I bought a commercial EEG monitor called Mindwave ($100). I’m interested in two things:
(1) Can reliable, useful signals be extracted from EEG?
(2) If so, can this (bio)feedback be used for training mental states?
Extra info:
(2) This appears to be the main gimmick for selling these devices. It includes proprietary measures of “attentiveness” and “meditation”. You can record these signals over time and see if you can improve your mental state. I was also intrigued by a connection with a program called “Vitamin R”, which suggested that you could leave the EEG on while working and it would find your peak productivity times and warn you when your attention wanders.
I’ve tried the meditation program and it does seem to be legitimately measuring something about meditation. I’m able after a few minutes to keep a consistently higher meditation score. An experienced meditator I know (>15 years practice) was able to max out the reading in under 10 seconds, and keep it there indefinitely.
(1) Multiple electrode EEGs have seen a moderate amount of success in the literature. People are able to control up/down/left/right and use this ability to type a few characters per minute. Single electrode EEGs (like Mindwave and several other companies produce for $100-$200) have a more limited capability. Research suggests that they probably do provide some signal that can not be chocked up solely to muscle movement (blinks, etc affect the EEG reading), but these results are quite limited.
I’d like to explore more, since some of my academic research is about extracting signals from noisy data, and this is a very interesting noisy data source (to me), my brain! Eventually, I’d like to do a longer post about brain-computer interfaces for less wrong. Wikipedia article
From my experience: Mindwave has unusually good algorithms for filtering out blinks and facial movements, but the utility of a single electrode EEG is limited. Larger EEG arrays don’t have algorithms as good at filtering out blinks and facial movements.
Thanks! I think you’re right. I found a program that gave me “raw” EEG signals from Mindwave at different bands (theta, gamma...) and blinks and muscle movements caused huge spikes. But the Neurosky attention/meditation metrics were relatively unaffected. Unfortunately, this is a wheel I’ll have to re-invent to extract good signals. (I assume so, since neurosky’s metrics are proprietary.) Though I hope some literature exists about this.
I bought a commercial EEG monitor called Mindwave ($100). I’m interested in two things:
(1) Can reliable, useful signals be extracted from EEG?
(2) If so, can this (bio)feedback be used for training mental states?
Extra info:
(2) This appears to be the main gimmick for selling these devices. It includes proprietary measures of “attentiveness” and “meditation”. You can record these signals over time and see if you can improve your mental state. I was also intrigued by a connection with a program called “Vitamin R”, which suggested that you could leave the EEG on while working and it would find your peak productivity times and warn you when your attention wanders. I’ve tried the meditation program and it does seem to be legitimately measuring something about meditation. I’m able after a few minutes to keep a consistently higher meditation score. An experienced meditator I know (>15 years practice) was able to max out the reading in under 10 seconds, and keep it there indefinitely.
(1) Multiple electrode EEGs have seen a moderate amount of success in the literature. People are able to control up/down/left/right and use this ability to type a few characters per minute. Single electrode EEGs (like Mindwave and several other companies produce for $100-$200) have a more limited capability. Research suggests that they probably do provide some signal that can not be chocked up solely to muscle movement (blinks, etc affect the EEG reading), but these results are quite limited.
I’d like to explore more, since some of my academic research is about extracting signals from noisy data, and this is a very interesting noisy data source (to me), my brain! Eventually, I’d like to do a longer post about brain-computer interfaces for less wrong. Wikipedia article
From my experience: Mindwave has unusually good algorithms for filtering out blinks and facial movements, but the utility of a single electrode EEG is limited. Larger EEG arrays don’t have algorithms as good at filtering out blinks and facial movements.
Thanks! I think you’re right. I found a program that gave me “raw” EEG signals from Mindwave at different bands (theta, gamma...) and blinks and muscle movements caused huge spikes. But the Neurosky attention/meditation metrics were relatively unaffected. Unfortunately, this is a wheel I’ll have to re-invent to extract good signals. (I assume so, since neurosky’s metrics are proprietary.) Though I hope some literature exists about this.