Chinese researchers have developed a flexible electrode technology that could make non-invasive brain-computer interfaces more comfortable and practical for extended use. In reported tests, the system maintained more than 94 per cent accuracy in a visual typing task, while its electrodes supported stable brain-signal recording for several hours.
The development addresses one of the longstanding challenges facing brain-computer interface (BCI) technology: maintaining reliable contact with the scalp without causing discomfort during prolonged use.
The research centres on a new type of hydrogel electrode designed to imitate the gradual changes in stiffness found in human skin. Unlike conventional electrodes, which are often either relatively rigid or uniformly soft, the new design combines a softer surface that conforms to the scalp with a firmer outer section that provides structural support and maintains an electrical connection.
The researchers integrated the electrodes into a commercially available electroencephalography (EEG) cap, a wearable device that records electrical activity generated by the brain. According to the findings reported by the South China Morning Post, the system was tested over a 14-hour period while the participant was resting. It also maintained stable EEG recordings for more than eight hours while a human subject carried out everyday activities.
The visual mind-typing experiment achieved accuracy above 94 per cent, demonstrating that the electrode design could preserve usable brain signals over an extended period. However, this result should be understood within the specific experimental task. It does not mean the device can interpret unrestricted thoughts or translate any sentence a person imagines directly into text.
The electrode’s structure is inspired by the layered mechanical properties of human skin. Researchers used a manufacturing process involving the Marangoni effect, which describes the movement of liquids caused by differences in surface tension. By applying a controlled temperature gradient during fabrication, they created a hydrogel with varying levels of crosslinking, producing a gradual change in stiffness and pore structure.
This design also supports the sustained release of electrolytes, which help maintain electrical contact and signal quality between the electrode and the scalp. The researchers associate this feature with the device’s ability to record brain activity consistently over longer periods.
The findings, published in the journal npj Flexible Electronics under the title Skin-mimetic adaptive gradient electrode for ambulatory noninvasive brain-computer interfaces, point towards a potential improvement in wearable brain-monitoring systems.
The implications extend beyond experimental mind-typing. Non-invasive BCIs use sensors placed on the scalp to detect electrical signals without requiring surgery or implanted devices. They are being explored for applications such as assistive communication, rehabilitation, brain monitoring and the development of systems that can help people with neurological conditions.
For people who have lost the ability to communicate or control certain movements, more comfortable and reliable brain-signal recording could eventually improve access to assistive technologies. A system that remains stable during ordinary movement may also be useful for long-term monitoring, including research into detecting unusual brain activity associated with conditions such as epilepsy.
For healthcare providers, improved electrode performance could reduce some of the practical difficulties associated with repeated monitoring sessions. More comfortable equipment may make it easier for patients to tolerate longer recording periods, while stable signals could help researchers develop monitoring systems that operate outside highly controlled laboratory environments.
The technology could also attract interest from companies developing wearable medical devices, rehabilitation equipment and human-computer interaction products. If further studies confirm its performance, the electrode design could provide a useful component for future consumer and clinical BCI systems.
Nevertheless, the findings do not establish that the device is ready for widespread medical or commercial deployment. Demonstrating stable signals in a controlled experiment is different from proving that a technology can work reliably across large and diverse patient populations. Researchers will need to investigate performance across different scalp conditions, movement patterns, environments and extended periods of repeated use.
There are also broader challenges facing non-invasive brain-computer interfaces. EEG signals can be affected by movement, electrical interference and differences between individuals. Translating recorded brain activity into accurate commands requires specialised signal-processing systems, and performance can vary depending on the task and the person using the equipment.
The reported 94 per cent typing accuracy is therefore an encouraging research result, not evidence that all users can achieve the same performance in everyday situations. Further independent testing and clinical evaluation will be important before the technology can be considered suitable for routine patient care.
For investors and technology companies, the development illustrates how progress in the BCI industry depends not only on sophisticated algorithms but also on improvements in the physical hardware used to capture brain activity. More comfortable sensors, longer recording times and dependable signal quality could help move brain-computer interfaces closer to practical applications.
The latest research offers a promising approach to one of the industry’s central engineering problems. Whether it can translate into widely adopted medical and consumer products will depend on further validation, manufacturing scalability, cost and the ability to maintain performance outside controlled testing environments.
