'Brain Stethoscope' Turns Brain Waves Into Sound to Detect Silent Seizures
Stanford University researchers have developed a 'brain stethoscope' that can help detect non-convulsive epileptic seizures. Their algorithm translates the brain's electrical activity into sound so that even non-specialists can detect these silent seizures. The research team has started a company, Ceribell, to develop an EEG device that incorporates their brain stethoscope algorithm.
Transcript
00:00:00 Stanford University. [OSCILLATING TONE] Our senses are tuned to very different things. Whereas the visual system is very good at detecting the edges of objects and motion of objects across space and color versus dark, the auditory system is actually tuned to frequencies. And the frequencies of the brain are the critical information
00:00:25 we're trying to extract. This device is an EEG recorder that acquires the brain waves through some sensors that have to be coupled with this device. And once the brain waves are being recorded, you can click this button here. That's the brain stethoscope button. And as you do that, it will turn the brain waves into sound. It relies on the code and the algorithm that we have worked on. And it enables someone who doesn't
00:00:58 know how to read EEG to be able to actually get meaningful diagnostic information at the bedside. When things are calm, the sound is actually just almost a background hum. [ELECTRONIC HUMMING] When a seizure condition exists, there's no longer that calm background hum. It's very, very prominent, high-amplitude rhythmicity. You hear an intense sound. And you hear a kind of almost screaming.
00:01:30 [OSCILLATING TONE] I think the majority of people think all seizures are convulsive, but that's actually not the case, especially not the case in critically ill patients in the intensive care units. What we did in this study was to actually send the same EEG files to a neurologist and medical students and nurses. Neurologists were asked to review the EEGs exactly the same way as traditionally being done, namely
00:02:01 by visual inspection. And nurses and medical students were asked to review the EEGs by listening to the sound. The ability of an untrained medical student or a nurse to read an EEG is 50%. They're either going to say it is or it isn't there. And it's chance. What we found was that when just listening to the sound of the EEG, medical students could pick up a seizure when there was one 97% of the time.
00:02:30 So it was really effective for people who don't have as much EEG training to listen to the EEG and get the diagnostic information they need to make a real treatment decision. [OSCILLATING TONE] For more, please visit us at Stanford.edu.