BCIwiki — Brain-computer interface news, research and industry database

Researchers recorded magnetoencephalography (MEG) from 18 right-handed participants as they imagined sounds and pictures, then compared two decoders: a convolutional neural network (CNN) and a linear logistic regression model. The CNN decoded both tasks above chance and exceeded 70% accuracy for visual imagery. It still decoded significantly when trained only on cortical regions unrelated to the task, suggesting imagined content is spread across partially overlapping networks rather than confined to a single sensory area, an experimental basis for feeding auditory and visual information into BCI decoders together.

Microsoft and INBRAIN Neuroelectronics said they will work together on AI-driven brain-computer interface therapeutics, pairing INBRAIN's graphene-based neural interfaces with Microsoft's cloud computing and data infrastructure to speed development of treatments for neurological disorders. INBRAIN, a Spanish neurotechnology company, says its system combines graphene electrodes with AI algorithms to read and modulate brain signals with high precision, on a platform it describes as bringing together neuroscience, bioelectronics and AI.

INBRAIN Neuroelectronics has announced a partnership with Microsoft to apply the software company's large language models and data analytics tools to make its brain-computer interfaces adapt in real time. INBRAIN builds graphene-electrode BCIs for neurological disorders, spanning invasive and non-invasive applications, and recently partnered with the Mayo Clinic to speed clinical development and commercialization of its BCI therapies, with investigator-led clinical research under way.

Barcelona-based INBRAIN Neuroelectronics said it will work with Microsoft on AI for neurology and brain-computer interfaces, combining INBRAIN's neuroscience, bioelectronics and AI work with Microsoft's cloud and AI infrastructure. INBRAIN's graphene-based neural interface is being developed as an adjunctive therapy for Parkinson's disease; it received FDA breakthrough device designation in 2023 and entered its first-in-human study in the fall of 2025.

The Chinese Association of Rehabilitation Medicine will hold a workshop on EEG-driven closed-loop TMS in Guangzhou, southern China, on July 17, 2026, staged under its Naoke China (脑客中国) series and run by its committee on brain function monitoring and modulation in rehabilitation. The technique reads EEG in real time and triggers a TMS pulse when a target brain state appears, a 'brain-state-dependent stimulation' approach that pairs the spatial precision of TMS with the temporal resolution of EEG, and the program is built around moving BCI work out of the laboratory into neurorehabilitation and the precision diagnosis of brain disorders. Shenzhen Yingchi Technology, a subsidiary of Hanix United, is among the supporting organizations, and registration closes on July 17.

A collaborative study involving authors from Hangzhou Dianzi University and other institutions developed a three-dimensional helical stretchable neural probe and achieved 1024-channel signal acquisition in the pig cerebral cortex. The researchers also report single-unit recordings with long-term stability in the mouse deep brain, while flexible planar microelectrode arrays recorded subcutaneous EMG signals in mice. The probe is fabricated using planar microfabrication followed by thermally driven helical shaping, expanding wiring space and accommodating tissue motion. Mechanical simulations and cyclic stretching tests examined strain mitigation and electrochemical stability, respectively. These device and animal results do not establish long-term performance in humans.

The study involves no implant and tests no device. It asks 50 adults with acquired blindness what existing assistive devices get wrong, and what they would want from a vision-restoration implant. Outcomes span demographics and economics, quality of life, clinical features, satisfaction with current aids, and design requirements for an invasive visual brain-computer interface. Participants must be at least 18, have only weak or no light perception in both eyes, and have lost their sight after birth; congenital blindness is excluded.