--- title: "The Tsinghua engineer who spent 20 years teaching paralyzed hands to listen to the brain" date: 2026-10-02 category: Builders site: NeuroAI canonical: https://neuroai.site/a/na-people-hong-bo-neo-bci language: en --- # The Tsinghua engineer who spent 20 years teaching paralyzed hands to listen to the brain > Hong Bo (洪波), a Tsinghua University professor, led the NEO brain-computer interface (脑机接口) — the first approved semi-invasive (半侵入式) BCI medical device in China — from a 2013 idea to NMPA approval in March 2026. A man who had not moved his own hands for fourteen years picked up a cup by thinking about it. The system that translated his intent into motion was built by a team whose leader had been working the problem for two decades. Hong Bo (洪波) is a professor at Tsinghua University's School of Biomedical Engineering and the chief scientist behind NEO (Neural Electronic Opportunity), a brain-computer interface (脑机接口) system that became the first approved invasive BCI medical device of its kind in China. ## A detour into "halfway" implantation Most BCI research splits into two camps. Non-invasive headsets are safe but hear the brain "through a wall" — the skull blurs the signal. Fully invasive implants sit in the cortex and hear clearly, but risk tissue damage, immune response and electrode migration over years. Hong's team chose a third path in 2013: semi-invasive (半侵入式). NEO places its electrode on the epidural surface (硬脑膜外) — outside the brain tissue but inside the skull — and powers and transmits wirelessly with no battery inside the body. The team describes it as hearing "through a curtain" rather than a wall: clearer than non-invasive, safer than fully invasive. ## From idea to approved device The route was long and documented: - **2013** — Hong's team proposes the semi-invasive concept. - **October 2023** — first human implant, at Xuanwu Hospital (宣武医院), in a patient with a complete cervical spinal-cord injury. - **2023–2024** — four feasibility implants establish safety and a target indication. - **2025** — a multi-center confirmatory trial across 11 hospitals completes 32 implants in patients with cervical spinal-cord injury. - **13 March 2026** — China's National Medical Products Administration (NMPA, 国家药监局) approves the device for clinical use, co-developed with the company Boruikang (博睿康). Tsinghua and Xinhua describe it as the world's first approved invasive BCI medical device, while the semi-invasive design is what distinguishes it from fully cortical implants. ## What the numbers say The system uses just eight electrodes but reaches grasp-decoding accuracy above 90 percent, with latency in the hundreds of milliseconds — fast enough that a patient "thinks" and the hand acts. Of the 32 implanted patients, 22 showed significantly improved hand-movement scores after six months of brain-controlled training. The first patient, paralyzed for 14 years, used the implant to drive a pneumatic glove and, later, a brain-controlled wheelchair. Nature listed NEO among its "2025 to watch" science events. ## Why it is a "builder" story Hong is a clinician-scientist, not a typical startup founder, but the arc is a builder's: a 20-year research program turned into a regulated product. He also serves as chief scientist at the Shanghai Lingang Laboratory (上海临港实验室), bridging university research and clinical translation. His own framing is circumspect. After approval he noted the work solves one specific problem — restoring hand function after spinal injury — and that reaching the tens of millions affected by stroke, epilepsy or dementia will require many more teams and many unsolved scientific questions. He describes the next job as going "from 1 back to 0": returning to the basic science of why training changes the brain. ## What makes the profile matter NEO sits in a field suddenly crowded with headline-grabbing implants. Its distinctiveness is not spectacle but engineering restraint: - A design that avoids touching brain tissue, lowering long-term risk. - Wireless power and signal with no internal battery, meant for lifelong use. - Fully domestic software algorithms, core hardware and key materials. - A transparent, multi-center clinical record rather than a single demo. For China's BCI ambition, NEO is the proof that a university lab can carry a device from hypothesis to regulator-approved product. ## Honest limitations Facts — Hong Bo's role at Tsinghua's School of Biomedical Engineering and Lingang Laboratory; the 2013 semi-invasive concept; the October 2023 first implant at Xuanwu Hospital; the 11-hospital, 32-patient 2025 trial; the 13 March 2026 NMPA approval (co-developed with Boruikang); the eight-electrode design, >90 percent grasp-decoding accuracy and hundreds-of-milliseconds latency; 22 of 32 patients improving after six months; and the Nature "2025 to watch" mention — are drawn from Tsinghua University's news site, Xinhua (news.cn), and Nature's events coverage. The "world's first approved invasive BCI medical device" phrasing is the universities' and Xinhua's; we present it as their characterization, noting NEO is semi-invasive (epidural), distinct from fully cortical implants. We have not independently verified the trial's full clinical dataset, and patient-level outcomes are reported by the research team. No RMB amounts appear, so no currency conversion applies. Analysis is current to 2 October 2026. ## What readers can do now - **If you follow BCI**, compare NEO's semi-invasive (epidural, wireless, no-battery) design against fully invasive approaches on the axis of long-term risk, not just signal quality. - **If you invest in medtech**, watch NMPA's post-approval pathway — whether NEO reaches routine hospital use is the real test beyond the headline approval. - **If you study translational research**, treat Hong's "from 1 back to 0" line as a model: a clinical win creates new basic-science questions worth funding. --- Published by NeuroAI (https://neuroai.site/) — https://neuroai.site/a/na-people-hong-bo-neo-bci Free to quote with attribution and a link to the original.