Caixin Weekly | Brain-Computer Interfaces Connected to Reality (AI Translation)
Listen to the full version


文|财新周刊 徐路易
By Xu Luyi, Caixin Weekly
文|财新周刊 徐路易
By Caixin Weekly's Xu Luyi
51年前,美国加州大学洛杉矶分校计算机科学教授雅克·维达尔在《科学》杂志上发表了一篇划时代的论文《迈向脑机直接通讯》,提出“脑机接口”(Brain-Computer Interface,BCI)这一概念。他采用脑电图(EEG)信号来进行实验和研究,论文介绍了在人类大脑皮层捕捉、记录到的脑电波的各种特性,并探讨了计算机在处理和分析EEG数据方面的应用,期望实现脑机直接通讯。
Fifty-one years ago, Jacques Vidal, a computer science professor at the University of California, Los Angeles, published a groundbreaking paper titled "Toward Direct Brain-Computer Communication" in the journal Science. Vidal introduced the concept of the "Brain-Computer Interface" (BCI), using electroencephalogram (EEG) signals for his experiments and research. The paper detailed various characteristics of brainwaves captured and recorded from the human cerebral cortex and explored the applications of computers in processing and analyzing EEG data, aiming to achieve direct brain-computer communication.
看到屏幕上突然亮起一个光点,大脑会对此视觉信息做出反应,这种反应体现在人脑电信号上,就形成了视觉事件相关电位(ERP),即大脑针对某个视觉刺激专门发出的一种独特的“电信号回应”。
When a dot of light suddenly appears on the screen, the brain reacts to this visual information. This reaction is reflected in the brain's electrical signals, forming what is known as a visual event-related potential (ERP), which is a unique "electrical signal response" the brain emits in response to a specific visual stimulus.

- DIGEST HUB
- Jacques Vidal introduced the Brain-Computer Interface (BCI) concept in 1973, paving the way for modern advancements. This technology now allows paralyzed patients to regain motor functions by controlling devices via thought.
- BCI has a promising market, projected to grow from $2.35 billion in 2023 to $10.89 billion by 2033. Applications include treating disorders, improving memory, and enabling complex task controls.
- Xu Minpeng's team has led advancements in non-invasive BCI and developed systems with up to 216 instructions, aiding in neurological rehabilitation and enhancing BCI system effectiveness.
The concept of Brain-Computer Interface (BCI) was pioneered by Jacques Vidal in 1973, exploring the direct communication between the brain and computers using EEG signals. The potential of BCIs expanded over the years, with significant advancements in both non-invasive and invasive technologies, having applications in restoring motor functions and treating disorders [para. 1]. Initially focusing on electroencephalogram (EEG) signals, today’s researchers use various BCI technologies to assist paralyzed individuals and treat brain disorders [para. 3].
Advancements in BCI technology include non-invasive methods that require no surgery, using electrodes attached to the scalp to collect brain signals. This approach dominates the market due to its simplicity and safety, though skull conductivity issues present challenges in reading clear signals. Innovations by researchers like Xu Minpeng, who experimented with new paradigms to enhance signal clarity, exemplify cutting-edge non-invasive exploration [para. 9][para. 11]. Xu's work highlights a novel and highly precise method for capturing very weak brainwaves, thus advancing BCI capabilities. His team demonstrated their application by controlling drones and large instruction sets in technological advancements [para. 13][para. 15].
However, the complexity of BCI technology and individual variations in EEG signals present significant challenges. Xu is working to create a universal communication protocol to integrate the latest BCI technologies seamlessly, akin to HTTP for web interactions [para. 19]. At the intersection of artificial intelligence and BCIs, innovative platforms aim to provide standardized tools and frameworks for further research and application development [para. 21].
Invasive BCI procedures, on the other hand, involve more complex methodologies like implanting electrodes directly into the brain tissue. While these methods allow for high precision in capturing brain signals, they come with substantial risks, including surgical complications and potential long-term impacts. High-profile companies like Neuralink are pushing the boundaries with such technologies, but they also face ethical and safety scrutiny [para. 45][para. 47]. Trials like Neuralink’s unprecedented human implant experiments spotlight the richness of potential applications but reveal hurdles like device stability and interaction with brain tissue [para. 57].
Interventional BCI approaches present a less invasive alternative by positioning electrodes within the brain's vascular pathways, thus reducing the need for craniotomies and lessening infection risks. These minimally invasive methods capture clear neural signals and demonstrate promise in animal and preliminary human trials, offering potential breakthroughs in the field [para. 71][para. 75]. Research teams in China, including Duan Feng’s at Nankai University, are making crucial strides with minimally invasive innovations, indicating an evolving landscape for practical BCI applications [para. 87].
Another promising area is the semi-invasive technology, which implants electrodes on the meninges, offering superior signal clarity without directly contacting neurons. This method shows great promise in aiding individuals with paralysis, as demonstrated by Hong Bo’s team at Tsinghua University [para. 99][para. 101]. The accomplishments of patients like Mr. Yang underline the potential of semi-invasive methods in restoring meaningful mobility and sensory functions, marked by notable improvements in functional scores [para. 107][para. 109].
Overall, BCIs represent a converging point of various scientific disciplines, from neuroscience to computational engineering. Continued dialogue and coordinated efforts across sectors are indispensable for addressing the technical and ethical complexities in this promising field. As developments push forward, the trajectory suggests that practical, transformative BCI solutions might one day become integrated into everyday assistive technologies, enhancing the quality of life for individuals with disabilities [para. 111][para. 113].
- Brain Co Medical Technology Co., Ltd.
- The article mentions that Brain Co Medical Technology Co., Ltd. is jointly developing a high-channel electrode array and external machine with Tsinghua University, used in a semi-invasive brain-computer interface trial. The partnership focuses on advancing medical-grade BMI systems for clinical applications, enhancing rehabilitation for paralyzed patients.
- Neuralink
- Neuralink, founded by Elon Musk, has developed an invasive brain-computer interface. Despite initial setbacks, including a rejected FDA application and technical issues during its first human trial, Neuralink received approval in May 2023 for testing with paralyzed patients. The company aims for precise neural signal recording and has developed flexible electrodes to minimize brain tissue damage, using robotic surgery for accurate implantation.
- Synchron
- Synchron is a U.S.-based company working on brain-computer interfaces (BCIs) to aid patients with movement disorders. It uses a minimally invasive approach by implanting sensors via blood vessels. Synchron's "COMMAND" clinical trial in the U.S. started in 2022, following approval in Australia, with participants showing good signal stability and the ability to control digital devices. Their method reduces infection risks associated with traditional invasive procedures.
- OpenAI
- The article mentions that in 2024, the U.S. company Synchron, specializing in brain-computer interfaces for restoring motor functions in patients, announced the integration of OpenAI-driven generative artificial intelligence into its brain-computer interface platform.
- 1969:
- Eberhard Fetz used tungsten microelectrodes to penetrate the cerebral cortex of monkeys, paving the way for the development of invasive brain-computer interfaces.
- 1973:
- Jacques Vidal published the paper 'Toward Direct Brain-Computer Communication' introducing the concept of Brain-Computer Interface (BCI).
- January 2020:
- Xu Minpeng's team unveiled a brain-controlled drone system capable of continuous control across four degrees of freedom using a brain-computer interface.
- November 2020:
- Xu Minpeng's team released a BCI system with 108 instruction sets, marking the world's first system to surpass 100 instruction sets.
- February 2021:
- Neuralink disclosed that a test monkey died during prototype testing of brain-machine interface implants.
- 2021:
- Xu Minpeng was awarded the National Excellent Young Scientists Fund.
- May 2023:
- Xu Minpeng's team unveiled a next-generation high-speed non-invasive brain-computer interface system, achieving a record of 216 instruction sets.
- January 2024:
- Xu Minpeng's team launched Meta BCI, an open-source software platform for brain-computer interfaces.
- March 2024:
- Xu Minpeng's doctoral advisor, Ming Dong, presented a proposal at the Two Sessions to accelerate the development of brain-computer interface technology in China.
- PODCAST
- MOST POPULAR