Surgically implanted devices that allow people with paralysis to speak can also intercept your inner monologue.
That’s the conclusion of a study on brain-computer interfaces (BCIs) published in the journal Cell.
The discovery could lead to BCIs that allow paralyzed users to produce synthesized speech more quickly and with less effort.
But the idea that new technologies can decode a person’s inner voice is “disturbing,” says Nita Farahany, professor of law and philosophy at Duke University and author of the book: The Battle for Your Brain.
“The more we advance in this research, the more transparent our brains become,” says Farahany, adding that measures to protect people’s mental privacy are lagging behind the technology that decodes signals in the brain.
From brain signal to speech
BCIs are capable of decoding speech using tiny arrays of electrodes that monitor activity in the brain’s motor cortex, which controls the muscles involved in speech. Until now, these devices relied on signals produced when a paralyzed person actively tries to articulate a word or phrase.
“We’re recording the signals while they try to speak and translating those neural signals into the words they’re trying to say,” says Erin Kunz, a postdoctoral researcher in Stanford University’s Translational Neural Prosthetics Laboratory.
Relying on signals produced when someone with paralysis tries to speak makes it easier for them to mentally close their lips and avoid speaking too much. But it also means they need to make a concentrated effort to convey a word or phrase, which can be tiring and time-consuming.
So Kunz and a team of scientists set out to find a better way — by studying the brain signals of four people who already used BCIs to communicate.
The team wanted to know if they could decode much subtler brain signals than those produced by attempted speech. The team wanted to decode imagined speech.
During attempted speech, a paralyzed person does their best to physically produce comprehensible spoken words, even if they can no longer do so. In imagined or inner speech, the individual simply thinks of a word or phrase — perhaps imagining what it would sound like.
The team found that imagined speech produces similar but weaker signals in the motor cortex than attempted speech. And with the help of artificial intelligence, they were able to translate those weaker signals into words.
“We achieved up to 74% accuracy in decoding sentences from a vocabulary of 125,000 words,” says Kunz.
Decoding a person’s inner speech made communication faster and easier for the participants. But Kunz says the success raised an unsettling question: “If inner speech is similar enough to attempted speech, could it leak involuntarily when someone is using a BCI?”
The research suggested that this could happen, under certain circumstances, such as when a person silently recalls a sequence of instructions.
Password protection?
So the team tried two strategies to protect BCI users’ privacy.
First, they programmed the device to ignore inner speech signals. That worked, but it reduced the speed and ease associated with decoding inner speech.
Then, Kunz says the team adopted an approach used by virtual assistants like Alexa and Siri, which only wake up when they hear a specific phrase.
“We chose ‘Chitty Chitty Bang Bang’ because it doesn’t occur often in conversation and is highly identifiable,” says Kunz.
This allowed participants to control when their inner speech could be decoded.
But the safeguards tested in the study “assume that we can control our thoughts in ways that may not match how our minds actually work,” says Farahany.
For example, says Farahany, study participants couldn’t prevent the BCI from decoding the numbers they were thinking of, even if they didn’t intend to share them.
This suggests that “the boundary between public and private thought may be thinner than we imagine,” says Farahany.
Privacy concerns are less problematic with surgically implanted BCIs, which users understand well and which will be regulated by the Food and Drug Administration (FDA) when they hit the market. But that kind of education and regulation may not extend to future consumer BCIs, which will likely be worn like caps for activities like video gaming.
Early consumer devices won’t be sensitive enough to detect words the same way implanted devices do, says Farahany. But the new study suggests that capability could be added someday.
If so, says Farahany, companies like Apple, Amazon, Google, and Meta could find out what’s on a consumer’s mind, even if that person doesn’t intend to share it.
“We have to recognize that this new era of brain transparency is really an entirely new frontier for us,” says Farahany.
But it’s encouraging, she says, that scientists are already thinking of ways to help people keep their private thoughts private.
Source: npr.org by Jon Hamilton