You might want to note this in the README of that repository, and maybe update the tag-line description at the top of the github page. (I assume it's this one https://github.com/daeken/Emokit)
Woah, awesome! I know it's just a POC, but it's great to see people using Emokit -- I'm disappointed that I didn't have enough time to dedicate to doing anything with it when I started it. Let me know if you run into any issues.
Thanks! I also did some basic rhythm detection: https://github.com/agermanidis/emokit/blob/master/python/uti..., and added a simulation option so folks that want to contribute to the library but don't actually have it (I suspect that there are exactly zero of them, but anyway) can still test it on sample data.
Oh very cool -- I wrote some super basic stuff for this, but it didn't work well at all, and I never had any time to improve it. Is your stuff merged into the qdot branch yet? If not, you should definitely send a pull request, as I think it'd work well with what he's been working on.
This technology is cool, yes, but it works pretty terribly. I've done work with this device in my postdoc, and you really can't get any useful signals from it. There is another company similar to this one - NeuroSky, google it - and they're just as bad. The technology for reliably obtaining differentiable neural signals using non-glue/screw electrodes is just not ready for production yet.
It's definitely not research-quality (as I initially thought it might be), but there are still some cool things you can do with it. One interesting possibility is tying it into Max/Msp -- there's been some work along that front lately.
From what I understand reading about this a few years ago: EEG with this kind of external equipment can't read more than one strong impulse signal at a time(that would be a simple detection, binary on or off), and certainly can't locate signal origin.
So all the peripheral sensors detecting facial muscle movement do the real work. Seems pretty dishonest to claim brain interface.
Am I misinformed? We're all waiting for this technology to develop, that creates a big opportunity for misdirection.
Strictly speaking, every interface is a brain interface. The real test is whether it's possible to do things that can't be done with mouse/keyboard/touch/etc. Though the consensus currently seems to be "no", maybe future applications will surprise us.
That's not entirely true -- there are some applications of the Epoc that can actually perform detections which cannot solely be attributed to facial muscle movement. [1] Of course, the signal-to-noise ratio is much lower than that of research-grade devices, so a lot of preprocessing is required.
> The Dial Tim application works on similar principles to P300-spellerbrain-computer interfaces: the phone flashes a
sequence of photos of contacts from the address book and a P300 neural signal is elicited when the flashed photo matches the
person whom the user wishes to dial.
This is just as I've said above, one brain signal is used over time.
Another player to watch in this space is NeuroVigil. Right now they're mainly focused on EEG for healthcare applications, but their technology could be modified for other use cases.
I don't know much about this offering, and I am not a neuroscientist/md, but I do know many well known and respected neurologists/epeleptologists and have worked with them for quite a while (I currently work in the Dept. of Neurology at a very well known institution). Nevertheless, even I know that signals acquired from surface electrodes are very difficult to "read" let alone useful to manipulate anything. For one, a patient who is awake and moving will generate a very difficult to ascertain signal to noise ratio due to too much muscle artifact. The best/most reliable I have seen via surface electrodes are simple binary commands along the lines of "move the dot on the screen up or down" (my brother did that as his senior thesis for his undergrad bio-engineering major).
Now, the guys at emotiv may have devised a way to enhance the electrodes and/or filter out artifact but I would really have to put this to some of my scientist buddies who would actually know what is going on under the hood here. I am inclined to think they have not because I haven't heard any excitement from any of the neurologists in the epilepsy division. I've seen expert technicians actually apply electrodes both surface and implanted (aka. in the operating room with ones brain exposed). Let's talk surface. There is a lot of expert care and skill that goes into setting up electrodes (and in specific places). On top of that, the electrodes are generally coated with a gel. All of this to enhance the electrical signal. The headset from emotiv looks like it is dry and how do you know if it is positioned properly?
tl;dr, I would love to think this has real value but I just don't think so.
Thanks for all the posts, that point out, that the technology is still unreliable.
But com'on people, this is exciting! Who knows what great ideas might come out of this, now, that there is an affordable, commercial product. Ideas are made out of ideas. And so are applications made out of applications.
I wonder if playing games like "Spirit Mountain Demo Game" can change your mental awareness and maybe even concentration abilities.
18 comments
[ 4.2 ms ] story [ 34.6 ms ] threadEdit: Repo in the parent was previously pointing to my original, unmaintained library. This comment no longer makes sense :)
So all the peripheral sensors detecting facial muscle movement do the real work. Seems pretty dishonest to claim brain interface.
Am I misinformed? We're all waiting for this technology to develop, that creates a big opportunity for misdirection.
[1] NeuroPhone is an example: http://www.youtube.com/watch?v=tc82Z_yfEwc. Here's a peer-reviewed article describing the preprocessing involved: http://sensorlab.cs.dartmouth.edu/pubs/neurophone.pdf
This is just as I've said above, one brain signal is used over time.
http://www.neurovigil.com/
Now, the guys at emotiv may have devised a way to enhance the electrodes and/or filter out artifact but I would really have to put this to some of my scientist buddies who would actually know what is going on under the hood here. I am inclined to think they have not because I haven't heard any excitement from any of the neurologists in the epilepsy division. I've seen expert technicians actually apply electrodes both surface and implanted (aka. in the operating room with ones brain exposed). Let's talk surface. There is a lot of expert care and skill that goes into setting up electrodes (and in specific places). On top of that, the electrodes are generally coated with a gel. All of this to enhance the electrical signal. The headset from emotiv looks like it is dry and how do you know if it is positioned properly?
tl;dr, I would love to think this has real value but I just don't think so.
But com'on people, this is exciting! Who knows what great ideas might come out of this, now, that there is an affordable, commercial product. Ideas are made out of ideas. And so are applications made out of applications.
I wonder if playing games like "Spirit Mountain Demo Game" can change your mental awareness and maybe even concentration abilities.