Facebook developing device to type words you are thinking

FACEBOOK has announced a major breakthrough in its plans to allow people to type just by thinking of the words they want to use.
But the US web giant says it is still some way away from a finished product. In 2017 the former head of the US Defence Advanced Research Projects Agency (DARPA), Regina Dugan, asked Facebook’s developer conference: “What if you could type directly from your brain?”
Just over two years later a research paper published in the journal Nature Communications has revealed the company is now able to use records of brain activity while people speak, to instantly decode what they were saying into text on a computer screen.
This ability to decode – an admittedly small set of – words and phrases from brain activity in real-time is a huge leap forward in the field of brain-computer interfaces, or BCIs. Working with research subjects who were already undergoing brain surgery to treat epilepsy, Facebook and scientists at the University of California, San Francisco (UCSF) implanted electrodes to record patients’ neural activity.
At the moment BCI research is limited to invasive electrodes which need to be surgically implanted in patients’ brains, but the company is working on a wearable BCI device to monitor these neural signals.
“We’re a long way away from being able to get the same results that we’ve seen at UCSF in a non-invasive way,” the company stated, before adding they had started work on it with a system using near-infrared light.

What are the signals?
Think of the word “neuroscience” – you might visualise how it’s spelled or your thoughts might throw up associated words and ideas.
Whatever happens inside of the human brain when you think of that word – or any other – is still a mystery.
Human brain activity when people think of words isn’t identifiable and it would be impossible for researchers to simply capture those thoughts and transcribe them.
But if you were to say the word “neuroscience” aloud then your brain needs to send a messages to the muscles in your jaw and mouth to pronounce it.
These signals are very similar every time they’re sent, and potentially similar again person to person – and although these similarities might be too difficult for a human to recognise, they offer a problem which artificial intelligence technologies are perfectly suited to identifying the patterns for.
But understanding these signals remains a topic at the very furthest frontier of neuroscience and AI.
“The past decade has seen tremendous strides in neuroscience – we know a lot more about how the brain understands and produces speech,” Facebook explained.
The company added that new AI research has improved the capabilities of technology to translate speech to text, and AI is an area of research the company has poured a lot of investment into over the past few years, luring academics from even the most prestigious universities to its research team.
“Taken together, these technologies could one day help people communicate by imagining what they want to say – a possibility that could dramatically improve the lives of people living with paralysis.”
How does the device work?
“Like other cells in your body, neurons consume oxygen when they’re active. So if we can detect shifts in oxygen levels within the brain, we can indirectly measure brain activity,” the company explained.
A prototype device fitted with enormous sensors aims to use near-infrared light to measure these oxygen levels in a non-invasive way, therefore revealing brain activity.
Currently the prototype is “bulky, slow, and unreliable” the company acknowledges – and certainly less handsome than the Oculus headset which Facebook also owns.
While measuring oxygenation may never allow us to decode imagined sentences, being able to recognise even a handful of imagined commands, like “home,” “select,” and “delete,” would provide entirely new ways of interacting with today’s VR systems – and tomorrow’s AR glasses.
It could take a decade to develop alternative methods of measuring neural signals that provide the same accuracy as implanted electrodes without the same level of invasiveness, but Facebook says it thinks it can close the gap. — news.sky.com