Breakthrough Techniques to Improve Hearing with Engineer Jack Noble

WEBVTT
00:00:01.590 --> 00:00:09.240
Welcome to "Quantum Potential," a podcast dedicated to curiosity, discovery and the bold new ideas shaping our future.
00:00:09.929 --> 00:00:12.810
I'm Cybele Raver, provost of Vanderbilt University.
00:00:19.675 --> 00:00:26.699
Today's guest is a rising star whose work is at the frontier of what's possible for medicine, health and society.
00:00:27.585 --> 00:00:32.024
Jack Noble is an associate professor of electrical and computer engineering here at Vanderbilt.
00:00:32.594 --> 00:00:43.515
Importantly, Jack has been super successful in winning federal funding from the National Institutes of Health for his groundbreaking research on ways to restore hearing for people who have profound hearing loss.
00:00:44.204 --> 00:00:49.664
If you or someone you love has ever experienced severe hearing loss, this conversation is going to feel very personal.
00:00:54.479 --> 00:00:57.479
Jack, it's such a pleasure to have you as my guest today.
00:00:57.479 --> 00:01:06.450
I really look forward to diving deep into the details of your incredibly innovative work as a researcher at the intersection of engineering, surgical medicine and AI.
00:01:06.840 --> 00:01:11.234
But first, I wanted to give our audience a sense of the magnitude of the kinds of challenges you're taking on.
00:01:12.109 --> 00:01:20.689
Your recently federally funded grant frames your work by highlighting the ways that hundreds of thousands of people get surgery to try to restore their hearing.
00:01:21.019 --> 00:01:25.399
But the past effort that's been pursued by surgeons has been pretty imprecise.
00:01:25.879 --> 00:01:29.509
Tell us a little bit more about the scope of the problem that you're trying to solve.
00:01:30.019 --> 00:01:33.289
Well, first thank you for that very generous, uh, introduction.
00:01:33.840 --> 00:01:35.129
I'm really honored to be here.
00:01:35.640 --> 00:01:39.060
This is a wonderful opportunity to, to talk about this, this work.
00:01:39.450 --> 00:01:43.469
Cochlear implants are generally a very remarkable technology.
00:01:43.469 --> 00:01:49.469
They work, uh, remarkably well for, uh, so many people, but there's a lot of variability in outcomes.
00:01:50.069 --> 00:02:06.554
One thing that we know that contributes to this and impacts the ability of, of some individuals to be able to recognize speech is how the electrode array that's surgically implanted into the inner ear interfaces with the auditory nerve.
00:02:06.734 --> 00:02:07.094
Okay.
00:02:07.094 --> 00:02:18.525
And just to slow down enough so that I understand, an electrode array sounds like it's 12 to 24 tiny wires, what's going on there?
00:02:18.615 --> 00:02:26.984
I mean, I think about the electrodes that get stuck on my skin if we're doing an EKG, for example, but that's a big old patch that can't possibly be squeezed into my ear, right?
00:02:26.984 --> 00:02:35.449
So help me understand how this array of electrodes looks, and how it is possible to get all of those things inside one's ear.
00:02:35.539 --> 00:02:36.049
Sure.
00:02:36.079 --> 00:02:51.560
So the electrode arrays differ a little bit, uh, from one manufacturer to another, but basically they're all these kind of floppy tubes, uh, that have, uh, a sequence of contacts sort of placed along the length of the tube.
00:02:52.159 --> 00:02:54.649
And these contacts are sub-millimetric. They're tiny.
00:02:54.739 --> 00:03:03.419
Uh, the surface of them is maybe about 0.3 millimeters by 0.3 millimeters wide, and they're arranged on this tube that's maybe about a centimeter long.
00:03:03.569 --> 00:03:08.550
Um, and so in surgery, uh, the surgeon has to, uh, gain access to the inner ear.
00:03:09.120 --> 00:03:25.664
Um, this requires making, uh, an incision behind the ear, then drilling away, uh, some of the mastoid bone behind the ear to get access first to the middle ear, where they can then see the bony surface of the cochlea, which is also called the inner ear.
00:03:26.115 --> 00:03:32.775
So in surgery, and I've inspected thousands of, uh, of video recordings of, of surgeries.
00:03:32.775 --> 00:03:38.444
So I've had a chance to really get a, a good sense for, uh, what the surgeons can see and what they can't see.
00:03:38.865 --> 00:03:55.979
But in surgery, what they see is this bony surface anatomy of the inner ear, as well as one tiny membrane that's called the round window membrane, uh, which is what they call the entrance to the cochlea, and this is where they make a tiny incision to thread that tube, uh, into the inner ear.
00:03:56.340 --> 00:04:01.500
I think a lot of people would be familiar with sort of the shape of the cochlea, which is the snail-shaped organ.
00:04:01.949 --> 00:04:15.900
So as they thread this sort of floppy tube of electrodes into the entrance to the cochlea, it uh, sort of snakes into the inner ear, following this snail- shaped curvature of the outer walls of the cochlea for certain types of arrays.
00:04:16.139 --> 00:04:16.709
Incredible.
00:04:16.709 --> 00:04:22.259
And so we're talking about adults' ears. That space can't possibly be very large, right?
00:04:22.785 --> 00:04:33.194
The inner ear, uh, is composed of, of several intra-cochlear tubes, intra-cochlear cavities, uh, each of which is probably no more than about a millimeter wide.
00:04:33.644 --> 00:04:35.055
It's a very confined space.
00:04:35.324 --> 00:04:46.468
The surgery, uh, we, we call it a microsurgery because everything is, um, at a microscopic level, and the surgeon is actually looking through a surgical microscope while performing this whole procedure.
00:04:46.829 --> 00:04:49.408
And then, so there are nerves inside the ear.
00:04:49.408 --> 00:04:55.228
That is the essentially what is picking up the signal of sound and transmitting it into the brain.
00:04:55.228 --> 00:05:04.079
And somehow this array of electrodes has to be correctly placed next to or inside the nerve, but in ways that make the nerve be able to function.
00:05:04.845 --> 00:05:05.295
Right.
00:05:05.355 --> 00:05:05.654
Yeah.
00:05:05.654 --> 00:05:12.764
So in, in natural hearing, um, an acoustic wave is propagated through the ear canal, um, and then hits your eardrum.
00:05:13.004 --> 00:05:22.845
It's also what we call the tympanic membrane, and that transfers that acoustic energy, those sound vibrations, uh, to the inner ear, which is filled with a fluid.
00:05:23.144 --> 00:05:29.129
And so that fluid transmits the acoustic energy along the lumen, along the canal of the inner ear.
00:05:29.579 --> 00:05:35.459
And that vibrates, uh, small hair cells that are at the ends of, of auditory nerve fibers.
00:05:35.730 --> 00:05:42.600
And as those hair cells bend back and forth, uh, it allows the auditory nerves to be, uh, stimulated naturally.
00:05:43.110 --> 00:05:45.959
So we call this a transduction mechanism.
00:05:45.964 --> 00:05:51.853
So the inner ear is transducing acoustic energy into electrical signals that get sent to the brain.
00:05:52.245 --> 00:06:01.334
For folks who are experiencing severe to profound hearing loss, something has gone wrong in that acoustic transduction.
00:05:56.264 --> 00:06:06.464
Something's gone wrong with the ability of the body to take that acoustic signal and activate the auditory nerve fibers.
00:06:06.464 --> 00:06:24.300
So with the cochlear implant, we're really bypassing all of that acoustic part and doing direct electrical stimulation of the auditory nerve, which it sounds like it's singular when you say "the auditory nerve," but actually, uh, in a healthy ear there are about 20 to 30,000
00:06:24.689 --> 00:06:27.389
Wow.
00:06:27.389 --> 00:06:29.670
...that are, um, composing the auditory nerve.
00:06:30.225 --> 00:06:33.944
I mean, it's just so amazing as a person who likes to listen to the radio on the way into work, right?
00:06:34.274 --> 00:06:42.555
I don't think about all of the things that are happening in that moment, and when you break it down that way, it's just incredibly intricate, incredibly complex.
00:06:55.579 --> 00:07:02.625
When you are doing this work with surgeons, I'm super curious about how surgeons relate to the new technology.
00:07:02.654 --> 00:07:33.764
I mean, obviously the new technology itself is fascinating, but then you know, in addition you have both the user, the surgeon, and you have the patient or the client, um, and you know, I'm just immediately curious as to how that has changed both the life of the surgeon and the life of the person who now has a much better functioning piece of technology, um, very carefully implanted, uh, very close to their brain and, uh, assisting them in a different way than prior to that implement.
00:07:34.574 --> 00:07:34.875
Yeah.
00:07:34.875 --> 00:07:52.814
So, uh, with this augmented reality system that we're developing, we're still pretty early on, so I'm working closely with, with a handful of surgeons who are awesome clinical collaborators to try to design this system so that, uh, most surgeons will be excited to use it.
00:07:53.384 --> 00:08:21.029
We haven't tried this on a patient yet, so this, this is, this is still pretty early on, but, um, we certainly have a lot more experience in some prior and ongoing projects that aren't augmented-reality based, but rather trying to model how that electrode array is interfacing with the patient's auditory nerve to try to help, not the surgeons, but rather the clinicians who see the patients after surgery—who are professionals called audiologists.
00:08:21.269 --> 00:08:32.654
They're the experts in trying to figure out how to tune parameters on the implant processor to try to help the patients hear better, not only right after surgery, but really through the rest of their life.
00:08:32.715 --> 00:08:32.804
Mm-hmm.
00:08:33.345 --> 00:08:39.975
So that's a lifelong form of tuning and improvement in terms of getting that technology to work.
00:08:40.754 --> 00:08:48.225
And that AI piece that you are so expert at using, is really strengthening and improving that capability?
00:08:48.554 --> 00:08:49.063
It is.
00:08:49.063 --> 00:09:06.764
So these devices have hundreds, if not thousands of different settings that can be customized for individual patients, but it's a really challenging job for the audiologist to figure out which parameters to, to change and how to change them.
00:09:07.259 --> 00:09:09.360
It's a perfect problem for the computer, right?
00:09:09.870 --> 00:09:10.559
Yeah, it's...
00:09:10.559 --> 00:09:17.100
I mean, I just think about that multivariate space and how machines are so much better at solving multiple equations simultaneously.
00:09:17.610 --> 00:09:18.360
Indeed.
00:09:18.360 --> 00:09:24.450
And as I said, typically, you know, the, the surgeon can only see the surface anatomy in the operating room.
00:09:25.049 --> 00:09:34.590
They have to thread this array into the cochlea without really seeing where the electrodes end up, and that without really knowing how they're interfacing with the auditory nerve.
00:09:35.009 --> 00:09:38.850
So the, the electrodes, generally, they're not in direct contact with the auditory nerve.
00:09:38.850 --> 00:09:46.139
The auditory nerve is, uh, sort of in the middle of the cochlea, and the snail shape, uh, of the cochlea is wrapped around it.
00:09:46.500 --> 00:09:57.990
So these electrodes, they, they just basically, they use very simple, uh, stimulation pulses with the hope that they're stimulating the right parts of the nerve.
00:09:58.500 --> 00:10:02.188
And that's just how these things have operated for the last 30 years.
00:10:02.188 --> 00:10:06.884
So it's actually quite remarkable how well they do for a lot of people.
00:10:06.943 --> 00:10:10.424
But the variability in outcomes is a, is a big problem.
00:10:10.784 --> 00:10:25.813
The N-I-D-C-D, the sub branch of the National Institutes of Health, that is called the National Institute for Deafness and Communication Disorders, they estimate that only about 5% of people who could benefit from an implant actually receive one.
00:10:26.083 --> 00:10:26.563
Wow.
00:10:27.134 --> 00:10:27.554
And...
00:10:27.554 --> 00:10:32.054
So 95% of people who could benefit don't receive one.
00:10:32.294 --> 00:10:32.865
Correct.
00:10:32.924 --> 00:10:36.044
Because of expense? Or because it's so difficult to do this surgery?
00:10:36.044 --> 00:10:40.154
Or because of preference? Or what would you say is the biggest barrier?
00:10:40.484 --> 00:10:48.938
I think it's, it's a mixture of a number of those different factors, but I do think a big one is that there is a high variability in outcomes and so...
00:10:48.943 --> 00:10:52.500
Yeah, so why go through it if it isn't gonna necessarily yield benefit?
00:10:53.164 --> 00:10:53.615
Correct.
00:10:54.063 --> 00:10:54.154
Yeah.
00:10:54.154 --> 00:10:56.975
And even, you know, in the audiology community, right?
00:10:56.975 --> 00:11:05.943
So typically how patients end up with a cochlear implant is first having their, uh, hearing loss addressed with a hearing aid, right?
00:11:05.943 --> 00:11:15.693
And then once a hearing aid is no longer useful because the hearing loss has become so severe, then the cochlear implant becomes the preferred, uh, treatment.
00:11:15.754 --> 00:11:26.594
But even amongst hearing aid audiologists, this perception of variability and outcomes can lead them to not counsel their patients to seek that next step.
00:11:26.625 --> 00:11:26.924
Yeah.
00:11:27.764 --> 00:11:43.063
And when it does work, I would imagine that the individuals with, um, really successful cochlear implants are able to not only, you know, produce speech and hear speech, but truly, really engage in communication, right?
00:11:43.063 --> 00:11:49.464
That there's not, um, sort of that lag, that microsecond lag in engagement.
00:11:49.703 --> 00:11:57.804
Have you seen, um, that change over time or with this improved, um, ability to tune the machine to work more effectively?
00:11:58.105 --> 00:12:01.943
There's a number of ways that we can improve things.
00:12:02.154 --> 00:12:04.105
We started with talking about the surgery.
00:12:04.105 --> 00:12:10.198
So can we place the device better in the first place so that it better interfaces with the auditory nerve.
00:12:10.799 --> 00:12:12.240
We think we'll be able to do that.
00:12:12.240 --> 00:12:16.259
We're sort of, we have to build this system first before we can test it and, and prove it.
00:12:16.620 --> 00:12:26.309
We do have some preliminary data that, you know, is strong enough to suggest that we'll be successful, that we were able to secure this, this funding to build the system.
00:12:26.669 --> 00:12:28.740
So I'm confident that we will get there.
00:12:28.799 --> 00:12:37.860
But the, the other way is, after surgery can we help the audiologist figure out better parameters, uh, that are custom for that particular patient?
00:12:38.370 --> 00:12:48.419
And we have started, uh, clinical trials of various strategies to do that and have seen some really remarkable results for, uh, quite a few, uh, patients.
00:12:48.870 --> 00:12:52.529
Typically what we're doing is we're measuring speech recognition rates.
00:12:52.919 --> 00:12:59.323
Uh, what this means is when the patient comes and sees the audiologist, the patient will get into, uh, what's called a sound booth.
00:12:59.654 --> 00:13:17.745
So it's like a quiet room with speakers, loudspeakers that the audiologist can play a certain list of, of words or sentences, and the patient's asked to, uh, repeat back what they hear, and then the audiologist can score what percentage of words were correctly recognized.
00:13:18.149 --> 00:13:29.404
So that's one metric that we look at for can we help improve outcomes, is can we make some manipulations that will improve the number of, of words that they recognize in these speech recognition tests.
00:13:30.224 --> 00:13:32.953
Another metric is not so quantitative.
00:13:32.953 --> 00:13:34.844
It's more just the qualitative feedback and...
00:13:34.844 --> 00:13:34.934
Right.
00:13:34.934 --> 00:13:35.684
How's my life?
00:13:35.833 --> 00:13:35.984
Yeah.
00:13:35.984 --> 00:13:36.254
Right.
00:13:36.583 --> 00:13:36.764
Yeah.
00:13:36.764 --> 00:13:40.693
And, and what kind of impression that patients get from, from the changes.
00:13:41.054 --> 00:13:47.924
One thing that makes it it challenging is that the implant is just one part of the hearing equation.
00:13:48.360 --> 00:13:50.730
We're delivering signals to the auditory nerve.
00:13:50.940 --> 00:13:59.340
Those signals are being transmitted to the brain, and then the brain is learning how to hear in a completely new way, and it is very different from natural hearing.
00:13:59.340 --> 00:14:05.190
So I mentioned in natural hearing you have 20 to 30,000 different, uh, auditory nerve fibers.
00:14:05.549 --> 00:14:11.250
Those all individually correspond to very finely tuned pitches.
00:14:11.833 --> 00:14:13.875
Think of, you know, pianos on a keyboard.
00:14:14.144 --> 00:14:16.453
I don't know off the top of my head how many pianos on a keyboard?
00:14:16.784 --> 00:14:21.073
Well, your ear has something like 20 to 30,000—very finely tuned.
00:14:21.764 --> 00:14:32.865
We're replacing that with 12 to 22 electrodes, and each of these electrodes is stimulating thousands of individual auditory nerve fibers.
00:14:33.434 --> 00:14:41.894
And so you're basically going from having 20,000 unique pitches available to you, to maybe as few as 12.
00:14:42.345 --> 00:14:51.855
And it's actually even a little bit worse than that because since we're not doing direct stimulation of the auditory nerve, there's a lot of spread of excitation, and these channels are all kind of muddy.
00:14:51.855 --> 00:14:53.115
They're overlapping with each other.
00:14:53.565 --> 00:15:03.825
So imagine if in your car you tune your radio to a station, if your radio wasn't selective enough, you'd be hearing three stations at once, and everything would be really, really muddy.
00:15:03.884 --> 00:15:04.274
Right.
00:15:04.904 --> 00:15:20.490
So we had a lot to overcome to really, uh, be able to recover what, what I would call natural fidelity, but there is certainly room to be able to improve people's ability to just be connected, have conversations.
00:15:20.610 --> 00:15:23.789
Understanding speech and noise is, is one of the biggest problems.
00:15:23.879 --> 00:15:23.940
Yeah.
00:15:24.000 --> 00:15:28.200
So can you go with your friends to a noisy restaurant, have a conversation?
00:15:28.529 --> 00:15:30.750
It's a challenge for, for implant listeners.
00:15:40.664 --> 00:15:43.544
What makes a true story more compelling than fiction?
00:15:44.325 --> 00:15:55.169
In the next episode of the"Quantum Potential" podcast, I talk with documentary filmmaker and Vanderbilt Professor Mariah Kramer about the method and the magic of showcasing real life on film.
00:15:56.129 --> 00:15:58.590
But now let's get back to my conversation with Jeff.
00:16:05.399 --> 00:16:15.208
I really wanted to ask you more about this idea of digital twins, this idea that you can customize that model to each of the patients that come in.
00:16:15.568 --> 00:16:19.259
Can you say a little bit more about how, um, that methodology works?
00:16:19.500 --> 00:16:29.490
You know, I think we hear the phrase digital twin, um, increasingly as we think about the ways in which technological advances are gonna dramatically change medical care, for example.
00:16:29.490 --> 00:16:35.458
But just give us a little bit of that texture of what you mean when you say digital twins and how that works.
00:16:35.860 --> 00:16:36.269
Okay.
00:16:36.450 --> 00:16:40.679
Every patient has a unique situation in multiple ways.
00:16:41.190 --> 00:16:42.899
The shape of their anatomy is unique.
00:16:43.379 --> 00:16:55.500
The electrical properties of the tissues of their ear is unique. By the way, these folks are suffering from severe hearing loss, so oftentimes their auditory nerve may not be all that healthy.
00:16:56.129 --> 00:17:01.080
Um, and there may be some of those auditory nerve fibers that are no longer alive.
00:17:01.110 --> 00:17:04.950
We, we may not have 20,000 of them anymore to target stimulation.
00:17:04.950 --> 00:17:10.829
There may be maybe 5,000 and may, maybe the distribution is, is kind of patchy.
00:17:11.220 --> 00:17:21.269
Uh, we've seen a lot of evidence from that, from, uh, histological studies, that you may have some regions of, of nerve fibers that are healthy, and then some spots where they're not healthy.
00:17:21.345 --> 00:17:21.463
Hmm.
00:17:21.568 --> 00:17:29.279
And all of these things factor into how the electrical energy that's getting delivered by the electrode array is interfacing with the auditory nerve.
00:17:29.878 --> 00:17:41.279
So, um, one of the things that we've been working on is for an individual patient, create a digital twin that tries to capture all of these unique aspects of an individual.
00:17:41.339 --> 00:17:50.309
And of course, we can't cover everything, but we, um, we do try to cover account for, uh, as many of these variables as we can.
00:17:50.970 --> 00:18:11.279
So to capture a patient's unique anatomy, uh, what we use is a preoperative CT, uh, that's a computed tomography scan or an MRI scan, and that gives us a spatial map, a 3D volumetric map of where all of that individual's anatomical structures are located and what, what's the shape of their inner ear.
00:18:11.730 --> 00:18:14.819
It's basically a map of, of the different tissues.
00:18:15.359 --> 00:18:27.480
Postoperatively we can, or intraoperatively, we can use a, uh, a computed tomography scan to then find exactly where those, uh, electrodes ended up in surgery.
00:18:27.869 --> 00:18:40.904
And by using a process called registration—it's basically just aligning the preoperative and postoperative images together— we can see where those electrodes are relative to all of those, uh, tissues that we're interested in.
00:18:40.904 --> 00:18:45.494
So that kind of helps us do a, a spatial customization for an individual patient.
00:18:45.825 --> 00:18:51.494
And then there are a, a number of pieces of data we can collect from the patient's cochlear implant itself.
00:18:51.555 --> 00:19:08.565
We can take these electrical recordings of different types of, uh, or in response to different types of electrical stimuli, and that gives us a sense for the, uh, individual electrical properties of that patient's tissue that we can feed into our digital twin model.
00:19:08.775 --> 00:19:16.154
And we use, also use some of those electrical recordings to try to get a sense for the health of the auditory nerve as well.
00:19:16.845 --> 00:19:26.954
Um, and this is kind of a, an iterative process where we, we start with maybe some assumptions about an assumed neural health situation for the patient.
00:19:27.464 --> 00:19:38.084
We use our digital twin model to predict what measurements we should get if that was true, Then we can compare that to the actual measurements we get from the patient and then adjust...
00:19:38.115 --> 00:19:39.404
Totally fascinating.
00:19:39.404 --> 00:19:46.934
...until we, we've customized that model to produce the same or close to the same measurements that we, that we collect from the patient.
00:19:47.025 --> 00:20:01.394
Yeah, I mean, it never, it never occurred to me that the implanted technology could be evaluative and sending you information in an ongoing way, not simply stimulative or in some way transmitting, right?
00:20:01.515 --> 00:20:03.914
It's transmitting out as well as in.
00:20:04.095 --> 00:20:10.170
Thankfully um, the hardware is designed with this capability where we can collect a number of different uh, sources of data.
00:20:10.529 --> 00:20:24.299
For the most part, these different measurements have really just, uh, been used in the research community and haven't really been translated for a clinical purpose yet, um, because it's, it's difficult to really interpret them on their own.
00:20:24.900 --> 00:20:29.039
That's really where the digital twin approach has a lot of value.
00:20:29.069 --> 00:20:29.549
So cool.
00:20:29.549 --> 00:20:49.515
It's because instead of just seeing the individual measurements in isolation and not really understanding where they're coming from, in the digital twin, we're representing the individual auditory nerve fibers, and we have accounted for patient-specific variables, and we can see what's leading to those measurements and, and explain them.
00:20:50.355 --> 00:21:02.819
Does it take hours, days or weeks to make that digital twin model run at the level that you feel confident that it's a fairly accurate ... its, its predictions are relatively close to reality.
00:21:04.814 --> 00:21:05.564
Yes.
00:21:05.684 --> 00:21:07.424
Uh, so, um,
00:21:08.250 --> 00:21:08.709
All of those?
00:21:09.134 --> 00:21:25.664
Initial, uh, yes, our, our initial software that we have, have, have built to create the digital twin models, uh, leverages the, uh, computing cluster here and, uh, can take, uh, a couple of weeks to be able to construct the model.
00:21:26.054 --> 00:21:37.964
But we're making a lot, a lot of progress and speeding that up to where we hope we'll be able to actually do it within a few minutes in the same visit where we're collecting the electrical measurement from the patient.
00:21:37.964 --> 00:21:38.023
Yeah.
00:21:38.023 --> 00:21:42.375
If your provost would spend more on, on computing, maybe, uh, it would speed things up a bit.
00:21:43.154 --> 00:21:52.394
Well, yeah, the computing power is part of the problem, um, but also, you know, we can, we can improve the methods, as well, so they don't require quite so much computing power.
00:21:52.424 --> 00:21:53.594
No, it makes total sense.
00:21:54.375 --> 00:21:57.674
I'm really interested in the way in which you have 12 patents.
00:21:57.674 --> 00:21:58.365
Is that right?
00:21:58.784 --> 00:21:59.444
I think so.
00:21:59.444 --> 00:22:01.349
I, I don't, I don't keep a running total in my head.
00:22:01.680 --> 00:22:08.250
You know, last I checked your CV, you had 12 patents and a number of licenses for your software.
00:22:08.670 --> 00:22:42.154
And I'm just fascinated, um, by the way in which you know how to leverage the market, essentially, um, you know, how to partner with corporate partners, um, to be able to make this incredibly innovative work more impactful—that it's not gonna sit in a journal article for three to five to seven years, but that it's moving out into the world, not only in the way that you're training surgeons, but prospectively in the way that the technology is being designed and built and then, you know, marketed and, and, uh, manufactured so that many, many more people have access to higher quality technology.
00:22:42.184 --> 00:22:43.534
Can you say a little bit about that?
00:22:44.105 --> 00:22:50.314
Yeah, it's, it's a very, uh, complicated question because there, there's so many different variables that are involved.
00:22:51.075 --> 00:22:53.744
One of them is market forces.
00:22:54.315 --> 00:23:07.424
So these devices are basically, the money that the implant manufacturers make on these devices is really tied to, uh, insurance reimbursement rates, which are, are kind of fixed.
00:23:07.785 --> 00:23:18.839
And so, uh, the companies have a difficult time putting too much investment in new technologies that may still have some risk.
00:23:19.289 --> 00:23:25.589
And so this is, this is exactly why the NIH is such a wonderful thing ...
00:23:25.589 --> 00:23:37.244
because, uh, the NIH will provide funding for, uh, earlier-stage research that may not be completely proven, but they're fine with taking risks that companies won't.
00:23:37.454 --> 00:23:46.575
So they're essentially willing to place a bet on a really bold idea, where a company doesn't have the enough financial incentive to place that same bet.
00:23:46.994 --> 00:23:47.535
Correct.
00:23:47.865 --> 00:23:53.730
It's also a little bit different than, let's say, the market for a cell phone, right?
00:23:53.730 --> 00:23:57.930
So you could go to a store, they present you with a dozen different phones.
00:23:57.930 --> 00:24:09.269
You can try them, see what you like and then make a choice. With an implantable medical device, you don't really get to try out different devices and see which one you like the best.
00:24:09.269 --> 00:24:17.160
You kind of have to trust what you're reading online, which of course every company is gonna have data that says that their device is the best.
00:24:17.160 --> 00:24:18.329
So who do you believe?
00:24:18.839 --> 00:24:34.859
And so unfortunately, it's just a situation where, you know, for these companies, where they see the most profit is in, uh, sort of unique features, uh, like the ability to have Bluetooth connection or, uh, have a waterproof device that you could swim with.
00:24:35.220 --> 00:24:49.244
And so it's, it's just difficult for them to put a lot of money to invest in, in things that improve the operation of the device and improve, uh, speech recognition reach, which is, I think largely why these devices haven't changed much in the, in the past few decades.
00:24:49.394 --> 00:24:50.744
Totally fascinating.
00:24:50.954 --> 00:25:01.994
So another approach is with NIH funding, we, we can, we can develop the technology to a point where we, we prove it's working well, and it makes the risk much less to invest in.
00:25:02.565 --> 00:25:17.444
Um, but also it gives us an opportunity to build systems that can be just open source and freely available, and then we don't, we can just make it freely available to the end user, and then the companies don't really have to make an agreement on it, right?
00:25:17.534 --> 00:25:22.574
So there, there's sort of multiple avenues to try to get the technology, uh, out there.
00:25:22.964 --> 00:25:34.394
It sounds like get the technology out there at a lower price and with high enough usability that actually it drives people's faith and-or trust in the technology, right?
00:25:34.664 --> 00:25:36.494
That's the long-term goal, yes.
00:25:36.494 --> 00:25:36.763
Goal, yeah.
00:25:36.944 --> 00:25:49.094
I kind of like to think of, of of an end goal being, can I help a cochlear implant user hear well enough to, for example, serve as a composer for a symphony.
00:25:49.694 --> 00:25:52.844
If I can get to that point, I, I'll feel like I've accomplished a lot,
00:25:52.844 --> 00:25:53.654
An amazing goal.
00:25:53.970 --> 00:25:54.480
It's great.
00:26:10.589 --> 00:26:11.700
And how about your students?
00:26:11.700 --> 00:26:25.694
I mean, it's so cool to see that just as you were a research assistant, as an undergrad, you now regularly win awards, um, for supervising undergrads and make sure that they get funding for their summer research project in your lab.
00:26:25.994 --> 00:26:29.549
Can you tell us a little bit about what it's like to pass it forward in that way?
00:26:29.549 --> 00:26:54.630
Of course, for me, as your provost, it makes me super proud and, you know, really happy because I want undergrads to be excited and inspired to do research, but also I'm just really grateful to you for mentoring students and for taking the time and really committing that kind of effort to helping them pursue their dreams, whether it's to support, um, someone with hearing loss to be a composer or some other future awesome dream.
00:26:55.034 --> 00:27:03.105
It's really a privileged, uh, position to be in, to get to interact with, uh, the amazing students that we have here at Vanderbilt.
00:27:03.105 --> 00:27:05.115
They're all, uh, incredible.
00:27:05.115 --> 00:27:09.194
By the way, I don't think I could get into Vanderbilt now if I applied.
00:27:09.824 --> 00:27:12.914
No. You're supposed to say that you're not sure whether your roommate could have gotten in.
00:27:14.474 --> 00:27:21.750
But yeah, they're, they're all so, so brilliant and, um, the difficult part often is, um, keeping up with them, right?
00:27:21.750 --> 00:27:29.460
They're, they're so amazing and passionate and I feel like, you know, I get to learn more from them than they learn from me sometimes.
00:27:30.000 --> 00:27:36.825
In terms of learning curves, I think a lot about, um, your commitment to collaborating with people in very different fields.
00:27:37.035 --> 00:27:44.085
That's so true to the ethos that is, um, Vanderbilt's kind of foundational value in terms of radical collaboration.
00:27:44.414 --> 00:27:51.585
Tell me a little bit about what it's like to collaborate with, uh, physician scientists in the medical center at Vanderbilt.
00:27:51.914 --> 00:28:03.464
I see that you have all these co-investigators who do really fascinating work, but very much in a different field. And, you know, what do they think of you as an electrical and computing engineer.
00:28:03.464 --> 00:28:06.854
And how does it work, um, in terms of you understanding what they do?
00:28:07.273 --> 00:28:13.214
It's such a great opportunity here that we have, uh, when we're trying to innovate in the medical fields.
00:28:13.244 --> 00:28:41.128
So, uh, we have an institute that the acronym is VISE, it stands for the Vanderbilt Institute for Surgery and Engineering, um, and really it's a collective of more than 20 labs across the School of Engineering here at Vanderbilt and, uh, Vanderbilt University Medical Center with a shared research space that's provided, uh, within the medical center, housed within the medical center where, uh, the labs can work together on various projects.
00:28:41.460 --> 00:28:52.169
And there's, uh, for example, a mock O.R., operating room, where we can do experiments with novel image-guided procedures, which is just a, a fantastic resource.
00:28:52.169 --> 00:28:56.220
And there's not really many places like it anywhere else in the world...
00:28:56.279 --> 00:28:56.460
Yeah.
00:28:56.878 --> 00:29:07.829
...where, you know, I'm a five- minute walk from my office in the School of Engineering to, uh, the operating rooms where, uh, these implants are, are actually placed.
00:29:08.339 --> 00:29:33.869
And I honestly, I, I feel like Vanderbilt and VUMC have both really had a concerted effort over the last few decades to hire people and bring in people that are, uh, very collaborative, have a very collaborative attitude and want to work together to try to solve these important clinical problems without a lot of, you know, internal competitiveness, right?
00:29:33.869 --> 00:29:37.920
We, we, we wanna work together and, and create solutions to these problems.
00:29:38.250 --> 00:29:53.055
And so just being able to bring together in one place the faculty and students that we have in engineering who have the expertise to come up with these technical solutions to problems with the clinicians who have the pain points.
00:29:53.085 --> 00:29:59.444
They know what their pain points are in treating their patients, but may not have the expertise to know how to fix those problems.
00:30:00.509 --> 00:30:24.329
And just putting people in the same room, making sure that we have a lot of interactions with each other so that the folks who know how to create the solutions can learn what the problems are and work with the folks who are, uh, on the leading edge, you know, the, the clinicians like the audiologists and the surgeons, to try to come up with ways to improve outcomes for patients and clinical workflows of, of various types of interventions.
00:30:24.419 --> 00:30:26.699
Yeah, it's an amazing set of collaborations.
00:30:26.699 --> 00:30:29.549
It's really, really impressive.
00:30:29.669 --> 00:30:39.929
And, you know, kudos to you and to your collaborators for seeing the strengths in each other's, um, expertise rather than feeling like, "Well, that person just doesn't speak my language.
00:30:39.929 --> 00:30:49.619
I don't know, you know, how to relate." It's, it's tremendous in terms of sort of a spirit of curiosity and open-mindedness, but also commitment, shared commitment to problem-solving.
00:30:50.460 --> 00:30:54.809
I'm really interested in, um, what makes you hopeful as you think about the future.
00:30:55.200 --> 00:31:03.419
You've described so many of the incredible challenges, Jack. What gives you hope as you think about the next three to five to 10 years in your own field?
00:31:04.650 --> 00:31:14.880
So I think, in general, I'm really encouraged by this trend towards, uh, developing open-source, freely available solutions.
00:31:14.880 --> 00:31:19.109
And the boom in AI is, is really helping to do that.
00:31:19.710 --> 00:31:34.919
So I really hope that a, a lot of these, a lot of these challenges that often, you know, government funding goes to projects that do the work, complete the work, but maybe it never leaves the bench to make it to the bedside and get, get translated.
00:31:35.189 --> 00:31:52.824
And often that's not for a technical reason, it's, it's more because there's just, uh, such a, a long sequence of events that has to happen to cross the regulatory hurdles or put the right people in the room to create that startup company that's actually going to market the product.
00:31:53.564 --> 00:32:04.125
But with AI, I think we're gonna more and more see a lot of these solutions that are just developed and put up freely available where anyone can use them.
00:32:05.099 --> 00:32:09.029
And I think that's gonna open the door to a lot of new advancements.
00:32:09.299 --> 00:32:19.170
So cool to think not only about acceleration in terms of discovery, but access—and widening access might also lead to acceleration.
00:32:19.170 --> 00:32:21.329
That's, it is really hopeful.
00:32:21.388 --> 00:32:25.888
Anything else that you'd love, um, our audience to know or anything you'd like to leave us with?
00:32:26.789 --> 00:32:32.549
I just feel really, uh, fortunate to have been sort of put in a position here to be able to do this research.
00:32:33.179 --> 00:32:38.819
There is a, a quote that I like to think about when I think about the work that we're doing.
00:32:39.329 --> 00:32:44.159
Uh, it's a quote from, uh, Helen Keller who said, "Blindness separates you from things.
00:32:44.189 --> 00:32:59.398
Deafness separates you from people." And, uh, it's just really, uh, fun and and rewarding to be able to do research in an area where I think we're able to do things that, that help people stay connected and communicate with each other.
00:32:59.759 --> 00:33:00.509
I love that.
00:33:00.569 --> 00:33:01.859
Great way to close.
00:33:01.859 --> 00:33:02.865
Thank you so much.
00:33:04.845 --> 00:33:19.184
Jack, thank you so much for opening our eyes or really opening our minds and our ears to a world most of us don't really fully understand, and certainly a part of the body, I think in terms of anatomy, that we really can't see in terms of the inner ear.
00:33:19.784 --> 00:33:29.785
It's inspiring to see Vanderbilt researchers pushing the boundaries of both medicine and our imagination, frankly. We can't wait to watch what you and your collaborators do next.
00:33:29.785 --> 00:33:32.424
And I'm thrilled to have had you, um, join us today.
00:33:32.424 --> 00:33:33.174
Thanks so much.
00:33:33.474 --> 00:33:33.924
Thank you.
00:33:33.924 --> 00:33:34.525
It's been an honor.
00:33:41.424 --> 00:33:46.974
Thank you to our guest, Jack Noble, who's the epitome of the trans-institutional researcher here at Vanderbilt.
00:33:47.575 --> 00:33:55.769
Jack is an assistant professor of electrical and computer engineering in the School of Engineering and a faculty fellow at the Vanderbilt Institute for Surgery and Engineering.
00:33:56.429 --> 00:34:12.480
Jack also holds an impressive four secondary appointments, including biomedical engineering within the School of Engineering, computer science within the new College of Connected Computing, and two appointments within the School of Medicine, including hearing and speech sciences and head and neck surgery.
00:34:19.409 --> 00:34:22.530
The "Quantum Potential" podcast comes from Vanderbilt University.
00:34:23.130 --> 00:34:28.679
I'm your host, Cybele Raver, provost and vice chancellor for academic affairs with Vanderbilt University.
00:34:29.730 --> 00:34:33.960
This episode was edited and mixed by the team at Sound On Sound Off.
00:34:34.019 --> 00:34:37.980
You can learn more about their work at Sound On Sound Off dot com.
00:34:38.804 --> 00:34:44.114
For more information about"Quantum Potential," go to vanderbilt.edu/quantum potential.
00:34:45.224 --> 00:34:48.255
Thank you for joining us today for the "Quantum Potential" podcast.
00:34:48.945 --> 00:34:54.135
Stay tuned for more conversations with the thinkers and innovators reshaping our world.
00:34:54.824 --> 00:34:58.784
I'm Cybele Raver, and this has been "Quantum Potential."