Earlier this year, we wrote about a brain-computer interface (BCI) developed by Stanford University researchers that enabled a paralyzed individual to turn his thoughts to written sentences. The Stanford BCI platform starts with chips implanted in the patient’s brain. The chips include 100 electrodes that sense hand movement signals in the motor cortex. A computer receives and analyzes the signals, decoding them to produce letters or characters. Salt Lake City-based Blackrock Neurotech recently announced plans to commercialize the Stanford BCI platform in 2022.
The Blackrock platform employs five machine learning-driven decoders to convert the incoming signals from the patient’s brain. The decoders can process thought patterns that represent symbols, virtual keyboards, and handwriting movements. In the simplest terms, if the patient thinks of the letter A, thinks of typing it on a keyboard, or thinks of handwriting the letter, the decoders recognize and pass on the data to create an A on a computer. In testing at Stanford a patient using the BCI was able to create up to 90 characters per minute with 94% accuracy boosted to 99% when run through auto-correction algorithms. As a point of comparison, the best computer text-to-speech conversion services are less than 86% accurate.
Blackrock’s goal is to help patients with ALS, paralysis, and spinal cord injuries regain lost ability to communicate through text. Blackrock also intends to make the BCI platform available for further research. According to Blackrock, the BCI technology could also equip patients with the implants to control screen cursors, chairs, vehicles, robotic arms, and keyboards.
