Epigraph Vol. 22 Issue 5, Fall 2020

Podcast: Sonatas and seizures - can listening to Mozart help to treat epilepsy?

Over the past two decades, several studies have suggested that listening to music—specifically, Mozart’s K448, Sonata for Two Pianos—is associated with seizure reduction. Recently, a one-year study from the Krembil Brain Institute in Toronto found that three months of daily listening to a section of K448 was associated with a 35% reduction in seizures, compared with listening to a control piece.

 Podcast icon  Listen to Podcast

(Podcast opens with the first movement of Mozart K448 – Sonata for Two Pianos)


This is Mozart’s sonata for two pianos in D major, also known as K448. It’s one of his most recognizable compositions, but he probably never imagined it would have such a prominent – and controversial – place in neurology and neuroscience.

Over the past 20 years, research has examined the potential association between exposure to K448 and a reduction in seizures. A recent study from the Krembil Brain Institute in Toronto, Canada, found that in people with epilepsy, listening to K448 every day reduced seizure frequency by an average of 35 percent.

Marjan Rafiee:

My name is Marjan Rafiee, I’m a research assistant at the Krembil Brain Institute at UHN and I work on understanding how listening to music can be used to reduce seizures in people with epilepsy.

This was a one-year study in which we asked individuals to listen to Mozart K448, and we showed that listening to K448 can reduce seizures in these individuals in comparison with a control piece lacking the rhythmicity.

We had 13 people in the study; in one group they started by listening to K448 once a day for three months, and then they switched to the scrambled Mozart without the rhythmicity. The other group did the opposite; they listened to the scrambled Mozart once a day and then switched to the actual K448. Both groups had an initial and final three-month period in which they weren’t listening to anything, so we got a one-year period to look at their seizure frequency and how it changed.

We found that individuals had a reduction in seizures during the three months of listening to Mozart K448, but not while listening to the scrambled Mozart.

Taufik Valiante:

My name’s Taufik Valiante, I’m a neurosurgeon at the Toronto Western Hospital, associate professor at the University of Toronto and a scientist at the Krembil Research Institute.

What Marjan and I was particularly surprised at, and I was a bit nihilistic about this, was how much of an effect size we saw. As a surgeon, as much as I like to talk about surgery being a panacea, it’s not, and individuals need other ways of managing their epilepsy and I think it’s probably a combination of many things, but we’re pursuing a strong neurobiological hypothesis.


Music as neuromodulation isn’t all that controversial. Of course listening to music can change us – it can make us feel happy or sad, energize us, or help us to relax. With more on music and the brain, here’s Philip Pearl, who’s director of epilepsy and neurophysiology at Boston Children’s Hospital, a professor of neurology at Harvard Medical School, and a professional musician.

Philip Pearl:

So we have epilepsy surgery patients and they have implanted electrodes, and when you play music, certain neurons respond, and they’re different than the neurons that respond to language. They’re in the same neighborhood – the superior temporal gyrus – but they’re different neurons. The music could be instrumental or vocal, or even a drum solo, which some people may argue is not music. I was a percussion major at the Peabody Conservatory of Music when I was at Johns Hopkins, so I’m biased toward percussion. But there are different centers in the brain activated by music. Different neurons also are activated by different timbres of music, different tones, and different frequencies and volumes.

And if you play the same note louder vs softer, it’s not that the same neuron is activated more by the louder music – different neurons are activated for the louder versus the softer music. The brain handles music in very different ways, neuron for neuron.

Anthropologically there’s been this longstanding debate, are music and language the same? They’re not the same. Maybe they both served as communication, and for other animals music is part of mating, such as bird calls. But we have demonstrated from intracranial EEG in epilepsy surgery patients that music activates different neurons than language, and different aspects of music activate different neurons.


In 1993, researchers found that college students who listened to Mozart K448 improved their scores on spatial reasoning tests. That finding, as well as related ones, evolved into claims that listening to a few minutes of Mozart would raise IQ – which naturally has provoked skepticism.

Philip Pearl:

The reason there’s been controversy is because it morphed into the belief that if a baby in the womb listened to Mozart during pregnancy, the IQ would go up in the baby. Or if you gave children music lessons very young, their IQs would go up, because Mozart must have been a genius. He started playing piano at 3, and his compositions were incredible by the time he was about 12. He wrote K448 at age 25; it’s a brilliant three-piece piano sonata with its center in D Major the whole way through. The first movement is allegro and the second is andante and then it goes back to allegro.

When people listen to the first couple of minutes, they’re only hearing part of the first movement – that’s supposed to increase your IQ if you listen to it when your mother’s pregnant with you? The skepticism has to do with extrapolating these limited studies to infancy and when we’re embryos.

However, some analysis does show that K448 is unique in certain ways, though there’s no solid evidence that its unique features are tied to seizure reduction.

Marjan Rafiee:

Listening to music can change brain activity in so many ways. One way to characterize music is to look at features like pitch, rhythm, and tempo and how these change with respect to one another mathematically over time.

Mozart pieces are famous in terms of having long-term periodicity, and they have some of the most unpredictable rhythmic structures mathematically. You can categorize musical pieces and compare them – this one has the least predictable rhythmic structure, this one has the most predictable, and then look at the effect of the pieces on seizures. So I would say one of the most important things is the features musical pieces have, and we know that Mozart is different in some ways, compared with other classical composers.


The Krembil Brain Institute study is only the second randomized clinical trial of K448 for seizure reduction. The other study included 36 people who listened to Mozart periodically throughout the night. That study found a 24 percent decrease in seizure frequency, with about one-fourth of participants having no seizures during this phase.

It’s difficult to conduct large studies in this area, given all the variables. The Krembil researchers assessed more than one thousand people for participation, and only 13 were included.

Marjan Rafiee:

We had a wide range of exclusion criteria because we wanted to make sure it was controlled. We had to make sure medications would not change in the upcoming year, and that they were not interested in trying other treatment options during that time. We also were limited to looking at people in Toronto because we asked them to come to the hospital every 3 months. So that affected the numbers.


Animal studies provide an opportunity to explore the mechanisms behind Mozart in a more controlled environment. At the University of Utah, Cameron Metcalf and colleagues recently assessed the effects of Mozart on mouse models of pain, as well as epilepsy kindling models. A kindling model uses repeated electrical or chemical stimulation to induce seizures.

Cameron Metcalf:

So with kindling we said let’s kindle the animals while they’re being exposed to music, or the control group, and ask if we see major differences in seizures. We really didn’t; we hypothesized that the animals exposed to music would have fewer seizures, that they would take longer to kindle, and we didn’t see that. What was remarkable though is that we saw improved survivability.

With CF1 kindling you can lose up to half the animals in any given run; that varies from run to run but it’s somewhere between a third and a half. And we saw that in the animals exposed to music we had improved survivability. We don’t understand the mechanism for that but it’s something we’re going to continue to look at to see if we can pick up on why that may be occurring. If it’s unique to the animal strain, is there an indication of an anti-inflammatory effect or any behavioral effects – animals that are fully kindled have behavior deficits as well, so that’s something we’re going to follow up on.


Metcalf and colleagues are using strategies adapted from pharmacology to try to identify if certain aspects of the Mozart sonata are responsible for changes in seizure frequency or improved survival.

Cameron Metcalf:

I did a lot of peptide research and one of our controls in peptide studies is we say okay here’s this peptide that has an analgesic or anti-seizure effect. Now let’s scramble it just a little bit, so the pharmacophore is a little bit different. Amino acid sequence is just a bit different. Then you can show that the activity is dependent on the active site.

We don’t really know which part of the music is providing the real benefit so we do plan on taking a control musical segment that’s just a little bit different, doesn’t have the same architecture as the original Mozart piece and we’re collaborating with our colleagues at the Gifted Music School to be able to help to design that.


Research on K448 and other musical pieces may ultimately offer an additional way to help people with epilepsy.

Taufik Valiante: As much as we talk about medical interventions and absolute reduction in seizures, it’s about how the individual perceives the benefit. We know from the surgical literature that particularly in the context of resective surgery, there’s a huge benefit to becoming seizure free. As much as we try to approach that in whatever we do, when we have individuals who may not benefit from surgery, we’re always looking for these other ways of helping.

Marjan Rafiee: I wanted to mention that this was a multidisciplinary work, and I wanted to thank the epilepsy clinic at Toronto Western Hospital, the neurophysiology lab and most importantly Epilepsy Ontario and all the members of our lab at the Krembil Brain Institute. I couldn’t have done this without their support. On top of all of this we had some of the most amazing individuals in this study – I learned a lot from them and I’ll be always grateful.

(Mozart K448 plays)