Late in 2023, two research groups accomplished an amazing feat. Using a brain–computer interface, they provided speech to two individuals who had lost the ability to communicate due to paralysis or amyotrophic lateral sclerosis. Both studies described “brain-to-speech” neuroprostheses capable of converting a person’s thoughts into speech at over 60 words per minute with an error rate of 25%, albeit with vocabularies of limited sizes. And one of the studies analyzed a participant’s facial expressions to control an avatar that could convey that individual’s emotional nuances related to the sentences made for them.
These remarkable findings, supported in part by grants from the National Institute on Deafness and Other Communication Disorders (NIDCD), sound like science fiction—but they are very real, thanks to modern tools. New neurotechnologies like this work by recording electrical activity from parts of the brain while a participant imagines speaking but cannot physically verbalize the words. A computer then decodes this brain activity and directs a speech synthesizer to articulate not just individual words, but whole sentences. A few years ago, similar Brain Research Through Advancing Innovative Neurotechnologies® (BRAIN) Initiative-funded research with a research participant paralyzed by spinal cord injury showed that a brain–computer interface could “compose thoughts” via mental handwriting that converted brain activity into text.
Collectively, the findings build on years of BRAIN Initiative-funded projects refining brain-recording techniques, studying the motor circuits underlying speech generation, and developing sophisticated artificial intelligence (AI) methods to analyze and decode brain activity patterns and convert them into computer-generated speech.
If you asked me 10 years ago, I would would not have imagined this to be possible today—still, it’s thrilling to see such significant progress. In fact, neurotechnologies in general are starting to make a significant impact on people’s lives, as I wrote about in my last blog post, Personalized Cures on the Horizon—A Fresh Take on Established Technologies. Following up in this blog post, I will explore the human-machine interactions a bit more.
It’s important to recognize that newly tested speech prosthetic devices remain in an early experimental phase and can restore only a basic level of speech expressed by a few words, like “I’m cold,” or “Please hand me a drink of water.” But still—the opportunity to communicate is dramatic for individuals who are unable to speak with family and friends due to a cruel disease. It’s possible that we’ll see more widespread adoption of speech neuroprostheses in the next 5 to 10 years using interfaces that are less invasive than the brain implants used now.
The Human and the Machine
The BRAIN Initiative funds substantial work on brain–computer interfaces—the cutting edge of neurotechnologies. Other examples include prosthetics that translate brain activity to movement of a limb, communication devices, and visual prosthetics, which aim to restore some level of sight for people who are blind. Bypassing limbs, the mouth, or the eyes, these machines directly interact with the brain, potentially affecting functions central to the organ that defines humanness such as thoughts and emotions. It may thus seem obvious that human-centered technologies need to be grounded in the perspective of the actual humans who will be using these machines to restore function lost to disease or injury. Although powerful neurotechnologies can quite literally restore a big part of an individual’s identity, they may also invoke changes to that identity. For this and other reasons, it is therefore essential that research using neurotechnologies in humans proceeds cautiously and fully acknowledges known potential risks and benefits to people who are using them.
Although all clinical research must adhere to these principles, there is something very special and unique about humans studying the human brain, since we have the potential to alter—in a direct fashion—an individual’s agency and ability to make decisions about the interventions themselves. And as more and more brain activity data are being collected, with rapid advances in AI, what may appear to be a jumble of data today could in the future be used to re-identify an individual from whom the data were originally obtained. This was one of several topics covered in a July 2023 BRAIN Workshop on Ethics of Sharing Individual Level Human Brain Data Collected in Biomedical Research. As I keep reminding myself, today’s science fiction may soon be tomorrow’s reality.
Rest assured that neuroethics remains firmly embedded across the BRAIN Initiative research portfolio, including a dedicated scientific project team on neuroethics. In addition, the BRAIN Neuroethics Working Group, a group of non-NIH experts, meets regularly to assess BRAIN Initiative research risks and promise, provide input on current neuroethics challenges, as well as consider current and future neuroethics opportunities for research.
Neuroethics Guides BRAIN Research
BRAIN currently funds a number of neuroethics research projects to guide the study (and eventual broader use) of neurotechnologies like brain–computer interfaces. For example, neuroethics researchers are studying the impact and use of brain–computer interfaces on people with neurologic disability by using focused discussions, structured vignettes, and video-supported interviews to evaluate user perspectives about the benefits, risks, and ethical challenges of these technologies.
Other teams are studying the impact of neurostimulation on goal setting and personality characteristics. This research is evaluating participants’ views of individual control as a result of non-invasive procedures such as magnetic resonance imaging (MRI)-guided focused ultrasound surgery to treat essential tremor (ET) compared to deep brain stimulation (DBS), an invasive treatment for the same condition. Neuroscientist-ethics teams are also working directly with “BRAIN pioneers,” people who take on significant risk as participants in first-in-human or early neurotechnology studies for the sake of helping to further science. This important research explores how participants, family members, and researchers articulate the role, benefits, and limitations of family and researcher care and support before, during, and after a neurotechnology clinical trial.
We are also very intentional about gathering views from research participants living with disabling conditions, who have unique and valuable perspectives about use of neurotechnologies given the seriousness of their conditions. Last year, I was humbled to share the stage with a research participant at a briefing hosted by the Congressional Neuroscience Caucus. This individual—who was a participant in a BRAIN-funded study conducted at the Baylor College of Medicine—courageously shared how she got her life back from an innovative application of DBS that controlled the debilitating manifestations of her obsessive-compulsive disorder (OCD).
As described in my last blog, Personalized Cures on the Horizon - A Fresh Take on Established Technologies and echoed in this one, we are living in a magical time for neuroscience. Cures are on the horizon, with the opportunity to give people back elements of living that have been stolen by disease. At the same time, it’s essential that we move ahead carefully, listen to what affected individuals and their loved ones think about this research before it begins, and also consider vital but difficult questions about equitable access to such technologies. We will proceed accordingly in the interests of these and all future participants willing to advance our search for treatments and cures.
With respect and gratitude,
John Ngai, Ph.D.
Director, NIH BRAIN Initiative