Hacker News new | ask | show | jobs
by ingsoc79 2458 days ago
As someone who works with EMG, the hype CTRL-Labs has been pushing regarding their tech is over-the-top ridiculous.

In short, it's a wearable band that uses electrodes to measure your forearm muscles' activity (electromyography, or EMG), coupled with accelerometers to capture motion. It's awesome technology, but contrary to all the "neuro" buzzword bingo, you can't "control computers with your mind" – you have to physically contract the muscles in your forearm, just like if you were using a mouse or keyboard. This video gives you an idea of what the device is actually capable of: https://www.youtube.com/watch?v=oD4efk8T2X4

Either way, it's exciting to see this kind of tech make it into the consumer space. Even if it doesn't catch on with AR/VR, it will have a lot of positive effects for researchers as well as for users with accessibility and mobility challenges.

3 comments

Whether you have to move your muscles isn't so important, the goal here is to create a control surface that is usable in VR/AR, which this technology can totally still do.

Walking down the street with AR glasses right now sucks because you can't actually do anything without disruptive gestures (holding your hand to your face, talking out loud to nobody, etc). If you could tell the glasses what to do simply by making small imperceptible movements in your hand, that's a huge leap forward towards making AR actually something usable

No doubt it's an interesting product, and if they provide a good SDK for it then a lot of researchers and tinkerers will be happy :).
the question is if this device is more expressive than , say, wearing a glove.
Even if it was equally expressive, it's probably far less obtrusive to wear a bracelet than wear a glove (though that depends on the design of the bracelet I guess)
Despite not being a brain-reading interface, it's indeed very interesting that the device exists. I wonder if there is such a small emg reader that one can buy and tinker with, as this one seems to be tied to their developer software and apis (and will probably be tied to their cloud as is everything today)
The MyoWare sensor is a popular component for DIY projects (https://www.adafruit.com/product/2699?gclid=CJLytM3U0dICFQGT...).
They claim to detect the neuron action potential (without any physical muscle movements), the relevant section from their demonstration which is pretty cool: https://youtu.be/D8pB8sNBGlE?t=730

So, it's not just the forearm muscles' activity, but the actual neuron impulses.

There is no chance they are detecting extracellular action potentials with their giant skin surface electrodes. This is 100% EMG, which is a potential produced by muscle tissue. They might be able to detect very subtle EMG signals that don't correspond to visual movement, but they for sure aren't detecting APs.