The human voice is an extraordinary tool for communication and self-expression, with a range of pitches and tones at our disposal. Because of this acoustic variability, a speaker can end up sounding quite different depending on the situation (e.g. a noisy café vs. a quiet meeting room or chatting with a friend vs. presenting to a colleague). This versatility can make it difficult to recognise a person from their voice alone; you must identify who is speaking despite the variation in how they sound.

In this new study, published in Current Biology, researchers from the Division of Psychology and Language Sciences used functional magnetic resonance imaging (fMRI) to ask how the brain responds to variability in the voices of other people, taking a specific look at how familiarity affects brain response patterns. 

Previous research by the team has shown that for some voices – particularly those of people we know really well like our partners and closest relatives – we can recognise the speaker with extremely high accuracy in all sorts of situations. For other less familiar people, we can find it much more challenging.

In the present study, researchers measured and compared brain responses to voices of differing familiarity: a personally-familiar voice (friend or partner), a voice that the participant learned to recognise through lab tasks (lab), and a new voice they hadn’t learned before (new). In the MRI scanner, 26 adult participants heard different voice clips from all three identities and had to respond to each clip to indicate which person was speaking. 

The team tested the hypothesis that brain response patterns would be more similar across different clips from the same identity compared to response patterns of two different identities). Researchers also tested whether response patterns to a specific speaker might be most similar for the personally-familiar voices compared to the less familiar voices (lab and new voices).

Across their analyses, the researchers found that brain response patterns to clips of the personally-familiar voice were more distinct from responses to the other two voice identities, but were also more distinct from each other. In other words, different clips from the same familiar voice identity generated the greatest difference in brain responses. Although this is counter to the hypothesis, the researchers explain that this shows that, for familiar voices, the brain is doing  better, more detailed job at encoding speech, allowing you to understand it more easily.

These results suggest that familiar voices are not just easier to recognise, but that our brains work harder and more precisely to process them, and the more familiar a voice is to us, the more sensitive our brains are to the unique features, helping us understand the people we know in different environments.

Lead author, Professor Carolyn McGettigan said: “This study adds to our understanding that personally-familiar voices are engaging our brains differently from other voices, even when we can recognise those other voices with high accuracy. I’m interested in what the possible benefits of this might be, for example for understanding and enjoying spoken language, and in terms of the rewarding value of hearing the voices of people we know and care about. There are also very interesting questions about how familiar identities might be replicated using advances in voice technology such as voice cloning, and whether these techniques can be used to harness some of those familiarity benefits.”

Links