A young man in his 30s was robbed of movement and speech by a devastating neurodegenerative disease known as amyotrophic lateral sclerosis (ALS). It’s the same crippling disease that the late Stephen Hawking suffered from. The man’s condition deteriorated so badly that he couldn’t even close his eyes, robbing him of his last tool for communicating with the outside world.

But in a stroke of genius, researchers at the University of Tübingen in Germany and the Wyss Center in Switzerland have devised a brain-computer interface (BCI) that translates brain activity into software commands on a digital keyboard. Although BCIs have been used before to restore or enhance communication in paralyzed patients, this is the first time that such a feat has been achieved for a patient in a completely locked-in state.

“To our knowledge, ours is the first study to achieve communication by someone who has no remaining voluntary movement and hence for whom the BCI is now the sole means of communication,” said Dr. Jonas Zimmermann of the Wyss Center in Geneva.

The man had been using an eye-tracking system to communicate since 2015. However, by August 2017, the patient’s condition deteriorated further, affecting the neurons that control eye movement. With no means to fixate his gaze, even this last-ditch effort to interact with the outside world was gone. Soon he would also lose the ability to even open or close his eyes on his own.

After they were contacted by the patient’s family, the researchers moved the man to a hospital, where they performed surgery to implant two microelectrode arrays into the motor cortex.

At this stage, despite the invasive surgery, no one was too hopeful. Previously, patients with advanced locked-in syndrome and no control over their eye movements couldn’t achieve communication even with BCIs. These failures suggested that the loss of physical movement must mean that the neural signals that enabled these muscle contractions were compromised as well. But neither the researchers nor the patient gave up.

After surgery, the patient had to undergo a series of grueling exercises in which he had to mentally imagine moving his eyes like he would select characters on a computer screen. But, as in previous other situations, the brain wave patterns weren’t consistent enough to decode them.

That’s when the researchers switched gears. In a moment of inspiration, they moved towards a neurofeedback approach. What this means is that an audio tone was played back to the patient, whose pitch corresponded to the firing rate of the neural activity. When the firing rate increased, the pitch rose as well, while a lower activity produced a lower tone. With some practice, the patient was able to concentrate better, following the tone until it increased in pitch.

After two weeks of training, the patient was able to focus his mental energy to follow two target tones, a higher frequency that translates to “yes” and a lower hum that means “no”. This ability allowed the patient to write words and sentences, though arduously slow.

The computer plays back each letter of the alphabet, then asks the patient if this is the character they want to use, before moving on to the next letter. It takes one minute on average to spell a single character. It’s slow but at the end of the day, it works.

Some of the messages the patient sent include instructions for his care like “everybody must use gel on my eye more often” and “when visitors are here, head position always very high,” but also personal messages for his family, like asking his son whether he wanted to watch Disney’s Robin Hood together, according to Technology Networks.

These were some of the successful messages, though. The system is plagued by noise so some communication attempts couldn’t always be decoded. According to the study, the BCI could output understandable phrases on 44 out of 107 days of monitoring. But with some improvements, it may be possible to enhance the success rate. Overall, the patient managed to produce 5,747 characters that would have otherwise been impossible to transmit.

For now, this proof of concept is not available to other clinics, but the technology could soon be modified for wider use. For the most challenging ALS cases, such therapy could mean the difference between hell on earth and a shot at a decent life.

“With communication possible, quality of life is good in ALS paralyzed patients,” Birbaumer says. “Patients tell us retrospectively that no communication is torture. [The] standard of care should improve dramatically if you can ask patients about pain, symptoms etc.”

The findings appeared in the journal Nature Communications.