We often take our ability to communicate with others for granted. The ability to verbally say “hi” or “I love you” isn’t something we think about every day. However, for those without the ability to speak, such a thing would be world-changing.
That change is upon us.
Researchers at UC San Francisco have succeeded in giving speech to a man denied that ability. Affected by severe paralysis and unable to form words at all, the man could communicate through a new speech “neuroprosthetic”. The device could translate brainwaves from the man’s brain to his vocal tract and then display that communication on a monitor.
According to the study, which was published in July 2021, the authors stated, “We decoded sentences from the participant’s cortical activity in real-time at a median rate of 15.2 words per minute, with a median word error rate of 25.6%. In post hoc analysis, we detected 98% of the attempts by the participant to produce individual words, and we classified words with 47.1% accuracy using cortical signals that were table throughout the 81-week study period.”
The study in question took over 80 weeks to conduct. Part of that was due to the need to implant a high-density electrode array over the patient’s speech motor cortex. After full healing, the team recorded 22 hours of neural activity within 48 different sessions. The patient was given 50 vocabulary words, and the electrode studied the result within the speech motor cortex.
Because the patient could not speak aloud, he was required to try to talk so that the brain activity would be visible to the electrode. The system was a huge success, even though there were some errors. The study’s authors are already planning to create an “auto-correct” function similar to that used on today’s smartphones to help combat those errors.
The incredible success of this initial trial will allow it to move forward and broaden. Eventually, other patients with severe paralysis and communication challenges will be trialed. With that being said, the study’s authors note that some improvements need to be made.
One of those improvements is to the device itself. Because it is very large and sits atop the head, the wearer cannot use it continuously. However, the size will decrease as development progresses. Another improvement needed is in terms of speed. At just 15.2 words per minute (median – the maximum was 18 words per minute), communication is very slow. Additional vocabulary words also need to be added.
However, one concern not voiced by the study’s authors that is perhaps being overlooked in the (rightful) excitement of the achievement is the potential for privacy invasions.
The speech neuroprosthetic discussed above is a tremendous leap forward. It promises to help tens of thousands of people communicate with their loved ones, friends, doctors, and more. The sheer potential for improved quality of life is astounding.
However, there is a downside here – potential privacy concerns.
For instance, who owns the speech being produced here? Does the paralyzed man own it? Or is it owned by the researchers who developed the neuroprosthetic? What about the software? Who has rights to this speech? Is it protected under the First Amendment? Can it be reused without the user’s permission? Does it represent the user’s intent?
To take these questions further, what happens if there is a malfunction or other error with the prosthetic? How will researchers prevent computers from accessing strong emotions tied to speech – the suppressed urge to call someone a name during a heated argument, for instance – that would never have slipped out were it not for the prosthetic device and software?
After all, retracted or unendorsed words still reveal something about the person speaking them. Where does the patient’s responsibility end, and where does the prosthetic owner/developer’s begin?
In the end, this advance is an inspiring moment. However, like all such technological leaps, it presents genuine concerns for user privacy and protection under the law.