The "Triple Gate" of Brain-Computer Interface Electrodes

Apprentice Reporter Yin Jingfei

A single electrode is increasingly becoming the “Battle of the Bulge” in the brain-machine interface industry. While the outside world is eager to compare channel counts and chip sizes, a more fundamental question remains unresolved: How long can this device implanted into the brain actually last? Before addressing “how long it can be used,” there is an even more pressing issue: what material should it be made of? In the invasive approach, after craniotomy, how to collect data, and how deep should it penetrate? The industry is actively exploring and seeking solutions.

The Rigidity Dilemma: Materials Must Be Soft

“The brain is as soft as tofu, but traditional metal or silicon-based probes are particularly hard. When inserted, they cut through brain tissue at a micro-scale, and with breathing, they drift, making stable signal tracking impossible. More seriously, hard materials can trigger immune rejection, leading to neuron death at the implant site, and the originally clear signals will eventually disappear,” industry insiders told Securities Times.

“Materials must be soft” has become a consensus in the industry. Based on this, two technical paths are diverging: one is to find entirely new soft materials, and the other is to engineer and optimize existing mainstream materials.

Liu Jia’s team chose to develop a naturally soft and tough new material—perfluoroelastomer, which is elastic like rubber and resistant to bodily fluids corrosion. In 2021, Liu Jia and others co-founded Axoft, whose products received FDA breakthrough device designation. Because the material is an elastomer like the brain tissue, it does not produce relative displacement during physiological activities like breathing, fundamentally solving electrode drift and immune rejection issues.

However, this is not the only path. “Currently, most mainstream invasive flexible electrodes abroad and domestically, including Neuralink and Jiatie Medical, use polyimide materials, reducing bending stiffness to achieve ‘physical flexibility,’” said Liu Xiaojun, head of the Brain-Computer Interface Project at the Yangtze River Delta Institute of Future Technologies, Life and Health at Peking University. This approach is the result of decades of laboratory exploration—biocompatibility and conductivity are relatively strong, and supply chains are mature with controllable costs.

Polyimide is not the ultimate solution either. Li Jianfu, Market Director of Shenzhen Weiling Medical, admitted that the effective lifespan of current mature technologies is only about two or three years, and achieving lifelong functionality still requires replacing materials in the brain.

Regarding the polyimide route, various companies are exploring processes and structures. Zhiran Medical has developed stretchable flexible electrodes that can follow the rhythmic movements of brain tissue through strain decoupling. Jiatie Medical has made electrodes at the cellular scale—only 1 micron thick, with a cross-sectional area about one-thousandth of a hair strand. Liu Xiaojun’s team innovated a “Swiss roll” structure, curling a two-dimensional flexible film into a needle shape, capable of integrating 1024 channels into a single needle, balancing high throughput and long-term stability.

Path Dilemma: “Insert” vs. “Attach”

Beyond material disputes, a more fundamental divergence lies in the approach: should electrodes be “inserted” into the brain or “attached” to the surface?

“Neuralink uses deep electrodes, inserting them into the cerebral cortex like hair strands,” said Tao Hu, founder and chief scientist of BrainTiger Technology. In contrast, BrainTiger’s cortical attachment method involves affixing thin-film electrodes onto the surface of the brain cortex. Deep electrodes face two major issues: immune rejection leading to signal attenuation, and physical damage from electrode movement. “We must prioritize patient safety.”

Shenzhen Weiling Medical has gone even further. Its self-developed high-density cortical electrode is only 10 microns thick and can be attached like a film to the uneven surface of the brain cortex. “After insertion, that part of the cortex is basically unusable. If the electrode fails, it cannot be replaced in the patient,” Li Jianfu emphasized. From an ethical standpoint, treatment should not cause secondary injury to patients—that is the bottom line. These “attached” electrodes can be removed with saline rinse, avoiding damage to brain tissue.

Thus, in Li Jianfu’s view, the industry is diverging into two value orientations: one is the “medical camp,” aiming for neural function reconstruction and replacement; the other is the “tech camp,” replicating Neuralink’s path, with features like controlling cursors or wheelchairs. He does not deny the technical difficulty of the tech camp but believes its clinical value is seriously exaggerated.

Risks and Challenges: From Clinical to Widespread Use

“At least in terms of hardware, domestic electrodes’ material technology has caught up with international frontiers,” said Liu Xiaojun.

However, “the last micron” remains a bottleneck.

“The core current challenge is whether the long-term stability of single-neuron signal collection can be guaranteed,” Liu Xiaojun said. The issue is not with the electrical performance of the electrode itself but whether, after implantation, it can reliably “listen” to the firing of individual neurons over months or years. “Animal tests have verified long-term stability, but validation in humans is still lacking.”

This is a huge gap. Animal experiments can last two years with impressive data, but the human brain’s microenvironment is more complex, with immune responses and glial scar formation—factors difficult to fully replicate in animals—that may gradually obscure signals over time. “The difficulty lies in the need for long-term observation, then using the results to iteratively improve materials, design, and surgical procedures,” Liu Xiaojun explained.

Regarding whether the industry is ready for large-scale promotion, experts agree: “The timing is far from ripe.” Most domestic companies plan to start clinical trials after 2025. As a Class III medical device with high regulatory risk, it still has a long road from clinical validation to market approval and widespread adoption.

Li Jianfu clarified misconceptions about surgery. “It’s not like Elon Musk claims that once implanted, the device can be used immediately.” Patients need a very long adaptation process to the presence of a foreign object in their bodies. “From a clinical perspective, one year is the minimum requirement.”

All interviewees agree that “its long-term safety needs time for validation.”

The dilemma is not only with the electrodes themselves. “Neural encoding and decoding also face talent shortages and data silos—brain electrical data are rarely open-source, and individual differences are huge,” Tao Hu added. Hardware is just the tip of the iceberg; beneath the surface lie longer-term collaborations involving algorithms, data, and clinical validation.

“Time will prove who can do it and who cannot,” said Liu Xiaojun.