Researchers develop a thought-writer and a brain machine science time

Illustration of a human brain | reproduction

The product mentioned in the title is not available in stores, one of those apps that write imposed sentences. I’m talking about brain-machine interfaces, which are attracting great interest from researchers due to the possibility of treating patients with disabilities in different functions, and transferring their ideas directly to a written basis. Think about how difficult it can be for someone with a spinal injury to make simple movements. Worse yet: imagine how painful it is when brain damage also prevents communication with others, due to motor difficulties with speech. A striking example is “The Escafandro ea Butterfly”, written through the eye movements of the paralyzed journalist Jean-Dominique Pope, who indicated yes or no to every letter given to him by his caregiver.

OK then. Nature reports a milestone in the development of brain-machine interfaces capable of writing at the command of the brain. The device was tested on a patient with severe brain damage, affected by its speed and accuracy. The interface, which is controlled by the patient’s brain, can type 90 characters per minute with an accuracy of 95%. It’s close to the hundred odd we get by typing on my cell phone—and with fewer errors than I made when writing the first version of this text on my notebook.

It works as follows: two small plates with a hundred very fine wires are placed on the surface of the patient’s brain, exactly in the area where the hand movements are programmed. The fine wires pick up the code signals of 200 neurons, leading to a computer capable of learning to recognize the neural commands of each letter of the alphabet and writing symbols, seen by the patient.

From the start, researchers have found that connected writing movements were faster and easier to decode than capital letters: the pen doesn’t take as much time out of paper as it does offline writing. From letters, it evolved into words and phrases: a classification algorithm was used to identify each sentence the patient imagined as going to be written by hand.

What then appeared on the screens was a handwritten version of the ideas the patient wanted to communicate. Even after nine years of complete paralysis, he kept “movement programming” engraved in his brain to write the twenty letters of the alphabet.

Obviously, the big problem with brain-machine interfaces is their invasive nature, as they require a brain implant. Thus, the benefits should outweigh the risks, which is what the researchers are betting. Also, for everyday use portability is a must. The processing power must be increased to accommodate multiple languages. Have you ever thought about the difficulty of translating thoughts into Tamil, one of the languages ​​of South India and Sri Lanka, with 247 letters?

In any case, using these machines to translate ideas directly from the brain into paper is not far behind. Rediscover writing, as the Sumerians did 5,000 years ago.