Scientists have developed a new technique to use the brain’s electrical activity to create new music that’s more than just a “beat” of sound.
The technique works by combining neural activity with music played on an audio device.
In some cases, this means using neural activity to play a specific piece of music that is different from the music that was recorded with the brain.
It’s the first time that researchers have used neural activity in a way that was similar to that which is found in humans.
The results were published online in the journal Nature Neuroscience.
“We wanted to show that there’s a brain mechanism that can generate new music, and it’s similar to music that humans do,” says lead author Peter Gauden, from the University of New South Wales.
“So the challenge for us was to build a system that was just as good as the human brain.”
The researchers used electrodes that can pick up electrical signals from specific brain areas, and then used computer algorithms to build an artificial neural network that matched that neural activity. “
But until now, we’ve had no way of getting at the neural circuitry that drives the music and how it’s produced.”
The researchers used electrodes that can pick up electrical signals from specific brain areas, and then used computer algorithms to build an artificial neural network that matched that neural activity.
They used this to produce music that sounded like a drum beat, which the researchers found to be very similar to human music.
The researchers also found that the neural activity generated music that mimicked the human music, even when the researchers used different neural networks for the two music types.
“This work shows that we can use neural activity from brain areas to produce a very specific and accurate musical representation of sound,” says co-author Michael Schafer, from University of Newcastle.
The team’s results have implications for music production, says coauthor and music theorist, Professor Brian Cox.
“It shows that there is an evolutionary benefit in having a brain that is able to produce an accurate representation of musical sounds,” he says.
“If you’ve ever been in a music studio and you’ve got the brain in your head, you can actually hear music coming from that area, so it could be a very useful tool for musical production.”
In addition to the two experiments, the research team has built two more artificial neural networks that use neural networks to produce complex sounds, including an orchestral version of a piece of classical music.
These new systems are expected to be used in music production.
“The goal was to show a way of generating a musical representation that was really close to what the human auditory system is capable of producing,” says Schafer.
“And it’s also possible that this could be applied to creating a new type of musical work that is based on a different kind of neural network.”
The study was supported by the Australian Research Council and the NSW Medical Research Council.