中 / EN

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

Latest

September 2026
09-10 Papers Foundations & Methods

Passenger EEG Helps Self-Driving AI Spot Road Risks Early With 95.3% Balanced Accuracy

Researchers recorded passengers' EEG as they watched driving scenes in a highly automated vehicle, then trained models to judge whether a risk lay ahead and where the hazard appeared. A 3D-CRNN reached 95.3% ± 2.7% balanced accuracy in risk prediction and raised hazard identification from 80.9% to 85.0%. In cross-subject tests on passengers the model had not seen, balanced accuracy fell to 64.9% ± 8.5%, showing the approach is still some way from deployment.
Why it matters Using the passenger, not the driver, as the signal source sets this work apart from most brain-controlled vehicle research, but the 95.3% and 85.0% figures are within-subject and drop to 64.9% across subjects, so the model remains sensitive to individual EEG differences; the paper has been accepted by Automotive Innovation but is currently a preprint that has not been peer reviewed.
View details →
09-10 Papers Foundations & Methods

EEG Motor-Imagery BCI Steers Wheelchair Prototype, With SVM Decoding at 90.7%

A research team built an EEG motor-imagery brain-computer interface and connected its decoder to a physical differential-drive wheelchair prototype, offering people with severe motor impairments a non-invasive means of control, though the uncertainty of EEG decoding has long hindered such systems from driving physical actuators. Using Filter Bank Common Spatial Pattern features on a public dataset, the team compared support vector machine (SVM), k-nearest neighbors and linear discriminant analysis classifiers; SVM reached 90.7% mean accuracy, and 85.9% when replayed on an independent dataset through a hardware-in-the-loop setup. Confidence-based command validation raised accuracy among accepted commands to 97.6%, at the cost of accepting only 60.7% of them. Layered safeguards (confidence gating, self-terminating steering, transition braking, communication timeout supervision and a hardware emergency cutoff) limit the consequences of decoding errors.
Why it matters The work takes non-invasive EEG decoding from offline classification onto a physical wheelchair prototype and reports an often-overlooked figure: with confidence validation, accepted commands are 97.6% accurate, but the system executes only 60.7% of them, leaving nearly 40% for the user to retry, so the bottleneck for brain-controlled wheelchairs is not just classification accuracy but whether the system will act when it is uncertain.
View details →
09-10 · Neuracle China Policy & Regulation Regulatory Approval

Neuracle Registers Cortical Electrode in China's NMPA Device Identifier Database

Neuracle Medical Technology (Shanghai) Co., Ltd. has registered a cortical electrode, model 1014-35, in the medical device unique identification (UDI) database of China's National Medical Products Administration (NMPA), under registration certificate Guoxiezhuzhun 20263120536. Neuracle's product line spans EEG acquisition equipment and implantable electrodes. The electrode belongs to an implantable EEG electrode kit (cortical electrode, blank electrode, electrode dilator, tunneler, torque wrench, lead fixation clip, fixation screws and electrode protective sleeve) that holds its own registration certificate and is also a component of an implantable BCI hand motor function compensation system. That system comprises a BCI implant, the electrode kit, an EEG signal transceiver, a pneumatic glove, a single-use surgical toolkit, and three software packages for EEG decoding, medical testing and clinical management.
Why it matters What was registered is not a complete BCI but a single cortical electrode, and the detail that matters is that its electrode kit is certified in its own right while also forming part of the implantable hand motor function compensation system, indicating the system is going through registration component by component; how complete its parts list is (implant, electrode kit, signal transceiver, pneumatic glove and three software packages) is a direct clue to how far it is from clinical use.
View details →
09-10 Brazil Papers Foundations & Methods

Random Forest Tells Real From Imagined Movements, Even on Consumer-Grade EEG

Can a machine tell from EEG alone whether a person is actually moving or only imagining the movement? A preprint study tested a Random Forest classifier on consumer-grade and research-grade EEG systems and found it could separate the two kinds of activity and identify which body part was involved. EEG is a basic tool for studying the brain's electrical activity during movement, and BCIs use that activity to build assistive technologies, particularly for people with physical disabilities; because extracting features and patterns from the signal remains complex, the task is often handed to machine learning. The study also found that differences in EEG between individuals drag down classification accuracy, meaning a model that works for one person may perform worse on another. The work is a preprint and has not been peer reviewed.
Why it matters The algorithm is not the story, since Random Forest is a mature method; what matters is that consumer-grade EEG hardware was part of the test. If low-cost consumer devices can support motor-intent classification, the barrier to building a BCI drops noticeably. The study also names the practical obstacle: person-to-person EEG variability lowers accuracy, so cross-user generalization is still unsolved.
View details →
09-10 Papers Foundations & Methods

More Components Can Hurt P300 Spellers, Full-Factorial Study Finds

P300 BCI spellers are often built by switching on every component that works on its own, on the assumption that more is better. A preprint study tested that assumption with a four-component full-factorial experiment and found that a component's value is conditional, not additive. Subject calibration was the strongest single contributor; Euclidean Alignment made up for the lack of calibration in zero-calibration settings; and stacking components that are each useful on their own could lower performance, an effect the authors call component anti-synergy. Language model support was not universally beneficial either: its effect depended strongly on how strong the underlying EEG pipeline was.
Why it matters The study tests a default habit in BCI engineering, stacking every algorithm module that works on its own, and finds it does not necessarily pay off: calibration mattered more than spatial filtering, and language models helped only when the EEG signal itself was clean enough. For teams building P300 spellers, that argues for fitting the pipeline to signal quality rather than piling on modules. The work is a preprint that has not been peer reviewed, and the findings still need validation on larger datasets.
View details →
09-10 Browse BCIwiki entries related to Germany: research, clinical progress and industry news, with original sources and related organizations. Papers Foundations & Methods

New Filtering Pipeline Keeps BCI Decoding at 76% During Brain Stimulation

Running a motor-imagery BCI while transcranial alternating current stimulation (tACS) is switched on has been difficult, because stimulation artifacts swamp the brain rhythms the decoder relies on and push accuracy close to chance. In a controlled test with 14 healthy participants, a real-time spatial filtering pipeline combining spatio-spectral decomposition with beamforming held accuracy at 76 ± 9% during stimulation, while a standard Laplacian filter managed only 58 ± 9%. That makes closed-loop "stimulate while reading" operation feasible at the signal level; whether it actually improves BCI performance or produces neuroplastic changes, the paper explicitly leaves to future research.
Why it matters BCI decoding and brain stimulation have usually been run separately, because once stimulation starts, artifacts bury the EEG and the decoder effectively fails. Using spatio-spectral decomposition plus beamforming as a real-time spatial filter, this study brought decoding accuracy during stimulation in 14 people from near chance back to 76%, roughly level with the 73% seen without stimulation. It does not show that stimulation improves BCI performance, which the authors themselves list as future work, but it delivers a real-time framework for stimulating and decoding at once, giving causal tests of bidirectional BCIs a usable tool.
View details →
09-10 · Huashan Hospital China Papers Foundations & Methods

OPM-MEG Beats EEG by 3.6 Points in Non-Invasive Speech Decoding

Ten native Mandarin speakers read aloud six single-vowel rhymes while optically pumped magnetometer magnetoencephalography (OPM-MEG) and EEG recorded the same task. In the 150-500 ms window after stimulus onset, OPM-MEG decoded significantly more accurately than EEG with all five classifiers tested. The best-performing combination, common spatial pattern features with a linear support vector machine, reached a mean accuracy of 60.3% for OPM-MEG versus 56.7% for EEG. That suggests speech BCIs that need no implanted electrodes and do not restrict head movement may have a clearer signal path than EEG.
Why it matters Speech BCIs have long been held back by blurry non-invasive signals. This study compares two non-invasive methods head-to-head on the same task, the same participants and the same analysis pipeline rather than in separate experiments, which is what makes the result comparable. The margin is small, 3.6 percentage points, but it points the same way throughout, and OPM-MEG tolerates head movement, which is exactly what speech tasks require.
View details →
09-10 Papers Foundations & Methods

Study Adds Protective Coatings to Pillar Electrodes in Retinal Prostheses

The study proposes coating the gold pillar sidewalls of the PRIMA retinal prosthesis with titanium dioxide or platinum while protecting surrounding device regions. The approach combines photolithography and electrodeposition to address tissue contact with exposed gold while preserving surrounding implant functionality. The authors report that both coatings were biocompatible in in-vivo studies and prevented adverse retinal tissue reactions to gold. Titanium dioxide had a lower specific capacitance than platinum, which the authors say better limits sidewall current and directs charge injection mainly through the pillar tops. These findings concern electrode materials and interfaces, not demonstrated improvements in patients' vision from the new coatings.
Why it matters The coating must separate gold from tissue without covering photosensitive regions or return electrodes. By combining selective protection with control of current paths, the study illustrates why coating an entire three-dimensional electrode device is not a straightforward solution.
View details →
09-10 Industry News

WABO Demonstrates EEG-Guided AI Image Generation at Qualcomm's Booth

WABO demonstrated Dreamer (Lingmeng), an experience that uses EEG signals to help condition AI image generation. Visitors wore a headband, viewed EEG feedback, and took part in generating images. According to the company, the system processes and analyzes recorded EEG signals, then uses the analysis results as conditions for image generation. This is not a direct reconstruction of pictures in a participant's mind. At Qualcomm's booth during the China Unicom Partner Conference, staff guided visitors through fitting the device and using the experience. The EEG display and generated artwork showed signal acquisition and creative output as two visible stages of the public product demonstration.
Why it matters The key distinction is that EEG analysis helps condition image generation; the artwork is not a mind-reading result. Showing both EEG feedback and images makes input and output visible, but the artwork alone cannot establish that the system identified specific thoughts.
View details →
09-09 China Papers Foundations & Methods

1,145-Patient Meta-Analysis Finds Stronger Evidence for Robotic Stroke Rehab Than BCI

Which new technology does most for upper-limb recovery after stroke: virtual reality, robotics or a brain-computer interface (BCI)? A systematic review and network meta-analysis placed all three in a single evidence network, pooling 25 randomized controlled trials and 1,145 stroke survivors. The authors searched PubMed, Web of Science, the Cochrane Library and Embase from inception to October 2025, used conventional physical therapy as the common comparator in a star-shaped network, and applied a Bayesian random-effects model to estimate relative efficacy and calculate SUCRA rankings. Robot-assisted training produced the most robust findings, with two studies supporting robotics plus conventional physical therapy and three supporting robotics alone; only one BCI study yielded extractable data, too little to judge efficacy. The authors caution that the top-ranked intervention, robotics combined with rehabilitative functional electrical stimulation, rests on a single trial and should not be read as definitive evidence of superiority.
Why it matters Robotic training and BCIs are often lumped together in stroke rehab but had never been compared directly; with robotics on firmer ground and only one usable BCI study, patients and rehab clinicians choosing BCI training still lack comparable efficacy data.
View details →

© 2026 BCIwiki.com Digest Topics Tips Subscribe Revisions About