Vocal biomarker tech startup Sonde Health has debuted a tool for someone’s mental health based on the sound of their voice. The company packaged the new Sonde Mental Fitness feature into an app for both iOS and Android and an API that healthcare providers and wellness services can incorporate into their products.  David Liu, CEO of Sonde Health, breaks down how the technology works and the other use cases for vocal biomarkers. 

Opening Bell

Sonde's Vocal Biomarker Technology Can Detect Health Issues 

here to discuss is David liu the ceo of sand health. Thank you so much for being with us today. I want to start off by talking about how does this technology to detect health conditions sounds so fascinating. It is well, we're based in boston and vocal biomarkers is a segment of a broader health category called biomarkers. And bio digital biomarkers have been around for many years and a lot of the things that you've been able to use through your Apple watch and detect through your Apple watch are part of this digital biomarker space. And vocal biomarkers is a large part of that category. And so anything that we do, we basically measure the vocal features that are coming from your voice. This is not natural language processing. This is absolutely analyzing the audio signal processing coming from your voice. And then we break down more than 4000 features and then analyze and and really target the features that give us indications that are tied to health conditions and disease. So can you break down how this technology works? Is that an app is is something you have to go to the doctor for you speak into the phone and how will college campuses use it to detect early signs of COVID? Well, we're really, you know, we did some work in the Covid space but we're really not focused in that area. We're focused in areas of asthma, other respiratory disease COPD. We do also give you indications uh, that were in that were also so analyzing indications for depression and anxiety and mental health. Um but in Covid we're not really focused in that area. So what we do is we were working through our platform and through our Api and this technology can live on any app, Not just our own. We have an app that you can download to the app store just to get a taste of what vocal biomarkers is like. And so you would give a voice sample very simply for respiratory disease. We'd be able to take a couple of 62nd voice samples, recordings. And then from there we would give you a score based upon your risk of having a respiratory disease through these symptoms that we analyze again through vocal features such as um timber or intensity. We do look at a lot of different features and then based upon the data that we have and we have the world's largest that we know of health labeled voice recording data set in the world based upon the the that data set and based upon what we've seen uh where doctors have told us that this patient has the disease and then we record that patient and then we compared against a control group. So let's talk about that right there because it's interesting that you have these vocal biomarkers. How do you think that it will really change the daily health monitoring that we're seeing happening more and more on public places. I'll tell you this is um this is the biggest, I would say area of innovation. You take a look at all the health trends in this multi trillion dollar industry that we're talking about just in the US alone, but globally remote patient monitoring and understanding what is actually happening when you're not in the doctor's office is quite possibly the biggest trend that we'll see in the next 20 years or so. And so what that means is that we can now thanks to where Parables, where it used to be about being able to collect longitudinal data through a device, right? Whether it be a wristwatch or or ring or anything else. So that's the first key with a voice. We are always speaking, we're always talking into a void bap. We're talking into a phone, you're you're able to be collecting voice 24 hours a day essentially with the user's permission. Right? And so with that longitudinal data, day after day data, we're able to now understand what is happening in those changes in your health to be able to then alert patients and users as to what is actually happening? That is the most critical public. Well, let's talk about the mental health side of things too. Because how does it determine the base level of one's mental health? Like maybe I'm just really tired and I sound like really down and it says I'm depressed. I'm like, no, I just need some sleep. So how accurate is it wouldn't that matter? Well, it's incredibly accurate when we're talking about measuring vocal features scores, we're not diagnosing for depression or anxiety. Let me be clear about that. You need a professional uh inclination to be able to do that. But what we are telling you is in the areas of smoothness with the technical term is jitter or control, local control. Liveliness, energy range, clarity, christmas. These are examples of local features that we are analyzing from your voice. So it at that moment in time your voice is telling us just like resting heart rate. Could be telling your apple watch that you may have a high resting heart rate. It doesn't mean that you have coronary artery disease but where the same thing with your voice? We're going to give you a readout as to what your voice is telling us. And in those local features that I just mentioned, studies have shown that those vocal features are tied to depression, anxiety and even mild cognitive impairment. And so we're beginning to look at areas of my M. C. I. Because that's a precursor for dementia and Alzheimer's disease. So these areas are all very very interesting because we can pull data and a lot of information from voice which is the rich source of health information. So how would you say that vocal biomarkers are used to detect other health conditions? What are some of them And what do you hope to see in the future when it comes to detecting other things that may exist in elements and diseases as well. Well think of this as an early warning system. And so while we're not a diagnostic and it's not meant to be a diagnostic. It is one of the things that we know that that health providers and health looking for, which is uh longitudinal advanced data and information on patients over time. That way when they have this information, they can act, they can get people to where they need to be and treat them much more effectively, much more proactively. And that really drives on the cost of health care. That makes everything much much better. So where I can see this going and we're already in the respiratory space with asthma COPD build diseases. We can give you an indication of symptoms that are beginning to show up earlier than if you were to then have to go into the doctor and imagine most people go in when you're you just can't stand it anymore and then you have to make that appointment right. And in mental health and areas of mental or mild cognitive impairment dementia. These are also areas that are very difficult to detect on your own or even through a provider in uh in in most cases. So what you want to be able to do is begin collecting this information early and people know what to do if the general public now has been conditioned to understand how wearables work, how to be able to have longitudinal data be able to interact with that information and then act upon that information when necessary. And so mental health is a huge area for us because we believe that not only are we in a crisis post pandemic, we believe that this will go on for a long time and there are segments of the population, not only in this country but around the world that are suffering from depression and anxiety and if you can get these people much sooner the results are going to be much much better. Well you know a lot of people these days are concerned about privacy and you know facebook just got in trouble for something like that. So does this tech maybe infringe on the user's privacy and how was that data store? Was the data used for things like that? That's a great question. And one of our key pillars for the company is to me to make sure that we are incredibly careful and respectful of people's data privacy when it comes to health data. And first of all, none of this data is being collected unless the user has given permission to our customers apps to us to be able to collect the data and analyze it for them. The data that we're collecting then can also be completely deleted. If that the jews are choosing to do so. So the user or the patient is in complete control, the provider is in complete control of this of this information and so it is stored in the cloud. It is being analyzed in the cloud and then not to. In the distant future, we will be doing this all on locally on the device on the edge. And this is why we have a relationship with Qualcomm and other chip set makers because we want to be able to record and analyze your voice right on the device and then it will be deleted right on the device. Wow, I cannot wait to see what the future of this tech goes and you know what's in store for it. So thank you David lucy of Sound Health.