How well can infants distinguish unfamiliar voices?

Infants can distinguish caregivers' and other familiar voices early in life, and can even tell the difference between two unfamiliar female voices. In this episode, we talk to Madeleine Yu (University of Toronto) about her research into infants' ability to distinguish the voices of unfamiliar male speakers.


Associated paper: Madeleine E. Yu, Natalie Fecher, and Elizabeth K. Johnson. "Learning to identify talkers: Do 4.5-month-old infants distinguish between unfamiliar males?" JASA Express Letters  4, 015203 (2024) https://doi.org/10.1121/10.0024271.

Read more from JASA Express Letters. 

Learn more about Acoustical Society of America Publications 

 Music: Min 2019 by minwbu from Pixabay. https://pixabay.com/?utm_source=link-attribution&utm_medium=referral&utm_campaign=music&utm_content=1022

Transcript

Kat Setzer  00:06

Welcome to Across Acoustics, the official podcast of the Acoustical Society of America's publications office. On this podcast, we will highlight research from our four publications. I'm your host, Kat Setzer, editorial associate for the ASA. Today we're talking with Madeleine Yu, who recently published the article "Do 4.5-month-old infants distinguish between unfamiliar males?" in the January 2024 issue of JASA Express Letters. The article has also been featured in an AIP Publishing Scilight. Madeleine, thanks for taking the time to speak with me today. How are you? 

Madeleine Yu  00:41

Good, how are you? 

Kat Setzer  00:43

Good. So first, tell us a bit about your research background. 

Madeleine Yu  00:46

Sure. So I'm a PhD candidate, working under the supervision of Dr. Elizabeth Johnson, who was one of the coauthors on this paper. My work mainly focuses on investigating human voice recognition. And so specifically, what that means is I research what factors affect our ability to accurately recognize some voices over others, and whether that is maybe the language someone is speaking or if the speaker has an unfamiliar accent, or if the speaker is male or female. And I'm also examining how voice recognition differs between adults, children, and infants-- so a whole wide range of ages. 

Kat Setzer  01:24

Very interesting, a lot of variety there sounds like. So I've heard a bit about this idea that babies learn to recognize their parents' voices in the womb. But your article mentions that voice recognition actually takes a while to develop. Can you give us some background about what's known about the development of voice recognition in children? 

Madeleine Yu  01:41

Yeah, so it's known that babies are actually quite good at recognizing their own mother's voice right from the start, and soon after they recognize other caregivers' voices as well, such as their father, however much less is actually known when it comes to unfamiliar voices. And work with older children indicates that this ability to accurately recognize a newly encountered voice is actually not yet adult-like until around early adolescence. And we know that other factors beyond how familiar the speaker is to the listener also affects the listener's ability to recognize a voice, such as language or even accent that the speaker possesses. So in fact, very little is actually known about the development of general voice recognition skills.

Kat Setzer  02:25

 Okay, interesting. Got it. So how does the language which a person is speaking affect a child's ability to recognize the voice, and is that similar to or different from adults' experience? 

Madeleine Yu  02:36

Quite a bit of research actually shows that both adults and children are much better at recognizing a voice when they're speaking in a language that's familiar to the listener. And this is known in the literature as the language-familiarity effect. And it's also been reliably replicated with many different languages and listener populations beyond English. And it's also another area of research that I am really interested in researching. 

Kat Setzer  03:01

So four-and-a-half months is really young. Why did you choose to work with kids this age?

Madeleine Yu  03:06

 Yeah, four-and-a-half months is really young. And that's perhaps one of the reasons why we were so interested in the first place. So work in our lab has shown that infants at this age already find it much easier to tell apart voices when they're speaking in a familiar rather than unfamiliar language; they show this language-familiarity effect. And this is super interesting, given that they're quite limited in their language experience at this age. And so given this, we're quite interested in seeing how maybe other factors might affect infants' ability to distinguish voices.

Kat Setzer  03:40

So why were you interested in infants' abilities to differentiate between unfamiliar male voices?

Madeleine Yu  03:45

Yeah, so as I kind of mentioned previously, so there's a lot of work that has examined how well infants can recognize familiar voices, such as mom and dad. And actually a study from our lab in 2019, by Fecher and Johnson, indicated that infants at this age are able to tell apart unfamiliar female voices. However, there's very little work to date on how well infants tell apart unfamiliar male voices. And so even the work that has been done, either doesn't specifically examine this question, or they only report data for just a few infants, which may not be a representative sample. So this question is interesting to us both on a theoretical and practical level. So not only do female and male voices differ in terms of the acoustic cues between them, but many infants, including the ones in our study, tend to receive much more input from female voices in their daily lives versus male voices.

Kat Setzer  04:42

So how did you go about assessing infants' abilities to differentiate between unfamiliar male voices? 

Madeleine Yu  04:47

Well, when we do this kind of work with adults, typically, you can ask the adults these questions directly, but obviously, infants at this age can't do this. So we have to be creative. And we have to set up tasks that allow us to infer  what they perceive just from their behavior alone. So here we use a standard looking time paradigm, in which we have infants seated on their caregivers' lap inside a sound booth. And there they're looking at a screen that features an animated image, which is accompanied by speech sample recordings by a male speaker. Right outside the booth, with the help of a camera, we are able to monitor when and how long the infant looks at the screen. And so basically, in the first stage of the task, we continue to play speech samples from the same male speaker over and over again, until the infant habituates, or basically gets bored of hearing the same speaker. And when their looking behavior -- we expect them to stop looking at the screen so much and maybe look around the booth. Once we notice, or we see that they're looking less and less, the test phase begins. And here some of the trials actually feature a second male voice that the infant has not heard before. And so crucially, if the infant is able to detect this change in the voices, then we expect their looking time or looking behavior to change when they hear this second voice. And so we tested a total of 48 infants on two different pairs of male voices using this task. 

Kat Setzer  06:16

Okay, so you first assessed how attentive infants were to male voices versus female. Why was this important? And what did you find?

Madeleine Yu  06:24

Oh, yeah, so this was important, because you want to see whether the results that we got were due to a difference in how much infants maybe just attended to male versus female voices. So maybe infants here simply did not find the male voices interesting to listen to. And the past research even has suggested this, that infants do pay more attention to female voices. However, we found that here, this wasn't the case. So infants use the same amount of trials to habituate to the male voice as the infants in the 2019 study, that they use to habituate to female voices. And we also found that the total average looking time during the first part of the task, so before the test phase, that looking time did not differ between the male and female voices. And so this suggests that infants find both male and female voices equally as engaging. 

Kat Setzer  07:15

So can four-and-a-half month old infants distinguish between unfamiliar male voices?

Madeleine Yu  07:21

Our studies suggest that no, they cannot. Four month olds here actually did not reliably detect a change between the male voices. Meanwhile, compared to our 2019 study, where infants did detect the change with female voices on the same task, so that was really interesting. 

Kat Setzer  07:38

Why do you suppose infants can distinguish female talkers more easily than they can distinguish male talkers? 

Madeleine Yu  07:43

We have a couple of ideas. And so first of all, we find it unlikely that infants just have weak voice recognition skills at this age. So in the 2019 Fecher and Johnson study, infants who were drawn from the same population were equally as good at telling apart two different pairs of female voices, but here they're equally as bad at telling apart two different pairs of male voices. And so we speculate maybe the results here have something to do with an increased sensitivity to female rather than male voices. And why this is is not clear, but we suspect that it could be related to the infant's own home environments. So we know that infants improve in their ability to tell apart routinely encountered stimuli. We see this with how they learn language and perceive music, as well as recognize faces. And this is called perceptual narrowing. And so in our study, even out of the 46, out of the 48 families who responded, all of them reported that the infant in fact had a female primary caregiver. And research as a sort of mentioned earlier indicates that North American infants today receive far more speech input in everyday life from female speakers than male speakers. So given this, we suspect that maybe the infants in our study had an improved ability to distinguish between female talkers more easily than they were able to distinguish between the less-encountered male voices. So as I mentioned before, this is also just speculation at this point. And of course, you need to collect more data too to see if this is supported.

Kat Setzer  09:15

Okay. Yeah. So it'd be interesting to see if you could study children with male primary caregivers. 

Madeleine Yu  09:22

Exactly.

Kat Setzer  09:23

Where do you see this research going next? 

Madeleine Yu  09:24

So, yeah, as you just mentioned, perhaps extending the testing to infants who have much more exposure to male voices early in life. And also given these differences that we see in how they, just between the two studies, the 2019 and the current study, that makes us think carefully about how we should approach research with infants from here on out, especially if they involve speech samples from different types of speakers. And our results also bring us back to the old nature-versus-nurture debate in understanding human development. So yeah, What else might matter in shaping voice processing abilities or other development? Is it how much is due to the infant's home environment or are there other factors in play? So there's so much that we can investigate from here. 

Kat Setzer  10:13

Yeah, it sounds like there's, you know, a lot of opportunity out there. Well, thank you again for taking the time to speak with me today. It's really interesting to think about how early perception is shaped and how young children start recognizing and differentiating voices. I wish you the best of luck in your future research and have a great day.

Madeleine Yu  10:32

Thank you.

Kat Setzer  10:35

Thank you for tuning into Across Acoustics. If you'd like to hear more interviews from our authors about their research, please subscribe and find us on your preferred podcast platform.