“Individual differences in the acoustic properties of human skulls” The Journal of the Acoustical Society of America 146, JASA-EL Special Section 191 (2019); https://doi.org/10.1121/1.5124321
Authors: Michael S. Gordon, Michael D. Hall, Jeremy Gaston, Ashley Foots, and Jitwipar Suwangbutra
Published in the September 2019 issue of The Journal of the Acoustical Society of America in its JASA Express Letters special section. Please note that as of January 1, 2021, JASA Express Letters has re-launched as an independent online-only gold-open access journal.
In this episode, we speak with co-authors Dr. Michael Hall, Professor in the Department of Psychology at James Madison University, and Dr. Jeremy Gaston, Division Chief at DEVCOM-Army Research Laboratory. We will discuss their research on the acoustic properties of skulls and how they might affect hearing. The authors will explain their method of using broadband noise projected through the skull, and then spectrally analyzed using a Fast Fourier Transform and in 1/3-octave bands, to compare acoustic patterns in each skull. We will delve into their methods, procedures, and future applications of this research.
Read more from JASA Express Letters.
Acoustical Society of America Publications page
Music: Min 2019 by minwbu from Pixabay. https://pixabay.com/?utm_source=link-attribution&utm_medium=referral&utm_campaign=music&utm_content=1022
Welcome to Across Acoustics, the official podcast of the Acoustical Society of Americans publications office. On this podcast, we will highlight office research from our four publications. The Journal of the Acoustical Society of America, also known as JASA. JASA Express Letters, Proceedings of Meetings on Acoustics, also known as POMA and Acoustics Today, I'm your host Malene', Walters. Publications, Business Manager at ASA. Today I will be speaking with the co-authors of "Individual Differences in the Acoustic Properties of Human Skulls”, published in the 2019 issue of the Journal of the Acoustical Society of America, in JASA Express Letters special section. Please note that has of January 1, 2021, JASA Express Letters has re-launched as the independent gold-open access journal, I would like to introduce our authors, Dr. Michael Hall, Professor in the Department of Psychology at James Madison University, and Dr. Jeremy Gaston Division chief at DEVCON Army Research Laboratory. Good afternoon, gentlemen, how are you doing today?
Good afternoon. Nice seeing you like seeing you.
Thank you for joining us.
I want to start off with your backgrounds. Dr. Hall, I'd like to start with you. Could you give us a brief synopsis of your background?
Sure. So, I did my graduate work at the State University of New York at Binghamton. The research laboratory that I was training in, the Department of Psychology, there was operated by Richard Pastore and it was a psychoacoustics laboratory that was looking at general auditory principles that could apply to both speech and music and environmental noises as well. And after I got my Ph.D. there, I did postdoctoral work at the University of Washington working in department of Speech and Hearing Sciences there, and went from there to professorship, first at University of Nevada, Las Vegas, before moving back to the east coast to Virginia, to James Madison University, and have been there ever since. The other things I can end up telling you that might be of interest would be that I'm a former president of PSI CHI, which is the International Honor Society in Psychology, and I'm currently serving as co-editor of a journal that I helped start called Auditory Perception and Cognition, which is looking to integrate information across all levels of auditory processing. And that's published by Taylor and Francis.
Oh, wow, very nice. Very interesting. And Dr. Gaston, could you give us a little bit about your background as well.
First off coincidentally. I also trained Binghamton University where I got my Ph.D. in cognitive psychology also in Dr. Pastore's lab there yeah, my concentration there was low-level auditory perception and that concentration on low-level speech perception, and graduating out of ahh, Binghamton, and I took a postdoc at the Army Research Laboratory, where I was a postdoc for about two years, where I was looking at natural, unforeseen perception. And then I became a full-time research scientist at Emory Research Laboratory. And after a few years, I became a branch chief and got the opportunity to actually supervise a number of research programs. And currently, I'm an active division Chief at our research laboratory, in my area where we have working infancy and looking at human interactions in complex systems. And I can say coincidentally, I've also kind of worked pretty hard all my life, and out of high school, I joined the Army National Guard, and I was a combat engineer for 10 years, all through undergrad and graduate school, before coming to Army research laboratory.
Oh, very nice. Now let's take a look at your research. Your current research is targeted to individual uniqueness in acoustic properties of skulls and how they might affect hearing. How did this research come about?
Well, the real kind of nod needs to end up going to the first author on this. Mike Gordon was actually first interested in the issue. He had actually presented some work at a meeting of Acoustical Society, I believe, that was interested in trying to look how the small amount of information that was being accessed through the skull instead of the normal chain of operations, might actually be impacting people's listening preferences. Something like a piece of music. They tried measuring that. But at the time, they had access to very little equipment. Even though they had access to a bone conduction microphone, he and I spent some time talking over his poster at a meeting of APCAM, which is the auditory perception, cognition and action meeting, I used to organize that meeting for several years. This is like a satellite meeting ahead of psychonomic society. And while he was describing this project to me, I said, Well, you know, I see where you found struggles with getting the measurement. And part of it was due to most of the people at the time, looking at like the single frequency at a time and measuring what the response characteristic would be to a single sinusoidal tone. And I said, Well, if you want to know the frequency response, as a result of bone conduction through the skull, why don't we use a broadband stimulus like noise, and then we'd have an idea of what this call is actually contributing across all frequencies at once. So, we started talking at that point, then we ended up connecting with Jeremy, when Jeremy got to speak with Mike became clear to Jeremy very quickly that well, wait a minute, you know, I have access to audiologists, we have equipment that could do this. And so, he came on board with the practical pieces of the puzzle, you want to pick it up from there, Jeremy?
Yeah, sure. So yeah, so coincidentally, um, you know, we, at our research laboratory at the time, had a sort of a program and looking at audio communications, and also noise mazzard. So, we did a lot of evaluations of different types of communication devices, and evaluation of different types of hearing protection devices. And so, sort of as part of my second head from doing basic research, in auditory perceptual issues was also just expertise in noise measurement. So, we had quite a bit of an expertise in making and recording bone conduction signals, in addition to sort of standard ideological measurements of human hearing. And we sort of use those for doing characterizations of sort of human profiles in, these different types of audio communications, and you know, I hear protection issues.
Oh, very nice. Now, I heard that you mentioned, there were your research kind of targeted, individual differences in bone conduction. So, in your research, instead of using cadaver skulls, you use in vivo or living samples, what was the benefit of this method,
What, to put it in simple terms, the skull is a filled object, right? And it's going to respond very differently. If you treat it like just a chamber an empty chamber, and work with cadaver skulls in that way, versus if it has everything else being present. If you really want to capture what kind of information could get through to the middle and inner ear through bone conduction, then it makes sense to end up doing it in a living human.
Okay, that makes sense. Now, you projected broadband noise through the skull and spectrally analyzed using a fast Fourier transform, and in 1/3 octave bands. This is a lot to unpack, and I want our listeners along with myself to fully understand. Can you explain what a fast Fourier transform is and explain the significance of 1/3 octave bands?
Absolutely, the idea of a Fourier transform Fourier analysis is based on the idea that you could take a complex signal that's made up of many frequencies, and you could break them down into terms that are simpler for us to describe and understand the nature of the sine wave. So, each frequency is an individual sine wave having its own frequency, amplitude and phase information, a Fourier analysis can actually decompose that complex signal into those individual frequency components. A Fourier transform is a mathematical technique that will actually try to isolate all of those within the complex signal at once. And that way, we could end up having a graphical or numeric output giving access to how intense was contribution from each frequency component. Where that's helpful for this project is then we could end up putting a broadband stimulus in there that covers the entire frequency range. And we could see then which frequencies come out as appearing to be amplified which ones appear to be reduced or attenuated. And that'll tell us like, what different things that the individual score might be doing.
Yeah, and then the third question was when we get to the third-octave band, so if we think about when we're trying to understand sort of the perceptible transformation of sound right is what we want to take into account is how the human auditory system actually parses that information. So, you know, the auditory system in a very simplistic nutshell, sort of analyze the sound in these overlapping auditory filters. And one very succinct engineering way to look at them is in terms of third-octave bands, which also map on to musical scale. And the significance of the third-octave bands is that they are good rough approximations of what the size of those auditory filters are, as you actually move up through the frequency scale. So, some of your listeners may be familiar ones are a little bit more familiar with the auditory system is that you know, one typical way of looking at the auditory filters is in terms of what the critical bandwidth is what they call it. And that sort of defines sort of the edges of a filter, that process information as you move up through frequency.
And now, I just want to talk more about your experiment. In the experiment, you tested 30 participants, 18 women, and 12 men. Could you explain your procedure for testing?
Yep. So, bear with me. Sorry. Yeah. So essentially, we took two methods for testing, right. So one is that we collected a set of subjective measures. So, this is more typical ideological measures. So, we did an air conduction threshold testing. And then we also did a complimentary bone conduction threshold testing. So, this is where you actually had a broken bone conduction transducer, and it was placed on the mastoid. And you're playing basically the tones through the system, or through the head to measure those subjective thresholds. Now, the reason you would actually do those, typically, in an audiological exam, do those two different types of measurements is it'll tell you if you have a conductive hearing loss, so there's something wrong with something with the middle ear, but conduction through the peripheral auditory system, or if it's a sensory neural hearing loss, something related to the inner ear. So, what we did is we did these typical audio electrical measures as sort of baseline subjective performance. And then we had sort of our objective measures is where we had placement of the bone conduction transducer on the back of the ear on the mastoid, that was producing the broadband noise that then was traveling through the skull. And then we had actually an experimental bone conduction microphone that had been developed in summer work with some of our partners to collect basically, the response as it traveled through the skull, from the dish, and on the forehead. So basically, with these participants, we made a number of different measurements for the objective measurements paired with those subjective measurements, that sort of provided the base data set for our follow-on analyses.
Yeah, just to add on to that, and piggyback on what Jeremy said a little bit, which is that the nature of the noise stimulus as a white noise stimulus, this is something that is going to have the same amplitude on average across all frequencies. And so, the question that we're essentially testing is like, if that you draw that as a graph is basically a flat spectrum, right? You know that the amplitude stays the same across the entire frequency range. But if you present that now through the skull, what you actually measure through the bone conduction microphone will come back as something different than that. The question is, how far is it deviating from that flat spectrum, certain frequencies are going to be reduced in or filtered out and other frequencies might come out looking augmented. So that was kind of the goal there. In addition, we also took measurements on essentially the recording of that stuff without the skull, so that you can end up finding out what the actual measurements from the equipment involved was. And we could subtract that out at the end. So, any information that was coming from just the limitations of the equipment to be able to record could be accommodated for and our outcome. What we ultimately reported then was adjusting for that. So, it was really reflecting what was coming from the skulls.
Oh, I see. Okay, that makes sense. So understandable. And now, um, after testing, what were your findings? Was there anything that surprised you? Things that you thought, Okay, this is exactly what I thought I would find out.
I was kind of in disbelief a little bit. Because, you know, if you take a measurement of decibels, we were kind of recording out of these third-octave band measurements, for example, where we can now put in kind of a relative decimal reading from each of those third-octave bands, and the changes within a third-octave band from what we started with from the white noise, we were submitting What we actually got when it was conducted through the skull was dramatically different. And not only was it dramatically different, but it also changed drastically from one person to the next. And that was kind of the most shocking point, I started graphing out the range of differences that we were seeing in our individual listeners, and people go to the paper will end up seeing this depicted there. And we're talking easily 2030 decibel ranges, and six decibels, every six decibels are doubling the pressure that's hitting your ear to give you a kind of a comparison. And so, in terms of mathematical or levels of magnitude difference, this is huge, even though it's only making up a small portion of what we hear that portion is drastically different from one of us to the next. The other thing that was kind of interesting, though, to me is that if you look at the graph as a whole, you can also end up seeing that there are certain areas that the human skull, in general, seems to end up endorsing or rejecting. And so, there's certain points in the certain frequencies, where we seem to end up boosting or amplifying the sound coming from conduction through the skull, and other places where we seem to get more kind of valleys and picking up on that information. And that's across this wide set of listeners, that was drastically different. Jeremy, you want to add to that at all?
Oh, no, I think that was a wonderful explanation.
And I'm curious. So, with your, with the varying amounts of men and women, were there any anomalies? Did you find any differences in men's skulls versus women's skulls? Or is it the size of the skull, that makes a difference?
Yeah, I'm relying on memory here at the moment without going back to my looking at individual data in front of me. But my memory for this is that Mike, our coauthor actually looked for those differences. And they weren't systematic, but we did have anomalies. And so, in the paper, there's actually one depiction there of several listeners to kind of highlight just how drastically different some of these listeners were. And I would characterize at least one or two of those individual listeners as being very unusual. And so, we were just trying to highlight the level of individual difference. So ultimately, I wouldn't attribute it to being man or woman, you know, the one thing I would end up saying there's going to change is the average size of the skull itself could change. And that could impact any resonant properties that the skull itself might have. But in terms of how it comes up across the spectrum of what's being emphasized, or de-emphasized, no unified pattern was coming from it.
Oh, okay. I see. Understood? And what are the implications of the findings as they relate to perception? Could you go a little bit into that for us?
Yeah, so the big thing in connection to the threshold data was that essentially what we were observing in terms of what was measured, going in and coming out, that that pattern was predictive of the nature of the bone conduction thresholds, but not the air conduction thresholds. And this makes sense, given the nature of what it is we're measuring in the first place. And given that it's a relatively small part of the bigger picture of how the person is ultimately going to perceive things. What we were interested in launching two at the time, that became like kind of a whole second study that we initially looked at, and that we are now just starting to revisit, is trying to actually use the pattern of data that we got to be able to generate a series of filtered productions of sound for people to listen to, and see, are there differences in listening preferences for complex things like an excerpt of a piece of music. And we found some very broad trends to suggest that if we just played them, just the bone-conducted version, and ignored everything else. That's really not fair, because that's exaggerating the contribution of that part. There were some general trends towards people actually preferring the sound of things that were not like the things that would come through their own skull, is that think of it your skulls already may be emphasizing certain frequencies, we don't need to end up like then boosting that more and give it to them for listening purposes. Right. And so, we've actually been talking recently about how we could end up doing that in a more kind of full-fledged fair test. And we're expecting to do that soon. We actually have some software to essentially do third-octave man filtering of the signal. To mix with the original and be able to end up testing this out. So, I hope to end up having like a more complete perceptual answer for you very soon in terms of what this is good for in terms of our own individual listening. But for now, we know at least it predicts one of the sets of thresholds that Jeremy was describing.
Right, and you kind of delved into my next question has four ongoing steps. Is that the only area that you're taking the research in? Or are there any other upcoming or ongoing next steps for the research that you'd like to share? Or that you can share with our listeners?
Yeah, so in talking to Mike, what our first step is going to be is to essentially generate a series of filtered recordings, based on the data set that we have turned up having a set of examples of what it would sound like, conducted through those types of skulls to play back to people so they could hear what the end product would really be like. And then have them judge those things relative to each other to tell us whether perceptually this is something we have to really be concerned about. We're talking typically in bone conductance signals through the skull, I think it's somewhere on the order of like five to 10% of what we ultimately get is coming from that, which is relatively small contribution. But if these differences are so drastic, then it's possible that that could still have perceptual meaning to us. Right? And if so, the implications of that are huge, then I would think that the next step, if that comes out supporting what we think could be happening, then the next step would be to think in terms of like, all the real world applications where this could end up having some impact, you know, you start to think about like, should we individually differentiate our listening devices to be able to end up handling? What's coming from our own skull, but that won't be necessary unless we can demonstrate very clearly that that's going to make a meaningful perceptual contribution.
I see. I'm, I'm so interested to see if that would be something if we would all have individual, you know, has he said auditory devices based on you know, how our schools are interpreting sound so that I'm so curious, in the long term to see if that pans out? And now, lastly, are there any closing details you would like to share anything else that you'd want? The listeners to know about the research?
I'm giving Jeremy, first crack on this one.
Sure. Yeah. I mean, it can add like a context thing here to kind of kind of segues in with what Michael was talking about, is so when you think about it is sort of the analog that would be obvious to folks is, you know, what your voice sounds like when we hear it over the phone versus what you sound like yourself, right? Or not now in the world of virtual teleconferencing, right, we can get a better idea that. So, in basically, the reason that you hear those differences between what you hear yourself as somebody else does is these bone conduction pathways. So, they clearly change the quality of the tambour of the of the sounds that we hear. So, in something like audio communications, you know, the perception is pretty resistant to these changes. And if you're talking about issues of intelligibility, you're probably going to be okay, but I think what Michael was really alluding to, is that when you get into things like preferences for stuff like music, right, those changes in the quality of sound are, can be potentially huge, in terms of, you know, personal preference or personal enjoyment of those types of sounds. So just add sort of the context there for listeners, maybe they have that in their mind the analog of what, really that those bone conduction pathways are adding into your audio auditory experience of the world.
And I would just add to that, that, traditionally, I would think that acoustic research and auditory research has done a very good job of paying attention to individual differences when it comes to something like a clinical focus. If we're trying to establish that somebody has a deficit in hearing or some impact of an adverse event that they need to end up dealing with, then we've been very good at quantifying that kind of stuff. But we haven't really paid attention to the impact of individual differences on kind of everyday listening activities. So, I think that this at least opens the doorway to that kind of possibility. We don't know exactly what the answer is yet, right. This is a preliminary step. But if we continue down this road, we might get to get a full answer as to like whether people should be thinking about individual differences, as it applies to all aspects of listening.
And again, as I said before, a very interesting, I'm definitely looking forward to hearing more of that. So I'd like to thank you both for joining us on the podcast. It was a pleasure speaking with both of you, meeting you and also learning more about your research.
Well, thanks for having us. Real pleasure.
Yes, thank you. Thank you so much. Thank you for tuning in to across acoustics. If you would like to hear more interviews from our authors about the research, please subscribe and find us on your preferred podcast platform.
Transcribed by https://otter.ai