No evidence that 'Mozart effect' can relieve symptoms of epilepsy

Unfounded authority. Underpowered studies. Non-transparent reporting. These three factors have perpetuated the popular myth that Mozart's music can reduce symptoms of epilepsy, conclude the authors of a new review and meta-analysis. Sandra Oberleiter and Jakob Pietschnig at the University of Vienna pull no punches in their paper in Scientific Reports. The fact that data from several key studies that have supported this effect were not available for scrutiny was "concerning," they add.

All kinds of claims have been made for benefits from listening to Mozart. As the researchers point out: "To this very day, the public interest in the Mozart effect and the supposedly beneficial effect of Mozart's music on intelligence remains unabated." But Oberleiter and Pietschnig homed in on one specific claim: that listening to Mozart's sonata KV448 reduces the symptoms of epilepsy. Such assertions were first made in the late 1990s, and since then two studies have reported replicating the original findings, with other studies reaching similar conclusions.

This idea has been widely reported in popular media outlets, the team adds, and it has attracted attention in part because not only do drug therapies often cause severe side effects, but they don't work for almost a third of people with epilepsy. So safe new treatments are badly needed. A few reviews of studies on this sonata and epilepsy have already been conducted, but they have had weaknesses, the researchers argue. For example, seven of the eight studies included in one review had the same first author.

For this, the first formal, systematic meta-analysis, Oberleiter and Pietschnig first identified the most relevant published journal articles. This gave them a total of 26 studies. But they had to exclude a massive one third from their analysis because they could not obtain the original data sets, or even the summary statistics. This was either because the corresponding authors didn't respond to their requests for this information, or because the authors reported that the data was no longer available.

This was despite the fact that all but one of these papers had been published since 2010, and so in an time when the importance of data openness and sharing is widely recognised — and despite the fact every one of the journals in which these papers were published "either mandated or at the very least encouraged authors to share their data", Oberleiter and Pietschnig write.

Of the 12 studies for which the necessary data was not available, 11 had reported a positive specific Mozart effect. So, while these studies contribute to the popular assumption that the effect is real, their data could not be independently re-analysed, and so, the researchers argue, their authority is unfounded. "We do not mean to suggest that results from underdocumented findings are necessarily flawed, but their value for answering specific research questions is uncertain."

Oberleiter and Pietschnig had to exclude another six studies from the original 26 for different reasons; for instance, four of these were single case reports so not useful in a meta-analysis. This left them with eight studies, which had a total of 207 participants. Though all of these studies investigated Mozart's sonata KV448 as a potential therapeutic, some also used other pieces of music. And while six of the studies investigated impacts on epilepsy, one study focused on levels of pain in premature infants and one looked for changes in blood pressure in stroke patients.

A few of these studies did find some beneficial effects from listening to the sonata. However, the sample sizes in these eight studies were small, ranging from 11-70 people. This "obviously raises concerns about the power to detect any meaningful effects," Oberleiter and Pietschnig write. Only three of the eight studies were gold-standard randomised-controlled trials — and they were all "insufficiently powered" to detect a meaningful effect, the pair reports. In fact, the pair's analysis across all studies showed no significant Mozart effect (or indeed any beneficial effect of listening to music more generally).

Their conclusion? "Our meta-analysis shows that there is no meaningful support for a beneficial effect of listening to Mozart's sonata KV488 on any medically relevant conditions, let alone a specific Mozart effect for epilepsy. Unfounded authority, underpowered studies, and non-transparent reporting appear to be the main drivers of the Mozart effect myth."