A team from the University of California, San Francisco (UCSF) has demonstrated the first “plug and play” brain phantom in paralyzed patients. Using machines to decode the brain’s electrical signals opens up more possibilities for future medical functional repair – such as improving Alzheimer’s disease screening, monitoring internal organs, and enabling paralyzed patients to regain control of prosthetic limbs. After being implanted into the brain in various forms, and powered by advanced algorithms, researchers can convert brain electrical signals into control inputs for various devices, such as prosthetics, complete exoskeletons, and even UAVs. The new technology developed by UCSF marks an important step forward in this field. It focuses on translating brain activity into software actions, and then trains with algorithms of machine learning. < / P > < p > by having paralyzed patients see cursor movements on the screen, they can track their imaginary neck or wrist movements. After the algorithm is continuously reset periodically (daily), the software can gradually learn the movement actions that match the user’s assumption. Although it takes several hours to experiment every day, this scheme can ultimately achieve the desired control. In addition, scientists are actively improving and exploring other applications. For example, adjust the algorithm to avoid daily training from the beginning. After continuous improvement, the final algorithm can allow users to immediately access and start using. Karunesh Ganguly, a practical neurologist at UCSF health, said: < / P > < p > by ensuring that algorithms are not updated faster than the brain can track, we can further improve the speed of machine learning, which is about every 10 seconds. < / P > < p > we believe that this is the establishment of a cooperative relationship between the two systems of brain and computer, which can eventually make the human-computer interface an extension of the user’s ability to freely control a mechanical prosthesis (hand or arm). < / P > < p > the brain computer interface (BCI) used in the experiment is called ECoG array, which consists of a pad of electrodes the size of a sticky note paper, which can be surgically implanted into the surface of the brain. < / P > < p > the results showed that even without any routine calibration, the performance of the system did not decrease after 44 days. Even if the user is down for a few days, the performance will only decline slightly. Karunesh Ganguly, M.D., senior author of the study, points out that over time, the user’s brain also optimizes its own activity to better control BCI and avoid daily recalibration. The details of this study have been published in the recently published journal Nature Biotechnology. Chinese version of K-car: reading a10e design drawing exposure