Sensors and AI restore the ability of ALS patients to express themselves

Sensors with Artificial Intelligence (AI) resources implanted in the brain could restore the ability to express oneself intelligibly in patients with amyotrophic lateral sclerosis (ALS). The technology, developed by researchers at Stanford University School of Medicine in California (USA), transmits signals from certain speech-related regions of the brain to software that decodes brain activity and converts it into text displayed on a computer screen.

The solution has already been tested on a 68-year-old patient, a former human resources director who practiced horse riding every day. She was diagnosed with ALS, a progressive neurodegenerative disease that attacks the neurons that control movement. However, in some ALS patients, the disease can lead to speech difficulties. The patient in question can no longer use the muscles in her lips, tongue, larynx and jaw to pronounce phonemes clearly.

In a surgery, two small sensors were implanted in the patient’s cerebral cortex, the outermost layer of the brain. The sensors were located in two different regions, but both worked on speech production. They are made up of square arrays of tiny silicon electrodes, and each array contains 64 electrodes, arranged in 8-by-8 grids and spaced about half the thickness of a credit card. The arrays are attached to thin gold wires that come out of the skull and are connected to a computer.

These sensors act as an interface between the brain and the computer (iBCI – Intracortical Brain-Computer Interface). Combined with decoding software, they can translate the brain activity that accompanies the patient’s speech attempts into written words.

The AI algorithm receives and decodes the electronic information coming from the patient’s brain, learning to distinguish the distinct brain activity associated with attempts to pronounce 39 phonemes that make up the English language. To train the algorithm, the patient underwent around 25 sessions, each lasting around four hours, in which she tried to repeat randomly chosen phrases taken from samples of conversations between people talking on the phone.

About a month after the surgery, the researchers began twice-weekly sessions to train the software that interpreted the patient’s speech. After four months, the speech attempts were converted into written words at a speed of 62 words per minute, more than three times faster than the previous record for BCI-assisted communication, according to the group of scientists.

“The patient’s rate of speech is starting to approach the natural speed of conversation, of around 160 words per minute among English speakers,” says surgeon Jaimie Henderson, adding that this shows that it is possible to decode speech by recording the activity of a very small area of the brain’s surface.

In 2021, Henderson and other researchers co-authored a study on the successful conversion of a paralyzed person’s imagined handwriting into text using an iBCI, reaching a speed of 90 characters, or 18 words, per minute, a world record until then for an iBCI methodology.

The results of the current research have been published in an article in Nature. The researchers warn that the work was a scientific proof of concept and is not a real device that people can use in everyday life, but it is a breakthrough in restoring rapid communication between people with paralysis who cannot speak.

Medical sensors market

The value of the global medical sensors market is expected to reach US$7.5 billion by 2032, growing at a compound annual growth rate of 13.4% during the forecast period, according to a report by Emergen Research.

This class of sensors is essential in modern healthcare, which remotely monitors the vital signs of patients with chronic diseases, such as heart rate, blood pressure, temperature, glucose levels and respiratory rate, among others, all in real time in order to facilitate the diagnosis of various types of medical conditions. In addition, the growth of the elderly population and the greater demand for home healthcare are driving the adoption of medical sensors. Advances in sensor technologies, such as miniaturization, wireless connectivity and wearable designs, have also contributed to the growth in demand.

According to the survey, this market is already fairly consolidated, with a few players accounting for most of the revenue. Various strategies are underway, such as mergers and acquisitions, strategic agreements and contracts around product development, testing and launch processes.

One of the main challenges in the medical sensor market is the strict regulatory standards that seek to guarantee safety, precision and reliability. In addition, issues involving interoperability and a lack of standardization in data formats can limit the efficient use of sensors for monitoring and decision-making. Security and privacy are other concerns surrounding the data generated by medical sensors. This scenario can impose some barriers to entry for new players and increase the time and cost of product development.