I’ve long lamented the lack of a direct brain-to-device interface. So here we are, finally: a simple bit of EEG output, snapped at a time where I was relaxing rather than highly focused (those are specifically measured metrics):

The above was done using a Neurosky MindWave headset and a bit of Python code (on github) written by Andreas Klostermann run on my Linux desktop.

The Neurosky hardware comes with some Windows and Mac software but the protocol is public so interfacing with it from Linux or anywhere else is not a problem. The headset I have uses a wireless interface with a USB dongle that shows up as a serial device. Other models have Bluetooth (enabling use directly on Android and iPhone) but the essense is the same. You can read a defined output stream, and people build more convenient interface libraries on top of that.

Apart from visualising/recording your own EEG, at home, at very low cost (about $150) and being able to not just view but play with all that directly, the one-sensor headset decodes “meditation” and “attention” states. These signals are decoded by an signal processor developed by Neurosky, based on the different frequency bands used by the brain. There’s a MindBand (not yet available for general sale) which has two sensors instead of one, enabling additional uses for controlling devices. But even with one sensor you can actually do a fair bit already.

The system uses a dry surface sensor (metal strip) on the forehead, and an earlobe clip to provide a baseline as the system needs to filter out the myriad of environmental noise including the 50Hz buzz from the power grid, and much more.

Ok, it’s obviously not yet at the stage where I can simply “braindump” my thoughts into a computer, but it’s a modest start to some really useful stuff. I reckon it’s one of those enabling technologies… what was previously available only in labs with expensive equipment, we can now play with directly. Such tools change a lot of things in a very dramatic way. Watch this space!