The music is eerie, if not altogether aesthetically pleasing. Like a soundtrack moments before a film’s horrifying twist, the sounds of the brain in a state of seizure betrays the plot with little more than a skin-prickling crescendo.
This music, the electrical activity of the seizing brain translated to sound, is a merger of art and medicine, the work of Stanford’s Dr. Josef Parvizi, an epilepsy specialist, and Chris Chafe, a composer and music researcher.
Their initial idea was to take recordings of epileptic patients’ brain activity, turn them into music, and see what they sounded like. What began as a curiosity project then quickly evolved into something more: “sonifying” a seizure in real time, they found, could be a way to more quickly and easily diagnose a patient in the midst of a seizure.
“Imagine that you have a device that is playing the sounds of the brain in action,” Parvizi said. “From a clinical perspective, that’s invaluable. You can’t see inside the brain, but you can listen.”
They described that device as a “brain stethoscope,” a portable headset with electrodes feeding information to a mobile app that would then convert the brain’s electrical signals into audio in real time.
Much like a stethoscope can listen to sounds in different parts of the chest, it could be moved over different parts of the skull, listening in on different parts of the brain. With this technology, they imagined, a mom with an epileptic child, for example, could easily tune in to her child’s brain activity and detect a seizure without a trip to the hospital.
The esoteric music made from a seizure was as valuable to science as it was art.
One piece of music begins as a woman with epilepsy sits in her hospital bed, her brain activity normal in the moments before a seizure. Twenty seconds into the recording, a seizure begins in the right hemisphere of the brain, in the medial temporal lobe.
The right side of the brain, represented by a feminine-sounding synthesized singing voice, begins to oscillate wildly. The left side, a lower, masculine sound, is steady and percussive, its neurons firing off as usual. The sonic chaos subsides as the seizure winds down and the brain enters what’s known as the postictal state, the confused, altered consciousness a person experiences after a seizure.
This particular recording, or electroencephalogram, was made while the woman was a patient at Stanford Hospital, awaiting surgery to remove problematic, seizure-causing brain tissue. The sequences of data from that recording then became the music, an algorithm and custom computer software translating them into an otherworldly, humanoid song.
An electroencephalogram, or EEG, uses electrodes to record the electrical activity from the brain’s billions of neurons, measuring voltage fluctuations and translating them into visuals of wave patterns much like a seismograph records vibrations of the Earth. It is widely considered the best tool for diagnosing epileptic seizures and identifying their cause, but the complex visual waveforms require a highly trained eye to read.
In Parvizi and Chafe’s musical seizure, though, the epileptic activity is easily identifiable to even the untrained ear: When the right side of the brain begins seizing, the song becomes an opera singer on the fritz. Just like almost anyone can pick out an off-key note in a song, the ear can easily discern audio patterns and sonic structures.
Those who care for people with epilepsy often struggle to discern when a seizure is occurring or is about to occur — sometimes a seizure has no dramatic physical effect at all.
Easily turning epileptic brain activity into audio, in real time, could make a seizure much easier to diagnose.
“When you listen to brain waves,” Parvizi said, “you can actually appreciate the dynamics of different stages of the mental state.”
San Francisco Chronicle writer Kristen V. Brown