Today we are talking about sound, communicating babies, alarm signals, epilepsy and how to measure brain functionality. While this may sound like a wild mixture of topics, it is all relevant in the context of what makes sound difficult to ignore, a topic I have been fascinated with for years. What way to better to start an auditory topic than imagining a sound stage.

Imagine walking down a busy street with a friend of yours. It is noisy but you are still holding a conversation. Cars and trucks are driving by. Once in a while there is a scooter with a modified exhaust. Now imagine hearing an ambulance in the distance. ‘Let’s hope it doesn’t come our way’, your friend says just before the ambulance turns into your street. It comes closer and passes you very fast only a short distance away. It is blaring its horn the entire time. Would you continue your conversation or wait until it has passed? Please feel free to play the audio file to help you imagine the situation.

Traffic and ambulance

What makes the ambulance sound different from other traffic noises, and how did it make you feel? Personally, I would likely wait until the ambulance has passed and I guess I am not the only one. Sirens can be very uncomfortable and difficult to ignore. What do you think makes this type of noise different?

If your answer is that the sound is louder, you are certainly not wrong, but it is only part of the equation. People sometimes enjoy conversations at loud places, like clubs and bars. Even when the siren has the same loudness as the other sounds, you will likely find the siren more intense and uncomfortable. There is something in the sound of a siren that intentionally makes it more uncomfortable and difficult to ignore. If you are listening closely to the siren part of the sound sample, you will notice that the siren sounds fluttery or rough.

It is this roughness that makes it stand out, but in order to understand why, we first need to understand what it is. We can introduce roughness in sound by changing its volume in quick periodic succession. We can for example create a sinusoidal volume mask, like the one you see below and modify the sound accordingly.

In the illustration above, sound levels are downregulated based on the vertical distance between the two sin waves that create the blue areas. Sound levels would not be altered at times where the two waves are furthest apart. At times in which the two waves meet there would be total silence. This modulation between silence and sound would happen 25 times in a second. How would that sound? Let’s test it out with a song snippet. The first sound file below is the original. The second sample is the same song but modulated with a 40 Hz sinusoidal volume mask.

Song sample – original

Song sample – rough

Listening to those two samples you likely recognize that they both play the same song. If that is not the case, you might want to listen again with a more capable sound system. Despite the heavy altering of the sound file, you can still hear the original rhythm pitch and loudness. Roughness can be introduced independently of those other sound properties. The modulation of the sound level is quick enough for the brain to still perceive the original sound as continuous. Of course, our perception does not completely stay the same, hence why we feel this rough texture. This can also be more subtle if the modulation mask is less extreme.  

Speaking a little more scientifically: roughness is introduced in sounds by a modulation of amplitude. Amplitude is related to how loud a sound is perceived. Higher amplitudes lead to louder sounds if all the other sound properties remain the same. To make a sound rough we must very quickly make it louder and less loud. How quick? 40 times a second seems to work best¹, and this is what is usually used. We do not need to use sine waves, we can also introduce abrupt breaks in the sound, or only amplitude modulate some of the sound frequencies.

Babies and roughness

But hold on! How does any of this relate to babies? This article even mentions them in the title, but instead we are only talking about some artificial sound modulations! The connection will become obvious once you have listened to the 3 following sound sample. The first one is a baby laughing, the second one is a baby sobbing, and the third one is a baby angrily crying.

Laughing baby

Sobbing baby

Angrily crying baby

If you have listened to the sound samples, you have most definitely realized how impactful babies use roughness. In case you did not listen to the sound sample, the angrily crying baby made its voice very rough to communicate their distress. In this case the roughness has the same effects as the sirens. I can hardly listen to it myself, because it makes me feel intense discomfort. There is roughness in the sobbing baby sample and even in the laughing baby sample but compared to the last one it is way less pronounced.

Babies use roughness to communicate. From an evolutionary standpoint it is understandable that tiny humans, that cannot take care of themselves, need a method to reliably get attention. On the other hand it happens way to often that parents are overwhelmed by stress, sometimes leading to them shaking their crying babies, and causing shaken baby syndrome². This syndrome is associated with traumatic brain injuries and can cause death³. If you are a parent suffering from stress related to the uncontrollable crying of the baby, look for help! Healthcare professionals will take it seriously and can point you to resources that you can use to help yourself and your baby.

So far, we have talked about natural and artificial rough sounds with a communicative purpose. Other examples of rough sounds include scooters. Not all of them, only the ones with combustion engines. Especially if the exhaust is modified. If you live in a city, you probably know this annoying br|r|r|r|r sound all too well. It is not surprising that this sound in combination with its roughness was found to be a major contributor to noise annoyance in cities ^4,5.

What does roughness do to our brains?

Enough! Let’s stop listening to discomfort inducing sounds. We need to talk about brains. If we understand how the processing of rough sounds is different from other sounds, it will become clear why they are so intrusive. Sound is initially processed in the auditory cortexes. They are located close to the ears. The following illustration will help you to understand their location.

When we monitor brain activity, we can see heightened activity in this region, if a person hears a sound. Louder sounds lead to higher activity. When sounds are perceived as more uncomfortable, we can sometimes detect heightened activity in the auditory cortex. Specifically, if the annoyance is related to the loudness of a sound. Remember the example in the beginning of the loud ambulance? Most definitely your auditory cortex would light up like a Christmas tree in a brain activity scan. But is it the same for auditory roughness?

Luckily there are amazing scientists like Luc Arnal, who investigated this question in a spectacular fashion. Typically, neuroscientific research on humans is rather restricted. While we do not necessarily have to cut open the brains of dead people, all modern imaging techniques have their specific limitation. You can make highly accurate structural scans of the brain with an MRI (magnet based). With some modifications you can even measure metabolic processes of brain areas and infer heightened brain activity from this functional MRI data. The downside is that it is rather slow. An alternative is to measure the electric brain activity with electrodes on the sculp. There is no faster way to measure brain activity, but it is rather difficult to pinpoint the activity to its location. It would be optimal if we could just put the electrodes directly inside the living brain. Fortunately, our ethical standards prevent us from using such unhealthy practices.

If nobody is implanting electrodes for the purpose of research, why even mention it? Because there are some cases in which it is done anyway, and there are good reasons for that. In severe cases of epilepsy, it can help to surgically remove parts of the brain that cause epileptic attacks. Epileptic attacks are very disruptive for a person’s life. During those attacks the brain does not function normally, and the person loses consciousness. I have just finished an interview about this topic, and I will upload it in the upcoming days, so keep an eye out for that.

Some epilepsy cases are very severe, and the available medicine does not help every person. In those severe cases, it can make sense to remove a bit of the defective brain tissue causing the seizure. Precision is super important in those cases, hence why patients sometimes get implanted with electrodes that measure their brains activity for some weeks. With the help of the electrodes and enough data the parts of the patient’s brain that cause the lesion can be identified and it can be avoided damaging functional parts of the brain as much as possible. While the electrodes are inside the brain, the patients cannot leave the room. The infection risk would be too great and if the electrodes are moved, it will cause brain trauma. If the patient is interested, they can participate in research studies unrelated to epilepsy.  

Luc and his team had the amazing opportunity to work with those patients and to let them listen to rough sounds⁶. Interestingly they found that rough sounds do engage deeper brain networks. Networks involved in attention and aversion processing. Those are usually not directly involved in sound processing, which is why it is meaningful that they react to roughness of sound but no other uncomfortable sound properties. The picture below shows one of the networks in question, specifically the default mode network.

Location of the default mode network

What makes the brain’s response pattern to rough sounds so special?

The way that rough sound engages those deeper parts of the brain is super interesting because it leads to very unique activity patterns. Before we get into that, let’s do a quick recap. Sound can be uncomfortable due to its loudness but also its roughness. A sound feels rough if its amplitude is changed periodically in quick suggestion. 40 little breaks per second in a sound (40Hz) are the most efficient way to create a rough impression. This auditory stimulation engages deeper brain networks, those brain networks are important for attention and aversion. In the following graphs I will show that it is not only the location of the brain activity, but also its pattern that is unusual.

The highlighted line in the first graph corresponds to participant brain responses to the following sound. It consists of 250 clicks over one second. A series of clicks would technically fall under the definition of roughness. However, the 250 Hz creates such a short interval, that the brain processes the sound more like it would a tone. It is still highly uncomfortable to listen to, but that is likely more a function of the energy in the sound. This energy translates to loudness, hence why I call this the “loudness sound”.  Please listen to the sound.

The graph shows the average brain activity of multiple participants in relation to the start of the sounds. The height of the line indicates how far away the brain activity is from the reference period (the voltage measured before the sound start). If the line moves away from 0 in either direction, it means that the brain has changed its activity in response to the input. Around 100 ms after sound onset there is a negative spike. This is the brain’s reaction to sound, and it originates in the auditory cortex. The height of this peak is influenced by the energy of a sound and can correlate with how uncomfortable sounds are. If you make an uncomfortable sound louder, you will have a more pronounced spike. The line corresponds to a rather uncomfortable sound, that is not perceived as rough. Next check out the rough sound and then we will talk about how the brain processes it.

You can see at the first glance that the line for the 40 Hz rough sound is different than the one for the 250 Hz loudness sound. The 40 Hz sound has a much lower peak at 100 ms after sound onset. Later on, we see a sustained wave that is absent in the 250 Hz sound from before. If you count how many waves there are in 100ms, you will get exactly 4. If you calculate how many that would be in 1 s, your answer would be 40. Wait, 40? Is that not the exact number of clicks in a second for this sound? And was 40 Hz not also the number at which roughness is the most uncomfortable? Yes! The brain is very sensitive to sound onset and the 40 Hz hit the sweet spot of how the cells trained for sound onset can react. They will synchronize to the periodic sound onsets created by the amplitude modulation, and this activity propagates to deeper structures of the brain. At the same time the brain still identifies the sound as something continuous, like we have seen in the song that we have added roughness to. This might create sensory dissonance in the brain.

How much this dissonance plays a role in the perception of the sound as uncomfortable is not clear. It seems highly likely that the 40 Hz fluctuation that results from cell activity resonating with roughness causes some of the discomfort. At the very least it explains why we cannot ignore sirens or crying babies. The roughness engages networks important for attention. Further it engages them in a continuous manner for as long as the roughness persists. This makes it difficult for the brain to down regulate itself.

Another mystery solved! I would like to say that, but there is still so much that we do not understand. At the moment we have correlations between the peak at 100 ms for loudness and uncomfortableness, as well as between roughness synchronization and uncomfortableness. But that does not mean we know the whole process. Does the forced attention on the sound created by the roughness lead to it being uncomfortable, or is it the other way round? Why are some people more affected? And I have not even started talking about the links with dementia research.

Nevertheless, I am leaving this topic be for the day. Certainty there will be interesting research going on in this field. I will update you probably more than once. I hope you enjoyed reading and that you have learned something.

Have a nice day.

References:

1.            Schneefeld, F. et al. Salient 40 Hz sounds probe affective aversion and neural excitability. Preprint at https://doi.org/10.1101/2022.02.26.482077 (2022).

2.            Goulet, C. et al. Development and Evaluation of a Shaken Baby Syndrome Prevention Program. Journal of Obstetric, Gynecologic & Neonatal Nursing 38, 7–21 (2009).

3.            Newton, A. W. & Vandeven, A. M. Unexplained infant and child death: a review of Sudden Infant Death Syndrome, Sudden Unexplained Infant Death, and child maltreatment fatalities including shaken baby syndrome. Current Opinion in Pediatrics 18, 196–200 (2006).

4.            Morel, J., Marquis-Favre, C. & Gille, L.-A. Noise annoyance assessment of various urban road vehicle pass-by noises in isolation and combined with industrial noise: A laboratory study. Applied Acoustics 101, 47–57 (2016).

5.            Paviotti, M. & Vogiatzis, K. On the outdoor annoyance from scooter and motorbike noise in the urban environment. Science of The Total Environment 430, 223–230 (2012).

6.            Arnal, L. H., Kleinschmidt, A., Spinelli, L., Giraud, A.-L. & Mégevand, P. The rough sound of salience enhances aversion through neural synchronisation. Nat Commun 10, 3671 (2019).