Artificial-intelligence systems have recently been hailed as a groundbreaking phenomenon for their ability to generate human-like speech from troves of data. Now AI has gone one step further: An AI-powered brain implant helped achieve a near miracle—capturing the electromagnetic signals of a patient paralyzed by stroke and translating them into digital words and facial expressions.
This otherworldly scenario is a culmination of decades of research in neuroscience and significant advances in modern computing. It is the latest example of the remarkable potential of human-machine interfaces—complex mechanisms at the intersection of technology and medicine designed to help people with paralysis live more freely as well as enable doctors to treat serious neurological diseases and impairments, such as amyotrophic lateral sclerosis (ALS) and spinal-cord injury; and augment human cognitive powers.
“For decades, technology has worked to bring computing to the level of human cognition, making processors, chips and other critical components progressively smaller, faster and safer along the way,” says Ed Stanley, Morgan Stanley’s head of thematic research in Europe. “The recent breakthrough with brain-computer interfaces (BCIs) could be a significant milestone toward commercially-available implantable devices that can restore sensory and movement functions and, perhaps someday, enhance human cognition.”
Roughly 1.3 billion people, or 16% of the global population, live with moderate to severe levels of disability associated with underlying health conditions and impairments, according to the World Health Organization. In addition to health inequities and discrimination, the limitations these individuals face often extend beyond day-to-day functioning and can constrain or can altogether close off access to necessary care, education and employment.
BCIs have the potential to be life-changing for this population. Early iterations of BCIs have been in development since the 1960s, and more recent technologies in the realm of so-called human-machine interfaces have aimed for more short-term solutions, such as helping people with quadriplegia use computers and mobile devices to live more independently.
But the recent AI breakthrough, which captured human thought and interpreted it into intended action and emotion, required coordination among biology and machinery that previously had no mode of connection. Scientists connected a network of more than 250 tiny electrodes onto the area of the brain associated with speech and facial expression to intercept signals meant to command muscle movements—for instance, to render a smile or show surprise. To bring the process full circle, the team used a novel neural network trained to predict speech by breaking the brain signals down into sounds instead of full words, helping the paralyzed patient deliver speech—animated through a digital avatar—at a rate of about 78 words a minute. Two years ago, the same group of researchers achieved 50 words per minute with a different patient using a similar but less complex implanted device. For reference, conversational English is roughly 150 words per minute.
A Long Path to Commercialization
A handful of small companies, not-for-profit organizations and universities are developing invasive and non-invasive versions of BCIs. Funding in the space is relatively sparse and significant technological hurdles remain, including miniaturization of electrodes and other key components. However, reports of human trials in the U.S. for a BCI-to-app interface to help paralyzed people control keyboard movements could draw more money to the space.
“Historically, direct neural stimulation has led to remarkable clinical improvements across a range of conditions, particularly in pain and movement disorders like Parkinson’s disease,” says Patrick Wood, equity analyst covering medical technology. “A more direct cerebral interface could open a host of new treatment avenues, though dramatic regulatory, ethical and clinical trial hurdles remain before this kind of tech is ready for prime time.”
Ethical concerns may well be the largest barrier, especially as algorithms rapidly improve and open the possibility of two-way communication between a human brain and the external world—leaving patients vulnerable to stimulus from the AI that is meant to be taking direction from them.
The long-term panacea via BCIs would be to find non-medical applications and create machine interfaces that are safe and powerful for the general population, to expand how humans interact with one another and the world—potentially replacing our everyday tech hardware like smartphones. “This appears still some way from reality,” says Stanley.
For a more detailed look at disruptive and ambitious technologies that could bring generational returns, ask your Morgan Stanley representative or Financial Advisor for “Moonshots” (Sept. 14, 2022).