After a stroke, brain damage leads to limb paralysis and language dysfunction, which seriously affects the patient's quality of life and communication ability. In order to help stroke patients restore their language function, the medical community has been working hard to overcome this problem. In the past, some patients with language dysfunction could communicate with the outside world by typing on the keyboard, but this required them to be able to control part of their limbs, and the typing speed was still too slow, and the communication was not smooth enough. Brain-computer interface technology and artificial intelligence technology have shown great potential in solving this difficult problem, and new research results continue to emerge. Brain-computer interface enables “talking with thoughts” Recently, a team led by Yulia Berezutskaya from the Donders Institute for Brain, Cognition and Behavior at Radboud University in the Netherlands and the University Medical Center Utrecht in the Netherlands published a research report in the journal Neuroengineering, saying that they temporarily implanted a brain-computer interface in some non-paralyzed volunteers. These volunteers were asked to speak out loud 12 given words, and the device recorded their brain activity signals. Image source: pixabay Author: CDD20 Subsequently, the researchers used advanced artificial intelligence to decode the volunteers' brain activity signals, establish a one-to-one correspondence between them and specific words, and convert the signals into speech output. The experiment was successful on these volunteers, and the results showed that this not only accurately identified the words corresponding to the brain activity signals, with an accuracy rate of 92% to 100%, but even more amazingly, the machine-converted speech was very close to the volunteers themselves in intonation and accent. This means that speaking through thoughts will become possible! Image source: piqsels Perhaps one day, people with speech disorders will only need to think about what they want to say in their heads, and the brain-computer interface will be able to capture brain activity signals and instantly output them as complete and fluent sentences through the machine. When ordinary people communicate with them, it will be as smooth as communicating with real people. Yulia believes that the model they have built so far has achieved certain results by using only a relatively small data set to better reveal the complex mapping relationship between brain activity and speech. However, they also pointed out that this model has limitations. Currently, they have only conducted experiments on volunteers who are not paralyzed and have only tested 12 words. Therefore, the scope of application of this model may still be very limited. To achieve the understanding of any word, complete sentence or even paragraph, a larger data set, more advanced brain-computer interface implants and more powerful artificial intelligence language model research and breakthroughs are needed. Opportunities and Challenges of Brain-Computer Interfaces Although there is still a long way to go, Yulia believes that they are moving in the right direction. She said that the ultimate goal of this technology is to enable paralyzed patients who have lost both muscle control and language ability and are completely unable to communicate with the outside world to convert what they want to say into voice output through a brain-computer interface and regain the ability to communicate. Image source: pixabay Author: geralt Technologies such as brain-computer interfaces, big data, and artificial intelligence provide new options for patients with brain diseases such as stroke to restore their ability to communicate. With further development of technology, we believe that they will be able to accurately decode any brain activity in the future, allowing more patients who have lost their language ability to regain the ability to communicate through voice, greatly improving their quality of life, and benefiting more ordinary people. At the same time, we also see that these technologies still have many unresolved problems, such as brain implants may cause intracranial infections, and big data and artificial intelligence technologies involve risks to citizens' privacy and information security. We hope that these problems can also be properly resolved in the future. References [1] https://www.hackster.io/news/brain-implants-and-artificial-intelligence-turn-thoughts-into-intelligible-speech-researchers-prove-96af60962d62 [2] http://news.eeworld.com.cn/medical_electronics/ic650064.html [3] Julia Berezutskaya et al 2023 J. Neural Eng. 20 056010 Planning and production Source: "Voice of Science and Technology Association" WeChat public account Author: Zeng Xinyue, a popular science writer Editor: Jin Yufen |
<<: The strange structure of the elephant's body can actually prevent cancer? | Nature Trumpet
>>: Mianyang, Sichuan, don’t be so low-key!
Q: I want to place Baidu information flow ads. Ca...
Following the popularity of the chicken-eating ga...
The way to play Douyin is constantly changing, an...
1. The show begins On the evening of November 10,...
"Mushrooms contain heavy metals and cannot b...
Author: Chen Wen Reviewer: Department of Neurolog...
Everyone probably has their favorite. A special s...
Planning and production Planning丨Ding Zong Visual...
In 1865, on the flat land outside Yongdingmen in ...
Social media giant Facebook seems to be really go...
As a winner of Toutiao's Qingyun Plan and Bai...
Although there was the MX4 before and the Meizu B...
There are always people living in a higher dimens...
On July 21, 2022, the World Conservation Union (I...