Brain computer interface Startup Paradromics announced today that surgeons successfully inserted the company's brain implant into the patient and safely removed it after about 10 minutes.

This is a step towards longer equipment testing, called connexus. This is also the latest business development in a growing number of corporate fields, including Elon Musk's Neuralink, which connects people's brains directly to computers.

With Connexus, the Austin-based paradox is seeking to restore speech and communication in patients with spinal cord injury, stroke, or amyotrophic lateral sclerosis, also known as ALS. The device is designed to convert neural signals into comprehensive speech, text and cursor controls. The paradigm was founded in 2015 and has been testing its implants in sheep for the past few years. This is the first time it has used the device in human patients.

The procedure was held at the University of Michigan on May 14 and is performed by people who are undergoing brain surgery to treat their epilepsy. The patient agreed to temporarily insert the Connexus device into its temporal lobe, which processes auditory information and encodes memory. To implant the device, the surgeon used an Epipen-like instrument developed by Paradromics. The researchers were then able to verify that the device could record electrical signals from the patient's brain.

"When someone is doing a major neurosurgery, there is a very unique opportunity," said Paradromics CEO Matt Angle. "They will open the skull and there will be a brain that is about to be removed. In this case, the marginal risk of testing brain implants is actually very low."

Paradromics' implant is smaller than a dime size and has 420 small protruding needles pushed into the brain tissue. These needles are electrodes that record individual neurons. Similarly, Neuralink’s implants are also located in brain tissue. (In comparison, it has more than 1,000 electrodes distributed in 64 thin, flexible threads.) Other BCI companies have fewer invasive methods. For example, precise neuroscience is testing implants that exist on the surface of the brain, and the synchronizer has developed a device that is housed in blood vessels and leans against the brain. Both devices collect signals from groups of neurons, rather than individual devices.

"By getting close to a single neuron, you can get the highest quality signal," Ang said. Getting high-resolution signals from the brain is important for accurately decoding a person's expected speech.

BCIS will not directly "read" a person's personal thoughts. Instead, they work by interpreting neural signals associated with motor intention. For example, like a Paradromics, BCI will decode the facial movement involved in speaking. A paralyzed person who cannot move his mouth can still try to make this movement produce unique neural signals in the brain. These signals are then decoded into speech.

In 2023, a team from Stanford University and UC San Francisco reported significant progress in voice decoding using BCIS. Among two paralyzed women, brain implants were able to decode expected speech at a rate of 62 and 78 words per minute. For comparison, people speak in about 130 words per minute.