Neuralink Developing Fully Implanted, Wireless Brain-Machine Interface (BMI) To Help People With Paralysis To Operate Computers

Neuralink, a neurotechnology company founded by Elon Musk in 2019, has developed technology to better capture human brain function. Starting with tests on rodents, they created a robotics-based surgical solution using wired leads and a connector through the skin. The tests used ultra-low-power custom application-specific integrated circuits (ASICs) to magnify and manipulate neural signals. 

In 2020, Neuralink released another update on a wireless version of this connection that can stream 1,024 channels of action potentials (nervous activity surges) in real-time. Researchers demonstrated the new connection’s capabilities by observing somatosensory (touch) signals in pigs as they explored their surroundings. The electrodes were mounted on the part of the brain that processes signals from the pig’s hypersensitive snout. Researchers were able to detect neuronal responses to sensory cues when studying the pig in its natural habitat.

Although working on pigs has helped show how sensory neurons function, researchers discovered that to modify these electrodes for human use, they wanted to test them on an animal with a neural topology similar to that of humansβ€”primates. The aim is to create a brain-machine interface (BMI) technology used for the arms and hands. This kind of closed-loop BMI is designed to make it easier for people with neurological disabilities to explore their surroundings.

Recently, the macaque variant of BMI has been used to research neuronic activity in macaque monkeys.Β

N1 links are two connections – one on the left side and another one on the monkey’s motor cortex’s right side. These cortices work together to interpret sensory stimuli such as touch and visual signals. For the most part, the technology uses these strategic positions to anticipate the subject’s planned movements. A model is created (i.e., “calibrate a decoder”) to determine the course and pace of an upcoming or expected movement by modeling the interaction between various patterns of neural activity and intended movement directions. This will go beyond merely forecasting the most likely intended action based on existing brain function patterns: by using these projections to monitor the motions of a screen mouse, or in the video below, a MindPong paddle, in real-time.

This demonstration was a small but significant step toward giving paralyzed people complete neural control of a virtual cursor.

Scientists are now working on a decoder that can connect a user’s desired movement and path to a computer cursor. The sensors used to process direction changes tend to be neurons fired in response to the subject’s action in that direction. The researchers noticed this after seeing the monkey use a MindPong, a paddle that was provided to the monkey to observe the neuron-to-sensor activity in action.

However, since paralyzed individuals may be unable to move any part of their body, experts are focusing on creating technologies that can read brain activity and interpret even simple human gestures.



🐝 Join the Fastest Growing AI Research Newsletter Read by Researchers from Google + NVIDIA + Meta + Stanford + MIT + Microsoft and many others...