You Don't Have to Be a Scientist to Own a Proper EEG Headset | Hacker News Reader

You Don't Have to Be a Scientist to Own a Proper EEG Headset

braindecoder.com

36 pointsbootload8 comments

Former neuroscientist here. Used to work with intracranial EEG (electrodes directly on the brains of epilepsy patients).

This is very cool, and there will definitely be some neat consumer applications.

But this is not a "proper" headset, compared to lab equipment. Lab equipment has more electrodes, positioning protocols, and electrode gel. As yoz-y mentioned, that electrode gel can boost signals by orders of magnitude. Good luck getting clean gamma signals from a dry electrode. (Power falls off non-linearly with the frequency, and gamma is the highest frequency we typically capture with EEG.)

The most likely scientific application of stuff like this is long-term/ecologically valid recording. E.g., let the user wear this all day with Google Glass. OTOH, even a small number of dry electrodes should suffice to train ML algorithms to turn on your computer with your mind.

They covered a little EEG ON National Geographic's Breakthrough series (Episode 2):

http://channel.nationalgeographic.com/breakthrough-series/vi...

I don't seem to hear much about this technology. Are we making any real progress?

When you say real progress, are you referring to the underlying technology itself, or something else?

With regards to the technology, I'm not sure if there's much more room for improving EEG itself, it seems like a solved problem (how do we measure electro-dermal (or brain) activity). However I'm not a neuroscientist, so I may be off base. Where I do see room for improvement is in systems that can automatically make sense of the signals we're reading. This requires gaining a better understanding of the brain, though.

You are correct. I worked with EEG as part of my PhD research on brain computer interfaces, and it is a super noisy signal regardless of the headset. This makes is very hard to do anything reliably with it outside a pretty narrow scope.

Relevant to that, there was a kaggle competition a little while ago on epilepsy using EEGs. The competition/relevant dataset can be found here: https://www.kaggle.com/c/seizure-prediction/data

A lot of the current research on EEG headsets concentrates on how to acquire signals with cheaper, lighter and more robust headsets in order to be able to push the technology to general public. There is still quite a way to go, especially since the signal quality coming from headsets with dry electrodes is miles behind the standard medical EEG with electrodes that use conductive gel.

Related prior discussion: https://news.ycombinator.com/item?id=8164595.

Got a friend who is a MS evangelist here in Michigan and actually has done presentations using an EEG headset and Azure machine learning to build a lie detector:

https://channel9.msdn.com/Shows/Lets-dev-this/LDT1602

Reply on news.ycombinator.com