imbalance and in what way.
Speaker 1 (05:15):
Good So this is their question, and then what happens next?
Speaker 2 (05:20):
So This first step is an MRI scan, and so
with these athletes or teams, we have ways to connect
them with an MRI machine, whether it be through an
imaging center that they partner with, or we've even actually
brought MRI mobile trucks to sites to make it.
Speaker 1 (05:45):
Like players run off the field and get an MRI
and go back and keep playing. Yeah, yeah, kind yeah, yeah.
Speaker 2 (05:51):
It helps just with the timing of things. But so
first we connect them there, so it takes about ten minutes.
Then they send those pictures up into the cloud into
our server and then we crunched through it and then
we send back a report on their muscles. We also
have what we call it interactive Viewer, and it's presented
(06:14):
in the form of a three D model. Three dimensional model,
so you actually see your own legs, the muscles and bones,
your own muscles and bones that we've identified from the
images going through a process called segmentation where we find
all the muscles and bones and then we reconstruct them,
so it's kind of like a digital twin of that
(06:35):
person that they can see on their computer. And so
along with it or a number or all these metrics
that helps them understand their balance. The development or strength
of the muscles and the health of the muscles.
Speaker 1 (06:50):
So tell me about this report, they get like, what
does it? What does it say?
Speaker 2 (06:55):
So the basis of that is actually a lot of
research that we did over many years, because you need
to understand where somebody falls relative to a normal essentially
to give them. Essentially, we have a scoring system for
the muscles and that's based on comparing with a large
data set of healthy individuals. And so we know for
(07:18):
a given person, based on their sex, age, height and weight,
how big we expect all the muscles to be. And
that's through a lot of previous research. So then we
can say, okay, here's where you land each particular muscle
compared to this nor what we call a normative database,
(07:38):
so we call it a spring box score.
Speaker 1 (07:40):
Do you do it for every muscle in the leg.
Speaker 2 (07:42):
Or we do it? So our primary product that we
started with was every muscle in the legs essentially from
belly to feet, no muscle left behind. They're all important.
Speaker 1 (07:54):
How many muscles are there.
Speaker 2 (07:56):
Thirty five per leg?
Speaker 1 (07:58):
So seven okay, okay, yeah, more than I would have guessed,
but a lot. Yeah, all important, okay. So, and so
it's basically, how strong and healthy is every one of
those seventy muscles relative to baseline?
Speaker 2 (08:16):
Right, And then the asymmetry comes where you can compare
side to side. So for each of the thirty five
muscles that they exist on each leg, we can say
which side is bigger, which side is smaller, and by
how much? And then we also have normative values for
that because we're all just slightly asymmetric.
Speaker 1 (08:34):
Uh huh, And presumably some muscles are more asymmetric than others,
And so you want to know kind of how how
asymmetric relative to baseline is this particular pair of muscles exactly? Yeah,
And and so in the in the case of this
uh soccer player who came to you who you know,
knew knew she had some kind of problem with her
(08:56):
quadriceps on one side, but didn't know what was going on.
What did you find?
Speaker 2 (09:01):
We found some imbalances, and actually not just in those muscles.
It turns out that, you know, it's all connected. So
if you have a weakness or an imbalance and one
set of muscles, usually some other set of muscles are
compensating in someone.
Speaker 1 (09:18):
Yeah, well, it's like when you like mess up. Even
if you're just a recreational athlete, right Like, if you
mess up something, you mess up your ankle, then you
start walking funny, and then like your back hurts because
you're walking funny. Right Like, that is a very anecdotally
apparent thing.
Speaker 2 (09:33):
Yeah, yeah, we all know that, but that you know
it shows through in the skin. But the thing is
that it's not very intuitive from the outside which muscles
have been affected, how they've compensated, and it looks different
for every single person.
Speaker 1 (09:46):
Huh.
Speaker 2 (09:47):
So that's why the report is very valuable because for
that person, they know exactly which muscles are the ones
that they really need to target, both the ones that
they already thought maybe were an issue, but then all
the other ones that showed up and they didn't really realize.
Speaker 1 (10:03):
And so in the case of this soccer player, was
it like one particular quadrcep on one side that was
like the core thing and you could figure out which
one it was.
Speaker 2 (10:12):
There was a few mess It wasn't just that. I
think there were at least one calf muscle and then
some in especially in the deep hip, those were impacted.
So yeah, it kind of shows up everywhere.
Speaker 1 (10:26):
And so you have this essentially diagnosis, right, a very
sort of fine grained kind of diagnosis. Do you also
have a have a prescription? Do you have sort of
particular kinds of training to address these very fine grained things,
or do you leave that to the trainers or whoever?
Speaker 2 (10:44):
We leave that to the trainers, because I think that
it's also important to have all the other information about
the athlete. We're not arguing that it replaces everything else.
And people pair it with lots of different other types
of measurements depending on the application or in the setting,
Like some people pair it with let's say, metrics of
(11:06):
jump performance. I'm shifting over to basketball here, but that's
just one that came to mind, where you can look
at the asymmetry about how how an athlete jumps, but
then you can also compare it to the asymmetry of
their muscles and get some insight. So it definitely, you know,
plugs in with a lot of other things.
Speaker 1 (11:25):
And to what extent can trainers or you know, strength
coaches develop programs that are sufficiently kind of fine grained
to match the kind of fine grained findings you're having, right, Like,
for example, if you find, as I understand you did
that a soccer player has one particular quadricep that is weak. Like,
(11:48):
are there workouts that target a single quadricep and not
the others?
Speaker 2 (11:52):
Yep, there are.
Speaker 1 (11:53):
That's cool for whichever quadricepp you're just like, just for fun,
give me an example.
Speaker 2 (11:59):
You know. One one way that it's very simple is
using something called biofeedback. Huh. So you can measure whether
you use something called EMG, which is a way to
measure how much electrical activity is a muscle, and then
you can see which muscles you're using for a given task.
So if you give people the feedback of which of
(12:20):
those muscles they're using and say, oh, no, you're not
using this one, use this one more, that actually works
very effectively.
Speaker 1 (12:27):
Oh really, So you can basically use your brain if
you're getting the feedback to focus on which quadricep you're training.
Speaker 2 (12:37):
Yeah, and there's other ways you can give the feedback
in other different ways, but yeah, our brains are very
good at that. Once they get feedback, they're very good
at learning.
Speaker 1 (12:46):
That's cool, especially somehow to think of what elead athletes right,
because they are already presumably like super dialed in in
terms of like the relationship between their brain and their
body at this very elite level exactly.
Speaker 2 (12:57):
Yeah. The other I was going to mention a lot
of players and teams use this not just one time,
but overtime. So they'll get a scan, figure out a plan,
work on that for maybe three months or six months,
and then do another scan and see how things are
progressing and adjust accordingly. So that's definitely another way to
(13:21):
in the long term see if what they're doing is
resulting in the change that they're hoping to see.
Speaker 1 (13:27):
So, what happened with that soccer player who had the
weak quadricep and other related Yeah, No, I.
Speaker 2 (13:34):
Think she's doing great, like staying healthy and and you know,
getting ready.
Speaker 1 (13:41):
Yeah. So I know you can't tell us her name,
but will we see her in the Olympics this time? Great? So,
as you were talking about that, I mean there was
a moment where it was like, Okay, the athlete goes
and gets the MRI and then you get the scan.
You get the scan, and then you said, like, you
crunch through the numbers and then you make the report.
Ssumably you crunching through the numbers is like the result
(14:04):
of many many years of work and kind of the
core of what your company does. So I want to
talk a little bit more about that and kind of
how you how you got here, Like how did you
come to start the company?
Speaker 2 (14:16):
Mm hmm? How long do I have a while?
Speaker 1 (14:22):
I mean it's not you know, it's not the radio,
it's a podcast.
Speaker 2 (14:25):
You. I am a professor too, so I can go on.
Speaker 1 (14:27):
Let me ask you this, What was the moment when
you decided to start the company?
Speaker 2 (14:32):
I can give you one moment and then we can
do a.
Speaker 1 (14:36):
Couple of moments. Yeah, that's what's what a good story is,
like three moments.
Speaker 2 (14:43):
In time, right right? Three not four?
Speaker 1 (14:46):
Four is a tricky number. We could do five, or
we could do one. I think I think is better.
Speaker 2 (14:52):
Yeah. Yeah. So I'm a professor. I run a lab
and form my entire career. I've been fascinated with must
and how it works, and fascinated by something that we
all in like the muscle field. We call form function relationships.
So the idea that the way a muscle is shaped
(15:13):
and the way it's structured and how big it is
influences how well it works or how well it functions,
how strong it is, how well it behaves. And there's
a lot too there, and there's a lot of nuance.
And that's like I've spent a career studying that in
lots of different ways. So I've always been interested in
quantifying muscle and figuring out how that influence is, how
it works, and both in healthy people or in athletes,
(15:36):
and also in different patient populations different In particular, I
have an interest in movement disorders, so neuromuscular diseases that
lead to impairments and mobility and movement ability. So one
of the light bulb moments for this was the fact
that I'd been using MRI to study muscle in my
(15:59):
research for a long time. It's kind of a ubiquitous
tool or like often use tool and research. But I
was struck by the fact that I was hearing from,
in particular a surgeon collaborator, and the surgeon was telling
me about his work and helping kids with cerebral palsy
(16:22):
improve their movement where they had hindered movement and largely
because their muscles are impacted. Not only do they have
an impaired ability to control their muscles, but their muscles
end up with impairments in their structure or their form,
which then influences how well they work. So surgeons have
(16:42):
to go in and do surgeries to try to change that.
They do things like modify tendons to try to make
muscles less stiff, or they transfer muscles to make them
do a new thing. But one of the big tricky
parts is that oftentimes some of those muscles are very week.
So if they choose the wrong muscle, then they'll make
(17:03):
a weak muscle even weaker. Huh, And that's catastrophic. So
it's a very fine line that a surgeon has to
figure out and they have to go in. You know,
we just talked about there's thirty five muscle muscles in
each leg, So which of the muscles are the ones
that should be operated on and which ones should be avoided?
And so what my collaborator, a guy named doctor Mark Abel,
(17:26):
fantastic surgeon. He was telling me, Yeah, like, it's very hard,
and I don't he didn't have a way to see that.
All that he could do is look from the outside.
No technology could give him the information he needed to
figure out which muscles he should focus on and which
ones to.
Speaker 1 (17:41):
Avoid, because it's not obvious by looking what's a strong
muscle and what's a weak one. Yeah, I guess that's
surprising to me on some level, Like I don't I've
never thought about it. But naively, I would think you
could look at a muscle and say it looks strong
or it looks weak.
Speaker 2 (17:59):
Not right, because you just see it from the surface.
You don't see it on the inside. And the other challenges.
For every joint, there's many muscles. So like we just
said that quadrceps has four muscles on the back of
the leg, ham strings, there's three hamstrings muscles. There's other
muscles that are in the thigh. So you're just seeing
(18:21):
what's an impairment and the overall movement, let's say of
a joint. But then there could be many muscles or
combinations of muscles that are leading to that, and you
don't know when you look from the outside. Our body
is designed that way actually to be somewhat we call
it redundant. We have more muscles than we need probably,
but if you think about imbalances, then any one of
(18:41):
those muscles could create it create some trouble.
Speaker 1 (18:44):
So, okay, so the surgeon describes this problem he's having,
and then and then what do you do?
Speaker 2 (18:50):
So then I was thinking, well, you know, that's the
information that we generate all the time. When we're doing
our research. We take MRIs, quantify, we identify muscles, we
create three dimensional models of the muscles, we figure out
how they're working from that. But I was struck by
the fact that none of that was something that a
(19:12):
clinician could use, despite the fact that MRI is obviously
ubiquitous in healthcare. Right, yeah, you can't go to a
hospital without finding multiple MRIs, but there's no way to
use those MRIs and the way that I was using
them for my research, And I thought, well, that's too bad,
because this would be very useful to the surgeon in
(19:33):
figuring out how to treat these patients. So that was
one light bulb at the beginning. So a lot of
it was figuring out how to take something that we
use in research for very specific, targeted basic science questions
and turn it into something that is useful clinically.
Speaker 1 (19:52):
And again this is like my ignorance, Like I might
have thought, well, you could just do an MRI and
see how big or not big the muscles are, and
infer from how big or not big the muscles are,
how strong or not strong they are. And that sounds straightforward,
but clearly it's not, like, why is it harder than that?
Speaker 2 (20:16):
So a couple of reasons. One is going taking the
MRI pictures and figuring out how big the muscles is
a very challenging problem. So in order to accurately get
how big the muscles are, you have to essentially generate
its shape in three dimensions, so you have to get
the whole length of the muscle, and so you do
(20:37):
that off of multiple MRI pictures. So the MRI essentially
kind of takes pictures through the body at multiple different
slices we call them, going from you know, the abdomen
all the way down to the feet, sort of going
cross sectionally we call it. And so we usually have
over two hundred images. So in each image you have
(21:00):
to find each muscle, and so for any given image
there's probably at least fifteen muscles or more.
Speaker 1 (21:09):
So it wasn't like you could just push the like
show me the muscles button on the MRI and it
would show you the muscles.
Speaker 2 (21:15):
Like.
Speaker 1 (21:15):
Nobody had done that, and there was no obvious way
to do it, certainly not for a surgeon. Ordering a
standard MRON just didn't exist.
Speaker 2 (21:24):
It did not exist.
Speaker 1 (21:25):
So okay, so you realize this, what happens? How do
you make it happen?
Speaker 2 (21:32):
So, you know, one of our first tasks was to
figure out how to get many muscles. So one of
the things that we had done on the research side
is really focus on a couple of muscles, but I
knew for this application that wasn't going to work. We
have to be able to identify any muscle. That was
really the problem is that like you don't know which
(21:52):
one's the problem, so you don't know which one to
look at. So you got to look at all of them.
And then the next task was to figure out all
those muscles and figure out a process to go from
the you know, identifying each and every muscle and each
and every image. So it's called developing an atlas.
Speaker 1 (22:08):
And is that an AI problem?
Speaker 2 (22:12):
So now we have an AI, and it's the type
of AI as supervised learning, where they can you essentially
train the computer to do what the person would do.
But in order to do that, you need to do
what the person would do first. And so we did
that all manually at first, in order to generate one
(22:32):
of these reports. At first, it took us about fifty
hours per person.
Speaker 1 (22:39):
Just going through image after image after image and saying
this is this muscle, that is that muscle exactly.
Speaker 2 (22:45):
So we needed to develop that. But the other piece
we needed is this back to this normative database I
talked about, because if I just told you how big
your muscle is in milli leaders in volume, what are
you going to do with that information?
Speaker 1 (23:00):
Like, oh, great, and nobody knew And it's interesting. It's
one of those things you always think, oh, surely there's
some data in the world that everybody knows X. But
so you're saying, nobody knew what was the kind of
medium size of a particular quadricept for whatever, a healthy
twelve year old boy or whatever. Nobody knew that at
that time. Well, all of.
Speaker 2 (23:18):
The information up until then, for the most part, was
based on dissecting cadavers.
Speaker 1 (23:24):
Uh huh.
Speaker 2 (23:25):
Was based on taking cadavers and dissecting muscles, weighing the muscles.
And you know, one of the big challenges with that
is usually cadavers are older adults, and so they're not
really representative of a younger, healthy population. And I will
(23:46):
tell you at that time that that was a lot
of work and we had I had people saying, like,
why are you doing that. Uh, like that seems like
a waste of time. That's crazy. You know, I had
this vision and I trusted that it was going to
turn into something at least useful to the research community,
and you know, I'm thankful that we stuck with it.
Speaker 1 (24:08):
There's lots more to come on the show, including but
not limited to the work Sylvia and her colleagues are
doing with major league pitchers, college football players, and patients
with degenerative muscle disease. Sylvia and her colleagues trained an
(24:33):
AI model to do what had previously taken a human
fifty hours for every person who got scanned, and they
expanded from working with patients with cerebral palsy to working
with elite athletes. Today, their clients include not just Olympic athletes,
but teams in the NBA and the Premier League. Also,
(24:54):
she told me they're working on a project with Major
League Baseball.
Speaker 2 (24:58):
Yeah, so we're working with the MLB studying pictures and
we're getting essentially a normative database, whole body scan of pictures.
Speaker 1 (25:05):
And is that partly because like pitchers mess up their
arms so badly? Is that kind of the motivation there, Yes.
Speaker 2 (25:12):
There's definitely a lot of issue with with injury and surgery.
And so the idea here is that by taking these scans,
we can really figure out where there might be weaknesses
and sort of potential areas for mitigating the injuries.
Speaker 1 (25:30):
So when you do work for a whole team, like
say the Bulls, you know, a basketball team, an NBA team, Like,
what's the nature of that of that work? What do
you do for a team like that?
Speaker 2 (25:42):
Yeah, we're they will do a baseline of the whole
team and.
Speaker 1 (25:47):
They basically tailor the athletes training presumably strength training in
particular on a muscle by muscle basis, based on the
reports that you're sending them.
Speaker 2 (25:59):
Correct.
Speaker 1 (25:59):
Yeah, yeah, And I mean you can imagine like better
performance being one outcome. Reduced risk of injury seems plausible, right,
Like it seems obvious that like a big asymmetry could
make you more likely to be injured. I mean, are
you at a point now where you can predict the
risk of injury?
Speaker 2 (26:22):
That's like a whole can of worms.
Speaker 1 (26:24):
I won't say, I mean, is that interesting to you?
Or is that like too much? Or yeah?
Speaker 2 (26:30):
No, no, no, this is something we think about a lot,
and let me I want to so first, I'll tell
you why it's a can of worms. Yeah yeah, And
I'll tell you what project.
Speaker 1 (26:42):
Tell me about the can. We'll look at it from
the outside of.
Speaker 2 (26:44):
It, so from the cannon. Like, there's a lot of
technologies out there that will say that they're predicting injury risk.
They'll give you numbers, and they're just not based on anything,
and so I don't know, it's there's.
Speaker 1 (26:58):
A lot of the yeah.
Speaker 2 (27:02):
Yeah, yeah, yeah. So that's not what we're about, Like,
we're about like providing actual things that matter. And so
the question is like can you do do these muscle scans?
Do they correlate with injury likelihood in some way? And
so we actually have a project to address that very question.
(27:24):
It's actually funded by the NFL. Uh. We're actually in
that project. We're working with college teams called college football
teams baselining entire rosters at the beginning of the season
and then tracking hamstring injuries and then if a if
an athlete gets injured, they come back for a scan
(27:46):
at the time of injury and then return to sport.
And so one of our questions is based on the
baseline scan, can we predict who's more likely to get
an injury, an initial injury index injury, and then then
the secondary question is can we predict who will be
re injured? I was saying that we often pairt people
(28:07):
pair it with other things. In this project, we're always
also doing that. Each athlete is getting an assessment of
their sprint mechanics, so kind of the biomechanics of how
they run, and then also assessment of their strength of
their hamstring muscles, so like kind of measured strength. Obviously
you can't do that when they have an injury, of course,
(28:28):
but you can you know when they're healthy.
Speaker 1 (28:30):
So that biomechanics piece seems like something that has been
developing in parallel with your work, also driven by computer vision, right, that,
like markerless motion captures, seems like a big world that
that overlaps with your world some Yeah, So tell me
(28:52):
about your work with female athletes versus male athletes and
how that plays a role.
Speaker 2 (28:59):
Yeah, I will say probably the biggest thing that we've
been focused on is making sure that our data addresses that.
So our normative database is uh separated by sex, so
and it is different because like women aren't small men, right,
So it's important that we have that basis to compare
(29:22):
that's like for women and not comparing to some average
or primarily male data sets. So that's that's one huge
important thing is that it's it's compared to the normative
values for the female population. And then in terms of
like working with the female athletes, I think, you know,
(29:43):
one of the big ones is is really just ability
to personalize and provide this like really accurate detailed assessment
of their of their bodies. And you know, a lot
of the you know, knowledge about like appropriate body composition
historically has been based on studies and men, and but
(30:06):
then we're applying them to women and making us feel
really bad about ourselves. So really motivated to move away
from that and sort of acknowledge the muscular physiology and
anatomy of the female and also the female athlete to
really understand understand that. You know, I think one thing
(30:26):
obviously that we've seen is, you know, acl injuries are
more common in women than men, and examining how these
like recovery profiles look on how they differ between men
and women. That's something that we're observing and seeing how
those things shake out. But we're motivated by really providing
that information that's specific to women.
Speaker 1 (30:49):
So what are some of the non sports things you're
working on, things you're trying to figure out.
Speaker 2 (30:55):
Yeah, I mean, one that's I'm really interested in is
this area that we're applying to in clinical trials for
muscle disease. So we've been working in a specific muscle
disease called fascioscapulo humoral muscular district f SHD, which is
(31:15):
a slowly progressing muscle disease genetic and basis, and so
eventually people with f SHD need a wheelchair. Just life
is very difficult, and so it's pretty devastating. But the
other exciting thing is there are some new treatments out there,
some in particular gene therapies coming online. And now the
(31:37):
challenges do they work because the problem is in these
diseases because they're pretty slowly progressing. If you want to
see if a drug is helping somebody, it's very hard
to see that in a slowly progressing.
Speaker 1 (31:51):
Right and the clinical manifestations are hard to pick up.
If it makes your muscles shrink more slowly, it's going
to be hard to see.
Speaker 2 (32:01):
It's very hard to see, especially from like rudimentary measures.
But with the MRIs. We've been able to provide this
really detailed insight about the disease date of each muscle
and how it's progressing over time. And so you know,
one of our goals is to really lean in on
IS and help figure out exactly how people should look
(32:21):
at all this data and figure out if a drug
is working or not. It's really profoundly important because without that,
these these clinical trials just won't move forward.
Speaker 1 (32:32):
What else are you sort of still trying to figure out?
Speaker 2 (32:37):
So we talked about predicting injury but having all the
data needed to show like if you do if your
scan looks like this, and if you do this, you
will be able to improve your jump high by height.
Speaker 1 (32:50):
By that, yeah, you'll be able to throw a fastball
two miles an hour faster. Like that would be wildly valuable.
Speaker 2 (32:57):
That would be very and we do have data in
our search. We were able to show that these muscle
scores correlate with performance metrics such as jump, hide, and speed.
So we for sure see that The question is then
the spin on like observing how how that plays out,
Like if you then strengthen the appropriate muscles, how how
(33:20):
much faster do you get? And you just need more
and more data to really like to go after that.
But that's one thing that I'm fascinated by. One of
the other interesting ones, Can I go off on a tangent.
Speaker 1 (33:32):
Anything you want?
Speaker 2 (33:35):
One of our research partners that's interested in how muscles
adapt to strength training and different interventions and what influences
that had a really interesting finding that I think is
quite profound but also obvious. So everybody if there, if
they're targeted training their quadrceps and hamstrings, those muscles got bigger,
(33:56):
that makes sense, but in a fair number of the people,
some muscles got smaller. And then you know, he had
done some controlling and in documentation of nutrition intake, and
he found that people that had higher caloric and protein
intake had less of that effect.
Speaker 1 (34:19):
So all that, all the Jim bros. Telling you to
eat a lot of protein are validated by this guy's study.
Speaker 2 (34:25):
Yeah, yeah, but they're not. It's not necessarily to make
that muscle that you're working bigger so you don't lose
the other muscles.
Speaker 1 (34:30):
Uh huh. That's a good one. And was he using
your scans to figure out that?
Speaker 2 (34:37):
Yeah? Yeah, he was using our scans and and the
thing that was cool. Is that Normally in research you
wouldn't bother looking at those other muscles. You would just
look at the ones that were targeted, right, because those
are the ones that you just think about. But by
getting the entire extent of the of all the muscles,
you see these impacts that you wouldn't necessarily.
Speaker 1 (34:55):
Have known, Like he wasn't even looking for it.
Speaker 2 (34:58):
Yeah, it's it's quite profound because somebody's strength training recovering
from an injury. That really means like that, the nutritional
elements important because you could be strengthening some muscles but
weakening others if you're not. If you're not, you know,
playing your cards right there.
Speaker 1 (35:18):
We'll be back in a minute with the lightning round.
M h, I want to finish with the lightning round.
I won't take too long. It'll be fun.
Speaker 2 (35:35):
Okay, what I don't know what that is?
Speaker 1 (35:38):
Well, you'll find out right now.
Speaker 2 (35:41):
Does this have to be fast?
Speaker 1 (35:44):
I could call it the random round? Okay, I like
that random. Have you scanned yourself?
Speaker 2 (35:52):
Yes? Well times? Oh of course that's what kind of thing.
Probably I've been in an MRI machine, I don't know,
maybe a hundred times.
Speaker 1 (36:01):
Like it's not radiation. Right, it's not like an X ray.
You could do it every day.
Speaker 2 (36:05):
As much as you want.
Speaker 1 (36:06):
Yeah, what'd you learn?
Speaker 2 (36:09):
So I actually used it? You know, I've learned lots
over the years, but I will tell you one anecdote.
I have a hip replacement. I have a genetic condition
that leads to early arthritis. And so I was before
I got my hip replacement, I got a scan. I
knew I was getting weak, but holy cow, was I
(36:31):
really weak on that side. What was profound was how
weak my hip flexers were very weak. And I you know,
I think a lot of times people talk about hip
flexures being tight, and that's kind of what I thought
was happening. I felt pain and I felt like I
was having a lot of tightness, but it was actually
(36:52):
weakness and they were like super small on both sides,
but really especially on the on the side that was
effect it. So that was one thing that I worked
on a lot.
Speaker 1 (37:04):
Do that genetic condition you have, did that influence your
your work and all your decision to go into the field.
Speaker 2 (37:13):
I mean it's like loosely maybe because my dad had
the same thing, which actually caused him to go blind. Oh,
it has like a multiple different issues, and so I
think that at an early age got me interested in
medicine and disabilities and like helping helping people, so that
(37:35):
that might be broadly speaking, I didn't and I knew
I had some eye problems. I didn't know the genetic thing.
We didn't discover that till later.
Speaker 1 (37:41):
But huh. Interesting. What's the most underrated muscle in the
human body?
Speaker 2 (37:48):
Hmm, that's a hard one. So I have a few
favorite muscles.
Speaker 1 (37:57):
Okay, what's your favorite muscle?
Speaker 2 (37:58):
Yeah, so the so muscle so as major it's it's
a hip flexer, okay, but it also it's really cool
it actually it's also a back lower back muscle, so
it attaches to the lumbar vertebrae. But then it also
crosses the front of your hip. It's really hard to
find because it's like really back deep in your hip.
(38:19):
It goes right over your.
Speaker 1 (38:20):
Your in the middle of your body kind of yeah,
right in the middle.
Speaker 2 (38:25):
It kind of connects everything, sort of connects your lower
extremity to the rest of your body in some ways.
Speaker 1 (38:31):
Okay, last one, why do you hate astrophysicist Barbie?
Speaker 2 (38:40):
I don't hate anything. I mean, well, it's too perfect.
It's kind of like this, you know idea that like, oh,
you know, you can, you can inspire girls to go
into science by showing them that Barbie does too. But
Barbie's like fictitious, so it kind of tells you that,
(39:02):
like it like promotes the idea of perfectionism in society
but definitely in girls. And you know, what we really
want to promote is almost the opposite of that is
taking risks and and not worrying about being perfect and
just doing something that matters to you. So yeah, I
(39:22):
don't know. I mean, I don't hate it. I had
Barbies when I was a kid, but I just it
kind of.
Speaker 1 (39:27):
Like wrote you wrote a whole column about it.
Speaker 2 (39:34):
That was I was very proud of that. Yeah, no,
and really that, you know, the astrophysicist part, honestly was
more of a hook. The article I had written already
before that Astrophysicist Barbie came to be. It was about
the issue of perfectionism and how that dissuades girls to
(39:55):
go into stem and research.
Speaker 1 (39:58):
Well, a good hook is important, and you found.
Speaker 2 (40:03):
Yeah.
Speaker 1 (40:07):
Sylvia Bleimker is a professor at the University of Virginia
and the co founder of Springbok analytics. Next week on
What's Your Problem, I'll be talking to Jimmy Buffy. He
is using AI to bring the insights of biomechanics to
professional athletes. Jimmy told me that before the advent of AI,
when biomechanics experts tried to work with athletes, it could
(40:30):
be somewhat awkward. So You've got like a picture in
his underwear with a bunch of little metal balls, Take
them and they're like, just pitch like you always pitched right.
Speaker 3 (40:39):
And the state of the art for tracking it was
an awful experience for the people you.
Speaker 2 (40:43):
Were trying to track.
Speaker 1 (40:45):
So what changes?
Speaker 3 (40:46):
The big inflection point was computer vision, basically using artificial
intelligence to identify where those joints are in a camera image,
rather than needing to paste those reflective markers.
Speaker 1 (41:00):
Today's show was produced by Gabriel Hunter Chang, edited by
Lydia Jean Kott, and engineered by Sarah Bugero. You can
email us at Problem at Pushkin dot FM. I'm Jacob Goldstein.
We'll be back next week with another episode of What's
Your Problem.
Hushkin, what's one surprising thing your work has taught you
about elite athletes?
Speaker 2 (00:22):
I never thought I would see muscles that were so
developed they broke our scale. Wow.
Speaker 1 (00:31):
Yeah, like it was just too big the machine, the
AI couldn't figure out what it is.
Speaker 2 (00:35):
Well, no, the AI found it, but we are like
our kind of rating system.
Speaker 1 (00:40):
Wow. Was there a particular athlete or a particular sport
or particular muscle?
Speaker 2 (00:44):
What?
Speaker 1 (00:44):
What? What muscle broke the scale?
Speaker 2 (00:47):
Uh? The gluteus maximus breaks it A.
Speaker 1 (00:49):
Fair kidding fantastic.
Speaker 2 (00:53):
Yes, it's a pain in my butt.
Speaker 1 (00:56):
Like because it's too big.
Speaker 2 (00:58):
Yeah, it's just so big.
Speaker 1 (01:06):
I'm Jacob Goldstein and this is what's your Problem? This
month a bunch of Pushkin podcasts are coming out with
Olympics inspired shows. Revisionist History has a series about America's
decision to participate in Hitler's Berlin Olympics in nineteen thirty six.
The Happiness Lab has an interview with a coach who
coaches coaches and here on What's Your Problem, We're going
(01:30):
to be talking with people who are working at the
frontiers of technology to help elite athletes perform better. For example,
today my guest is Sylvia Blemker. She's a professor of
biomechanical engineering at the University of Virginia, and she's the
co founder of a company called Springbok Analytics. Sylvia's problem
(01:52):
is this, how do you combine MRI scans and artificial
intelligence to generate new insights that can help both elite
athletes and people suffering from diseases that affect the muscles.
Springbox clients include medical researchers, Olympic athletes, major League baseball,
and several professional basketball and soccer teams. You'll hear about
(02:13):
all that on the show, but first we're going to
pick up where we left off in the conversation. We
were discussing the extraordinarily large muscles of elite athletes, and
then Sylvia told me something even more surprising.
Speaker 2 (02:30):
The other thing is that they have some tiny muscles too, Huh.
Speaker 1 (02:33):
Like they have like smaller than a normal person's muscle.
Speaker 2 (02:36):
Much smaller.
Speaker 1 (02:37):
Huh.
Speaker 2 (02:38):
They put their muscle where they need it.
Speaker 1 (02:40):
What's an example, Like what muscle is tiny and what
kind of athletes?
Speaker 2 (02:44):
Calf muscles are small in most fast athletes huh? And
ether you look at a sprinter or like a running back.
Speaker 1 (02:53):
It's just all quad no calf, all.
Speaker 2 (02:57):
Like thigh no calf, yeah, THI And it kind of
makes sense because you know, if you're trying to run fast,
you wouldn't want to put a lot of mass like
at the end of your leg. It's like as a
lot of inertia to like move your leg.
Speaker 1 (03:09):
Huh.
Speaker 2 (03:09):
Because you know, the muscles are important for sprinting, that's
the interesting thing, but they just don't they're small, very.
Speaker 1 (03:16):
H huh uh huh. So I'm particularly interested at this
moment in the sports piece of what you do. I'm curious,
by the way. Do you work with any Olympic teams
or Olympic athletes? Yeah?
Speaker 2 (03:33):
Yeah, We've actually been working with several different Olympic athletes.
The ones that probably that come to mind most are
multiple players on the US women's national soccer team.
Speaker 1 (03:46):
Oh cool, tell me, like, tell me the story of
that of that work. So they came to you, what
did what did they what do they want when they
came to you? Like, how did that? How did that begin?
Speaker 2 (03:57):
They came to us along with their team. So the
technology we provide, you know, an athlete could understand it,
but really with their team to help them figure out
how to keep athletes healthy.
Speaker 1 (04:10):
So what did they what did they say? What did
they say when they came to.
Speaker 2 (04:14):
So, for example, one athlete that's coming to mind had
a known imbalance side to side that based on a
history of injury, and they really wanted to know where
that imbalance was coming from.
Speaker 1 (04:30):
So the woman had had hurt one of her legs,
and that leg was even after she came back, that
leg was weaker essentially than the other. I mean, is
that the sort of gross you know, macro.
Speaker 2 (04:41):
Way, Yeah, exactly, that's a that's a nice way to
put it.
Speaker 1 (04:43):
Yeah, And and they wanted a sort of finder like, okay,
but we can see that, but what's going on on
the inside, like muscle by muscle tell us that, yes.
Speaker 2 (04:53):
Exactly, that's precisely what we do. We go on the
inside because on the outside you see perhaps that her
knee extents or quads seem weaker on one side than
the other. But there's four quads, quadre steps, four muscles,
and so it's not clear which of those muscles are
actually the culprit for that imbalance and in what way.
Speaker 1 (05:15):
Good So this is their question, and then what happens next?
Speaker 2 (05:20):
So This first step is an MRI scan, and so
with these athletes or teams, we have ways to connect
them with an MRI machine, whether it be through an
imaging center that they partner with, or we've even actually
brought MRI mobile trucks to sites to make it.
Speaker 1 (05:45):
Like players run off the field and get an MRI
and go back and keep playing. Yeah, yeah, kind yeah, yeah.
Speaker 2 (05:51):
It helps just with the timing of things. But so
first we connect them there, so it takes about ten minutes.
Then they send those pictures up into the cloud into
our server and then we crunched through it and then
we send back a report on their muscles. We also
have what we call it interactive Viewer, and it's presented
(06:14):
in the form of a three D model. Three dimensional model,
so you actually see your own legs, the muscles and bones,
your own muscles and bones that we've identified from the
images going through a process called segmentation where we find
all the muscles and bones and then we reconstruct them,
so it's kind of like a digital twin of that
(06:35):
person that they can see on their computer. And so
along with it or a number or all these metrics
that helps them understand their balance. The development or strength
of the muscles and the health of the muscles.
Speaker 1 (06:50):
So tell me about this report, they get like, what
does it? What does it say?
Speaker 2 (06:55):
So the basis of that is actually a lot of
research that we did over many years, because you need
to understand where somebody falls relative to a normal essentially
to give them. Essentially, we have a scoring system for
the muscles and that's based on comparing with a large
data set of healthy individuals. And so we know for
(07:18):
a given person, based on their sex, age, height and weight,
how big we expect all the muscles to be. And
that's through a lot of previous research. So then we
can say, okay, here's where you land each particular muscle
compared to this nor what we call a normative database,
(07:38):
so we call it a spring box score.
Speaker 1 (07:40):
Do you do it for every muscle in the leg.
Speaker 2 (07:42):
Or we do it? So our primary product that we
started with was every muscle in the legs essentially from
belly to feet, no muscle left behind. They're all important.
Speaker 1 (07:54):
How many muscles are there.
Speaker 2 (07:56):
Thirty five per leg?
Speaker 1 (07:58):
So seven okay, okay, yeah, more than I would have guessed,
but a lot. Yeah, all important, okay. So, and so
it's basically, how strong and healthy is every one of
those seventy muscles relative to baseline?
Speaker 2 (08:16):
Right, And then the asymmetry comes where you can compare
side to side. So for each of the thirty five
muscles that they exist on each leg, we can say
which side is bigger, which side is smaller, and by
how much? And then we also have normative values for
that because we're all just slightly asymmetric.
Speaker 1 (08:34):
Uh huh, And presumably some muscles are more asymmetric than others,
And so you want to know kind of how how
asymmetric relative to baseline is this particular pair of muscles exactly? Yeah,
And and so in the in the case of this
uh soccer player who came to you who you know,
knew knew she had some kind of problem with her
(08:56):
quadriceps on one side, but didn't know what was going on.
What did you find?
Speaker 2 (09:01):
We found some imbalances, and actually not just in those muscles.
It turns out that, you know, it's all connected. So
if you have a weakness or an imbalance and one
set of muscles, usually some other set of muscles are
compensating in someone.
Speaker 1 (09:18):
Yeah, well, it's like when you like mess up. Even
if you're just a recreational athlete, right Like, if you
mess up something, you mess up your ankle, then you
start walking funny, and then like your back hurts because
you're walking funny. Right Like, that is a very anecdotally
apparent thing.
Speaker 2 (09:33):
Yeah, yeah, we all know that, but that you know
it shows through in the skin. But the thing is
that it's not very intuitive from the outside which muscles
have been affected, how they've compensated, and it looks different
for every single person.
Speaker 1 (09:46):
Huh.
Speaker 2 (09:47):
So that's why the report is very valuable because for
that person, they know exactly which muscles are the ones
that they really need to target, both the ones that
they already thought maybe were an issue, but then all
the other ones that showed up and they didn't really realize.
Speaker 1 (10:03):
And so in the case of this soccer player, was
it like one particular quadrcep on one side that was
like the core thing and you could figure out which
one it was.
Speaker 2 (10:12):
There was a few mess It wasn't just that. I
think there were at least one calf muscle and then
some in especially in the deep hip, those were impacted.
So yeah, it kind of shows up everywhere.
Speaker 1 (10:26):
And so you have this essentially diagnosis, right, a very
sort of fine grained kind of diagnosis. Do you also
have a have a prescription? Do you have sort of
particular kinds of training to address these very fine grained things,
or do you leave that to the trainers or whoever?
Speaker 2 (10:44):
We leave that to the trainers, because I think that
it's also important to have all the other information about
the athlete. We're not arguing that it replaces everything else.
And people pair it with lots of different other types
of measurements depending on the application or in the setting,
Like some people pair it with let's say, metrics of
(11:06):
jump performance. I'm shifting over to basketball here, but that's
just one that came to mind, where you can look
at the asymmetry about how how an athlete jumps, but
then you can also compare it to the asymmetry of
their muscles and get some insight. So it definitely, you know,
plugs in with a lot of other things.
Speaker 1 (11:25):
And to what extent can trainers or you know, strength
coaches develop programs that are sufficiently kind of fine grained
to match the kind of fine grained findings you're having, right, Like,
for example, if you find, as I understand you did
that a soccer player has one particular quadricep that is weak. Like,
(11:48):
are there workouts that target a single quadricep and not
the others?
Speaker 2 (11:52):
Yep, there are.
Speaker 1 (11:53):
That's cool for whichever quadricepp you're just like, just for fun,
give me an example.
Speaker 2 (11:59):
You know. One one way that it's very simple is
using something called biofeedback. Huh. So you can measure whether
you use something called EMG, which is a way to
measure how much electrical activity is a muscle, and then
you can see which muscles you're using for a given task.
So if you give people the feedback of which of
(12:20):
those muscles they're using and say, oh, no, you're not
using this one, use this one more, that actually works
very effectively.
Speaker 1 (12:27):
Oh really, So you can basically use your brain if
you're getting the feedback to focus on which quadricep you're training.
Speaker 2 (12:37):
Yeah, and there's other ways you can give the feedback
in other different ways, but yeah, our brains are very
good at that. Once they get feedback, they're very good
at learning.
Speaker 1 (12:46):
That's cool, especially somehow to think of what elead athletes right,
because they are already presumably like super dialed in in
terms of like the relationship between their brain and their
body at this very elite level exactly.
Speaker 2 (12:57):
Yeah. The other I was going to mention a lot
of players and teams use this not just one time,
but overtime. So they'll get a scan, figure out a plan,
work on that for maybe three months or six months,
and then do another scan and see how things are
progressing and adjust accordingly. So that's definitely another way to
(13:21):
in the long term see if what they're doing is
resulting in the change that they're hoping to see.
Speaker 1 (13:27):
So, what happened with that soccer player who had the
weak quadricep and other related Yeah, No, I.
Speaker 2 (13:34):
Think she's doing great, like staying healthy and and you know,
getting ready.
Speaker 1 (13:41):
Yeah. So I know you can't tell us her name,
but will we see her in the Olympics this time? Great? So,
as you were talking about that, I mean there was
a moment where it was like, Okay, the athlete goes
and gets the MRI and then you get the scan.
You get the scan, and then you said, like, you
crunch through the numbers and then you make the report.
Ssumably you crunching through the numbers is like the result
(14:04):
of many many years of work and kind of the
core of what your company does. So I want to
talk a little bit more about that and kind of
how you how you got here, Like how did you
come to start the company?
Speaker 2 (14:16):
Mm hmm? How long do I have a while?
Speaker 1 (14:22):
I mean it's not you know, it's not the radio,
it's a podcast.
Speaker 2 (14:25):
You. I am a professor too, so I can go on.
Speaker 1 (14:27):
Let me ask you this, What was the moment when
you decided to start the company?
Speaker 2 (14:32):
I can give you one moment and then we can
do a.
Speaker 1 (14:36):
Couple of moments. Yeah, that's what's what a good story is,
like three moments.
Speaker 2 (14:43):
In time, right right? Three not four?
Speaker 1 (14:46):
Four is a tricky number. We could do five, or
we could do one. I think I think is better.
Speaker 2 (14:52):
Yeah. Yeah. So I'm a professor. I run a lab
and form my entire career. I've been fascinated with must
and how it works, and fascinated by something that we
all in like the muscle field. We call form function relationships.
So the idea that the way a muscle is shaped
(15:13):
and the way it's structured and how big it is
influences how well it works or how well it functions,
how strong it is, how well it behaves. And there's
a lot too there, and there's a lot of nuance.
And that's like I've spent a career studying that in
lots of different ways. So I've always been interested in
quantifying muscle and figuring out how that influence is, how
it works, and both in healthy people or in athletes,
(15:36):
and also in different patient populations different In particular, I
have an interest in movement disorders, so neuromuscular diseases that
lead to impairments and mobility and movement ability. So one
of the light bulb moments for this was the fact
that I'd been using MRI to study muscle in my
(15:59):
research for a long time. It's kind of a ubiquitous
tool or like often use tool and research. But I
was struck by the fact that I was hearing from,
in particular a surgeon collaborator, and the surgeon was telling
me about his work and helping kids with cerebral palsy
(16:22):
improve their movement where they had hindered movement and largely
because their muscles are impacted. Not only do they have
an impaired ability to control their muscles, but their muscles
end up with impairments in their structure or their form,
which then influences how well they work. So surgeons have
(16:42):
to go in and do surgeries to try to change that.
They do things like modify tendons to try to make
muscles less stiff, or they transfer muscles to make them
do a new thing. But one of the big tricky
parts is that oftentimes some of those muscles are very week.
So if they choose the wrong muscle, then they'll make
(17:03):
a weak muscle even weaker. Huh, And that's catastrophic. So
it's a very fine line that a surgeon has to
figure out and they have to go in. You know,
we just talked about there's thirty five muscle muscles in
each leg, So which of the muscles are the ones
that should be operated on and which ones should be avoided?
And so what my collaborator, a guy named doctor Mark Abel,
(17:26):
fantastic surgeon. He was telling me, Yeah, like, it's very hard,
and I don't he didn't have a way to see that.
All that he could do is look from the outside.
No technology could give him the information he needed to
figure out which muscles he should focus on and which
ones to.
Speaker 1 (17:41):
Avoid, because it's not obvious by looking what's a strong
muscle and what's a weak one. Yeah, I guess that's
surprising to me on some level, Like I don't I've
never thought about it. But naively, I would think you
could look at a muscle and say it looks strong
or it looks weak.
Speaker 2 (17:59):
Not right, because you just see it from the surface.
You don't see it on the inside. And the other challenges.
For every joint, there's many muscles. So like we just
said that quadrceps has four muscles on the back of
the leg, ham strings, there's three hamstrings muscles. There's other
muscles that are in the thigh. So you're just seeing
(18:21):
what's an impairment and the overall movement, let's say of
a joint. But then there could be many muscles or
combinations of muscles that are leading to that, and you
don't know when you look from the outside. Our body
is designed that way actually to be somewhat we call
it redundant. We have more muscles than we need probably,
but if you think about imbalances, then any one of
(18:41):
those muscles could create it create some trouble.
Speaker 1 (18:44):
So, okay, so the surgeon describes this problem he's having,
and then and then what do you do?
Speaker 2 (18:50):
So then I was thinking, well, you know, that's the
information that we generate all the time. When we're doing
our research. We take MRIs, quantify, we identify muscles, we
create three dimensional models of the muscles, we figure out
how they're working from that. But I was struck by
the fact that none of that was something that a
(19:12):
clinician could use, despite the fact that MRI is obviously
ubiquitous in healthcare. Right, yeah, you can't go to a
hospital without finding multiple MRIs, but there's no way to
use those MRIs and the way that I was using
them for my research, And I thought, well, that's too bad,
because this would be very useful to the surgeon in
(19:33):
figuring out how to treat these patients. So that was
one light bulb at the beginning. So a lot of
it was figuring out how to take something that we
use in research for very specific, targeted basic science questions
and turn it into something that is useful clinically.
Speaker 1 (19:52):
And again this is like my ignorance, Like I might
have thought, well, you could just do an MRI and
see how big or not big the muscles are, and
infer from how big or not big the muscles are,
how strong or not strong they are. And that sounds straightforward,
but clearly it's not, like, why is it harder than that?
Speaker 2 (20:16):
So a couple of reasons. One is going taking the
MRI pictures and figuring out how big the muscles is
a very challenging problem. So in order to accurately get
how big the muscles are, you have to essentially generate
its shape in three dimensions, so you have to get
the whole length of the muscle, and so you do
(20:37):
that off of multiple MRI pictures. So the MRI essentially
kind of takes pictures through the body at multiple different
slices we call them, going from you know, the abdomen
all the way down to the feet, sort of going
cross sectionally we call it. And so we usually have
over two hundred images. So in each image you have
(21:00):
to find each muscle, and so for any given image
there's probably at least fifteen muscles or more.
Speaker 1 (21:09):
So it wasn't like you could just push the like
show me the muscles button on the MRI and it
would show you the muscles.
Speaker 2 (21:15):
Like.
Speaker 1 (21:15):
Nobody had done that, and there was no obvious way
to do it, certainly not for a surgeon. Ordering a
standard MRON just didn't exist.
Speaker 2 (21:24):
It did not exist.
Speaker 1 (21:25):
So okay, so you realize this, what happens? How do
you make it happen?
Speaker 2 (21:32):
So, you know, one of our first tasks was to
figure out how to get many muscles. So one of
the things that we had done on the research side
is really focus on a couple of muscles, but I
knew for this application that wasn't going to work. We
have to be able to identify any muscle. That was
really the problem is that like you don't know which
(21:52):
one's the problem, so you don't know which one to
look at. So you got to look at all of them.
And then the next task was to figure out all
those muscles and figure out a process to go from
the you know, identifying each and every muscle and each
and every image. So it's called developing an atlas.
Speaker 1 (22:08):
And is that an AI problem?
Speaker 2 (22:12):
So now we have an AI, and it's the type
of AI as supervised learning, where they can you essentially
train the computer to do what the person would do.
But in order to do that, you need to do
what the person would do first. And so we did
that all manually at first, in order to generate one
(22:32):
of these reports. At first, it took us about fifty
hours per person.
Speaker 1 (22:39):
Just going through image after image after image and saying
this is this muscle, that is that muscle exactly.
Speaker 2 (22:45):
So we needed to develop that. But the other piece
we needed is this back to this normative database I
talked about, because if I just told you how big
your muscle is in milli leaders in volume, what are
you going to do with that information?
Speaker 1 (23:00):
Like, oh, great, and nobody knew And it's interesting. It's
one of those things you always think, oh, surely there's
some data in the world that everybody knows X. But
so you're saying, nobody knew what was the kind of
medium size of a particular quadricept for whatever, a healthy
twelve year old boy or whatever. Nobody knew that at
that time. Well, all of.
Speaker 2 (23:18):
The information up until then, for the most part, was
based on dissecting cadavers.
Speaker 1 (23:24):
Uh huh.
Speaker 2 (23:25):
Was based on taking cadavers and dissecting muscles, weighing the muscles.
And you know, one of the big challenges with that
is usually cadavers are older adults, and so they're not
really representative of a younger, healthy population. And I will
(23:46):
tell you at that time that that was a lot
of work and we had I had people saying, like,
why are you doing that. Uh, like that seems like
a waste of time. That's crazy. You know, I had
this vision and I trusted that it was going to
turn into something at least useful to the research community,
and you know, I'm thankful that we stuck with it.
Speaker 1 (24:08):
There's lots more to come on the show, including but
not limited to the work Sylvia and her colleagues are
doing with major league pitchers, college football players, and patients
with degenerative muscle disease. Sylvia and her colleagues trained an
(24:33):
AI model to do what had previously taken a human
fifty hours for every person who got scanned, and they
expanded from working with patients with cerebral palsy to working
with elite athletes. Today, their clients include not just Olympic athletes,
but teams in the NBA and the Premier League. Also,
(24:54):
she told me they're working on a project with Major
League Baseball.
Speaker 2 (24:58):
Yeah, so we're working with the MLB studying pictures and
we're getting essentially a normative database, whole body scan of pictures.
Speaker 1 (25:05):
And is that partly because like pitchers mess up their
arms so badly? Is that kind of the motivation there, Yes.
Speaker 2 (25:12):
There's definitely a lot of issue with with injury and surgery.
And so the idea here is that by taking these scans,
we can really figure out where there might be weaknesses
and sort of potential areas for mitigating the injuries.
Speaker 1 (25:30):
So when you do work for a whole team, like
say the Bulls, you know, a basketball team, an NBA team, Like,
what's the nature of that of that work? What do
you do for a team like that?
Speaker 2 (25:42):
Yeah, we're they will do a baseline of the whole
team and.
Speaker 1 (25:47):
They basically tailor the athletes training presumably strength training in
particular on a muscle by muscle basis, based on the
reports that you're sending them.
Speaker 2 (25:59):
Correct.
Speaker 1 (25:59):
Yeah, yeah, And I mean you can imagine like better
performance being one outcome. Reduced risk of injury seems plausible, right,
Like it seems obvious that like a big asymmetry could
make you more likely to be injured. I mean, are
you at a point now where you can predict the
risk of injury?
Speaker 2 (26:22):
That's like a whole can of worms.
Speaker 1 (26:24):
I won't say, I mean, is that interesting to you?
Or is that like too much? Or yeah?
Speaker 2 (26:30):
No, no, no, this is something we think about a lot,
and let me I want to so first, I'll tell
you why it's a can of worms. Yeah yeah, And
I'll tell you what project.
Speaker 1 (26:42):
Tell me about the can. We'll look at it from
the outside of.
Speaker 2 (26:44):
It, so from the cannon. Like, there's a lot of
technologies out there that will say that they're predicting injury risk.
They'll give you numbers, and they're just not based on anything,
and so I don't know, it's there's.
Speaker 1 (26:58):
A lot of the yeah.
Speaker 2 (27:02):
Yeah, yeah, yeah. So that's not what we're about, Like,
we're about like providing actual things that matter. And so
the question is like can you do do these muscle scans?
Do they correlate with injury likelihood in some way? And
so we actually have a project to address that very question.
(27:24):
It's actually funded by the NFL. Uh. We're actually in
that project. We're working with college teams called college football
teams baselining entire rosters at the beginning of the season
and then tracking hamstring injuries and then if a if
an athlete gets injured, they come back for a scan
(27:46):
at the time of injury and then return to sport.
And so one of our questions is based on the
baseline scan, can we predict who's more likely to get
an injury, an initial injury index injury, and then then
the secondary question is can we predict who will be
re injured? I was saying that we often pairt people
(28:07):
pair it with other things. In this project, we're always
also doing that. Each athlete is getting an assessment of
their sprint mechanics, so kind of the biomechanics of how
they run, and then also assessment of their strength of
their hamstring muscles, so like kind of measured strength. Obviously
you can't do that when they have an injury, of course,
(28:28):
but you can you know when they're healthy.
Speaker 1 (28:30):
So that biomechanics piece seems like something that has been
developing in parallel with your work, also driven by computer vision, right, that,
like markerless motion captures, seems like a big world that
that overlaps with your world some Yeah, So tell me
(28:52):
about your work with female athletes versus male athletes and
how that plays a role.
Speaker 2 (28:59):
Yeah, I will say probably the biggest thing that we've
been focused on is making sure that our data addresses that.
So our normative database is uh separated by sex, so
and it is different because like women aren't small men, right,
So it's important that we have that basis to compare
(29:22):
that's like for women and not comparing to some average
or primarily male data sets. So that's that's one huge
important thing is that it's it's compared to the normative
values for the female population. And then in terms of
like working with the female athletes, I think, you know,
(29:43):
one of the big ones is is really just ability
to personalize and provide this like really accurate detailed assessment
of their of their bodies. And you know, a lot
of the you know, knowledge about like appropriate body composition
historically has been based on studies and men, and but
(30:06):
then we're applying them to women and making us feel
really bad about ourselves. So really motivated to move away
from that and sort of acknowledge the muscular physiology and
anatomy of the female and also the female athlete to
really understand understand that. You know, I think one thing
(30:26):
obviously that we've seen is, you know, acl injuries are
more common in women than men, and examining how these
like recovery profiles look on how they differ between men
and women. That's something that we're observing and seeing how
those things shake out. But we're motivated by really providing
that information that's specific to women.
Speaker 1 (30:49):
So what are some of the non sports things you're
working on, things you're trying to figure out.
Speaker 2 (30:55):
Yeah, I mean, one that's I'm really interested in is
this area that we're applying to in clinical trials for
muscle disease. So we've been working in a specific muscle
disease called fascioscapulo humoral muscular district f SHD, which is
(31:15):
a slowly progressing muscle disease genetic and basis, and so
eventually people with f SHD need a wheelchair. Just life
is very difficult, and so it's pretty devastating. But the
other exciting thing is there are some new treatments out there,
some in particular gene therapies coming online. And now the
(31:37):
challenges do they work because the problem is in these
diseases because they're pretty slowly progressing. If you want to
see if a drug is helping somebody, it's very hard
to see that in a slowly progressing.
Speaker 1 (31:51):
Right and the clinical manifestations are hard to pick up.
If it makes your muscles shrink more slowly, it's going
to be hard to see.
Speaker 2 (32:01):
It's very hard to see, especially from like rudimentary measures.
But with the MRIs. We've been able to provide this
really detailed insight about the disease date of each muscle
and how it's progressing over time. And so you know,
one of our goals is to really lean in on
IS and help figure out exactly how people should look
(32:21):
at all this data and figure out if a drug
is working or not. It's really profoundly important because without that,
these these clinical trials just won't move forward.
Speaker 1 (32:32):
What else are you sort of still trying to figure out?
Speaker 2 (32:37):
So we talked about predicting injury but having all the
data needed to show like if you do if your
scan looks like this, and if you do this, you
will be able to improve your jump high by height.
Speaker 1 (32:50):
By that, yeah, you'll be able to throw a fastball
two miles an hour faster. Like that would be wildly valuable.
Speaker 2 (32:57):
That would be very and we do have data in
our search. We were able to show that these muscle
scores correlate with performance metrics such as jump, hide, and speed.
So we for sure see that The question is then
the spin on like observing how how that plays out,
Like if you then strengthen the appropriate muscles, how how
(33:20):
much faster do you get? And you just need more
and more data to really like to go after that.
But that's one thing that I'm fascinated by. One of
the other interesting ones, Can I go off on a tangent.
Speaker 1 (33:32):
Anything you want?
Speaker 2 (33:35):
One of our research partners that's interested in how muscles
adapt to strength training and different interventions and what influences
that had a really interesting finding that I think is
quite profound but also obvious. So everybody if there, if
they're targeted training their quadrceps and hamstrings, those muscles got bigger,
(33:56):
that makes sense, but in a fair number of the people,
some muscles got smaller. And then you know, he had
done some controlling and in documentation of nutrition intake, and
he found that people that had higher caloric and protein
intake had less of that effect.
Speaker 1 (34:19):
So all that, all the Jim bros. Telling you to
eat a lot of protein are validated by this guy's study.
Speaker 2 (34:25):
Yeah, yeah, but they're not. It's not necessarily to make
that muscle that you're working bigger so you don't lose
the other muscles.
Speaker 1 (34:30):
Uh huh. That's a good one. And was he using
your scans to figure out that?
Speaker 2 (34:37):
Yeah? Yeah, he was using our scans and and the
thing that was cool. Is that Normally in research you
wouldn't bother looking at those other muscles. You would just
look at the ones that were targeted, right, because those
are the ones that you just think about. But by
getting the entire extent of the of all the muscles,
you see these impacts that you wouldn't necessarily.
Speaker 1 (34:55):
Have known, Like he wasn't even looking for it.
Speaker 2 (34:58):
Yeah, it's it's quite profound because somebody's strength training recovering
from an injury. That really means like that, the nutritional
elements important because you could be strengthening some muscles but
weakening others if you're not. If you're not, you know,
playing your cards right there.
Speaker 1 (35:18):
We'll be back in a minute with the lightning round.
M h, I want to finish with the lightning round.
I won't take too long. It'll be fun.
Speaker 2 (35:35):
Okay, what I don't know what that is?
Speaker 1 (35:38):
Well, you'll find out right now.
Speaker 2 (35:41):
Does this have to be fast?
Speaker 1 (35:44):
I could call it the random round? Okay, I like
that random. Have you scanned yourself?
Speaker 2 (35:52):
Yes? Well times? Oh of course that's what kind of thing.
Probably I've been in an MRI machine, I don't know,
maybe a hundred times.
Speaker 1 (36:01):
Like it's not radiation. Right, it's not like an X ray.
You could do it every day.
Speaker 2 (36:05):
As much as you want.
Speaker 1 (36:06):
Yeah, what'd you learn?
Speaker 2 (36:09):
So I actually used it? You know, I've learned lots
over the years, but I will tell you one anecdote.
I have a hip replacement. I have a genetic condition
that leads to early arthritis. And so I was before
I got my hip replacement, I got a scan. I
knew I was getting weak, but holy cow, was I
(36:31):
really weak on that side. What was profound was how
weak my hip flexers were very weak. And I you know,
I think a lot of times people talk about hip
flexures being tight, and that's kind of what I thought
was happening. I felt pain and I felt like I
was having a lot of tightness, but it was actually
(36:52):
weakness and they were like super small on both sides,
but really especially on the on the side that was
effect it. So that was one thing that I worked
on a lot.
Speaker 1 (37:04):
Do that genetic condition you have, did that influence your
your work and all your decision to go into the field.
Speaker 2 (37:13):
I mean it's like loosely maybe because my dad had
the same thing, which actually caused him to go blind. Oh,
it has like a multiple different issues, and so I
think that at an early age got me interested in
medicine and disabilities and like helping helping people, so that
(37:35):
that might be broadly speaking, I didn't and I knew
I had some eye problems. I didn't know the genetic thing.
We didn't discover that till later.
Speaker 1 (37:41):
But huh. Interesting. What's the most underrated muscle in the
human body?
Speaker 2 (37:48):
Hmm, that's a hard one. So I have a few
favorite muscles.
Speaker 1 (37:57):
Okay, what's your favorite muscle?
Speaker 2 (37:58):
Yeah, so the so muscle so as major it's it's
a hip flexer, okay, but it also it's really cool
it actually it's also a back lower back muscle, so
it attaches to the lumbar vertebrae. But then it also
crosses the front of your hip. It's really hard to
find because it's like really back deep in your hip.
(38:19):
It goes right over your.
Speaker 1 (38:20):
Your in the middle of your body kind of yeah,
right in the middle.
Speaker 2 (38:25):
It kind of connects everything, sort of connects your lower
extremity to the rest of your body in some ways.
Speaker 1 (38:31):
Okay, last one, why do you hate astrophysicist Barbie?
Speaker 2 (38:40):
I don't hate anything. I mean, well, it's too perfect.
It's kind of like this, you know idea that like, oh,
you know, you can, you can inspire girls to go
into science by showing them that Barbie does too. But
Barbie's like fictitious, so it kind of tells you that,
(39:02):
like it like promotes the idea of perfectionism in society
but definitely in girls. And you know, what we really
want to promote is almost the opposite of that is
taking risks and and not worrying about being perfect and
just doing something that matters to you. So yeah, I
(39:22):
don't know. I mean, I don't hate it. I had
Barbies when I was a kid, but I just it
kind of.
Speaker 1 (39:27):
Like wrote you wrote a whole column about it.
Speaker 2 (39:34):
That was I was very proud of that. Yeah, no,
and really that, you know, the astrophysicist part, honestly was
more of a hook. The article I had written already
before that Astrophysicist Barbie came to be. It was about
the issue of perfectionism and how that dissuades girls to
(39:55):
go into stem and research.
Speaker 1 (39:58):
Well, a good hook is important, and you found.
Speaker 2 (40:03):
Yeah.
Speaker 1 (40:07):
Sylvia Bleimker is a professor at the University of Virginia
and the co founder of Springbok analytics. Next week on
What's Your Problem, I'll be talking to Jimmy Buffy. He
is using AI to bring the insights of biomechanics to
professional athletes. Jimmy told me that before the advent of AI,
when biomechanics experts tried to work with athletes, it could
(40:30):
be somewhat awkward. So You've got like a picture in
his underwear with a bunch of little metal balls, Take
them and they're like, just pitch like you always pitched right.
Speaker 3 (40:39):
And the state of the art for tracking it was
an awful experience for the people you.
Speaker 2 (40:43):
Were trying to track.
Speaker 1 (40:45):
So what changes?
Speaker 3 (40:46):
The big inflection point was computer vision, basically using artificial
intelligence to identify where those joints are in a camera image,
rather than needing to paste those reflective markers.
Speaker 1 (41:00):
Today's show was produced by Gabriel Hunter Chang, edited by
Lydia Jean Kott, and engineered by Sarah Bugero. You can
email us at Problem at Pushkin dot FM. I'm Jacob Goldstein.
We'll be back next week with another episode of What's
Your Problem.
Popular Podcasts
1. Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
2. Stuff You Missed in History Class
Join Holly and Tracy as they bring you the greatest and strangest Stuff You Missed In History Class in this podcast by iHeartRadio.
3. Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations.