Multimodal Brain-Computer Interface: A New Opportunity for Chinese Enterprises to Leapfrog in the Global Race

“The brain-computer interface (BCI) industry is currently in a unidirectional development phase, gradually evolving toward bidirectional BCIs and multimodal integration. Multimodal integrated BCIs are poised to become the next major growth track, offering Chinese enterprises an opportunity to overtake competitors on the bend.” On May 28, Zhang Yunyun, Deputy General Manager of CICC Capital, made these remarks at the launch event in Shanghai for the collaborative “Interventional Brain-Computer Interface Rehabilitation System” by OYMotion and Minimally Invasive Brain Science.

Zhang Yunyun introduced that multimodal fusion is a systematic engineering approach that integrates neural signals of different scales and modalities to efficiently decode patient intent and match corresponding stimulation and assistive technologies. It is not merely a simple stacking of signals; its implementation requires novel algorithmic architectures, chip design, and clinical trial systems, relying heavily on supply chain integration.

She stated that, globally, few companies have truly achieved the commercial deployment of multimodal products. Traditional leaders in the brain-computer interface (BCI) sector largely operate in isolation, developing full-stack solutions independently. In contrast, China boasts a vast patient population, rapid response capabilities, and a robust supply chain ecosystem, which can further amplify the industrial advantages of multimodal BCI technology. “The collaboration between OYMotion and Minimally Invasive Brain Science serves as an excellent model for multimodal integration.”

Future Product Concept

It is understood that the two parties have formally established a strategic partnership to jointly develop the “Interventional Brain-Computer Interface Rehabilitation System.” Throughout the collaboration, they will work together to tackle key technical challenges, including hardware-software integration, signal transmission, and user experience optimization, ensuring low latency, high stability, and high precision from neural signal decoding to exoskeleton action execution. The application scenarios will primarily focus on two major directions: first, deepening engagement in medical rehabilitation by providing comprehensive solutions—from rehabilitation training to functional compensation—for patients with conditions such as stroke and spinal cord injury; second, expanding the boundaries of human-computer interaction by exploring the integration of brain-computer interfaces with embodied intelligence devices, thereby uncovering greater potential for human-machine synergy.

At the product level, the first outcome of the collaboration between the two parties will be a multi-channel endovascular brain-computer interface (BCI) system delivered via an interventional approach. Unlike traditional craniotomy-based invasive BCI solutions, this system employs a minimally invasive endovascular route, delivering signal-acquisition stents through blood vessels to target brain regions. It enables the acquisition of high-quality intracranial electroencephalogram (iEEG) signals without the need for craniotomy, thereby balancing clinical safety with signal acquisition precision.

On this basis, the system integrates a lightweight exoskeleton.Robot, translating decoded intracranial neural signals into precise motor control commands to drive exoskeletons in assisting patients with movement execution, thereby bridging the disrupted “intention–execution” pathway in patients with ischemic stroke. Leveraging repetitive active movement training, the system establishes a complete closed-loop framework comprising “intravascular endothelial EEG acquisition → real-time decoding → exoskeleton servo execution → proprioceptive feedback → cortical plasticity remodeling,” facilitating neurological rehabilitation for patients with motor dysfunction.

Zhang Dongdong, Director of the Brain Science Division at OYMotion, stated that while individual signal sources each have their limitations, integrating multiple physiological signals can achieve complementary advantages, significantly enhancing the system's robustness and practicality. Taking the fusion of electroencephalography and electromyography (EEG+EMG) as an example, EEG reflects "intent," while EMG reflects "action execution." Combining the two not only captures the user's motor intent but also acquires residual muscular execution signals, substantially improving command recognition accuracy and response speed.

Hybrid Integration of Invasive and Non-Invasive TechnologiesThe hybrid model integrating invasive and non-invasive technologies combines the dual advantages of high-precision invasive electrodes and the safety and convenience of non-invasive devices. By employing implanted electrodes to capture fine-grained signals in core regions, while supplementing contextual information via peripheral modalities such as EEG, eye tracking, and speech, this approach establishes a “high-precision + broad-coverage” hybrid architecture. This ensures decoding quality while reducing reliance on surgical interventions. The essence of this integration lies in exchanging information redundancy for reliability: multimodal fusion reduces dependence on any single signal, alleviates the trade-off between precision and safety, and accelerates the large-scale adoption of brain-computer interface technology in both medical rehabilitation and consumer markets.

For OYMotion, this strategic collaboration represents not only an expansion of its technological pathways but also marks the company’s transition from non-invasive brain-computer interfaces toward becoming a platform-based BCI enterprise by integrating diverse technological approaches.

This article is an exclusive piece for Guancha.cn. Reproduction without authorization is prohibited.