Speaking By Thinking. How Ai And Bmi Help Paralyzed Patients  

Tuesday, October 31, 2023
Speaking without the aid of the body In September 2023, a breakthrough in medical news made headlines: two decades after suffering a stroke, a paralyzed woman had, thanks to a brain-computer interface (BMI) and artificial intelligence (AI), regained the ability to speak. Nevertheless, it's important to note that she wasn't cured; scientists had found a way to capture brain signals, interpret them, and convey them through a digital avatar. The research project results, carried out by researchers from the University of California, San Francisco, were recently presented in Nature ("A high-performance neuroprosthesis for speech decoding and avatar control"). At the age of 29, Ann experienced a brainstem stroke, which left her unable to speak. While conventional medicine offered no hope of her regaining her natural ability to speak, recent advancements in biomedical engineering provided assistance. Ann willingly participated in a study where scientists implanted a flexible, credit card-sized matrix with 253 electrodes directly onto her cortex, precisely where the brain generates and transmits electrical signals responsible for speech to the jaw, facial muscles, larynx, and tongue. Signals captured by the implant were transferred to a computer system via a port located on her head.
BMI and AI helped a paralyzed woman speak through thought 20 years after a stroke
In the initial phase, scientists developed and trained artificial intelligence algorithms to interpret these brain signals. Ann repeated words from a 1,024-word dictionary until the AI system accurately correlated electrical signals with words, or more precisely, phonemes (subunits of words). This approach significantly sped up speech interpretation, allowing the system to recognize 80 words per minute. Researchers took it a step further by replicating Ann's own voice using a speech synthesis algorithm based on a voice sample recorded prior to her paralysis. Her avatar not only spoke but also synchronized facial muscles, mouth, and lip movements. When Ann thought of words, her avatar acted as her personal translator. This had an emotionally profound impact on her, as she was able to communicate once more. Although the technology is currently in the clinical trial phase, researchers aim to provide assistance to paralyzed patients in the future. The primary challenge is developing wireless technology for transmitting brain signals to the IT system to make the solution more convenient for patients.

Will machines soon be able to read minds?

The development of brain-computer interfaces (BCIs) has a long history, with the first implantation of electrodes into the brain occurring in the 1990s. The concept of BCIs dates back to the 1940s and 1950s, but significant progress was only made with the advent of artificial intelligence systems. Notably, Stephen Hawking, who passed away in 2018 after being diagnosed with motor neuron disease, communicated with the assistance of a computer, which translated his cheek movements into voice output rather than directly interpreting signals from the brain. Ann's case represents a significant breakthrough, not only because of the technology itself but because of the rapid translation of signals into speech. Previous solutions achieved a rate of approximately one word every 20 seconds, severely limiting communication capabilities. Scientists remain cautious about their enthusiasm, as there are still many challenges to overcome. However, Elon Musk, the founder of Tesla, holds a different perspective. Neuralink initiated the first clinical trials of a brain-computer interface on individuals with quadriplegia caused by cervical spinal cord injuries or amyotrophic lateral sclerosis (ALS) in September of this year. While Musk declares his primary aim is to assist paralyzed patients, his vision extends far beyond this. In the future, he envisions that every individual will be able to communicate with computers and artificial intelligence without the need for a keyboard, significantly expanding cognitive abilities. To this end, Neuralink has developed a specialized surgical robot for implanting BCIs in the brain. The planned clinical trial will assess the safety and effectiveness of the implant (N1) and the surgical robot (R1), as well as their ability to enable paralyzed individuals to control external devices with their thoughts. Implants do have significant limitations due to surgical complications. Therefore, researchers are also exploring wireless alternatives. Instead of implants, methods like EEG, fMRI (functional magnetic resonance imaging), and fNIRS (functional near-infrared spectroscopy) are being used. Early experiments suggest that it is possible to read thoughts by measuring brain activity in this way and interpreting it with the help of AI algorithms. Nevertheless, this technology faces the limitation of requiring large brain imaging equipment.

Artificial intelligence will enhance cognitive capabilities

In the past year, scientists have encountered another puzzle that could expedite their quest to develop safe, non-invasive, and effective BMIs: generative artificial intelligence. This type of AI can create text, music, and images based on simple descriptions, known as prompts. While it's still too early to provide specifics, paralyzed individuals might be able to perform various tasks, such as creating art or writing books, simply by controlling AI tools like ChatGPT with their thoughts. Currently, efforts primarily focus on simpler solutions, such as controlling smart prostheses, exoskeletons, or robots to assist patients in regaining independence, often called neuroprosthetics. Additionally, solutions based on neurofeedback hold promise for treating conditions such as depression, ADHD, epilepsy and personalizing rehabilitation for stroke patients. Neuralink is not the sole contender in the quest to lead in the development of BMIs. CTRL-Labs, a subsidiary of Meta (formerly Facebook), aims to create applications in the entertainment industry, including virtual reality-based games and the metaverse. One question remains unanswered for now: are the commercial companies investing billions of dollars in BCI technology primarily interested in creating a new type of human integrated with a computer? Until that question is resolved, BCIs offer great hope for paralyzed patients to regain the gift of natural communication and independence.