A fundamental question in brain–computer interfaces (BCIs) is how much visual information can be decoded from time-resolved electrophysiological signals. Here, we propose FuzzyAlign, an alignment framework driven by fuzzy similarity, to establish a benchmark and explore the integration of large pretrained vision models with neural decoding. FuzzyAlign creates a shared latent space between large-scale electrophysiological activity and artificial visual representations, enabling similarity-weighted alignment. A convolutional model combined with fuzzy attention is used to capture temporal and spatial patterns across neural recordings. Using this fuzzy-enhanced framework, we achieve strong visual decoding performance with 1024-channel macaque multiunit activity and state-of-the-art results on human electroencephalography and magnetoencephalography, covering both object identification and image reconstruction via diffusion-based generative models. FuzzyAlign further resolves the spatial and temporal organization of primate visual object recognition, revealing biologically plausible hierarchical processing across brain areas and time. These findings demonstrate the effectiveness of incorporating fuzzy logic into computational brain models, offering a high-performing and interpretable approach for bridging neural and artificial vision systems.

For older adults with disabilities and aphasia, timely care remains challenging, particularly during the 8-h nocturnal period, creating a need for stable non-invasive EEG interfaces for assistive brain-computer interface (BCI) applications. Asynchronous steady-state visual evoked potential (SSVEP)-based control provides a practical strategy for nighttime assistive care, but its performance is largely constrained by the electrode interface. Here, we developed a comb-shaped, silicone-scaffolded PVA/SA/PEDOT:PSS hydrogel electrode for stable overnight EEG acquisition on hair-bearing scalp regions. The PVA/SA network provided soft and compliant contact, while PEDOT:PSS improved interfacial charge transfer. The comb-shaped silicone scaffold further enhanced conformal contact and effectively suppressed hydrogel dehydration. The electrode maintained stable scalp impedance for 8 h and enabled reliable EEG recording on both hairless and hair-bearing scalp regions. When used in an asynchronous SSVEP-based intelligent assistive care system, it achieved 100% triple-blink (three consecutive blinks) detection accuracy for system activation and 97.3 ± 0.4% SSVEP classification accuracy for command selection, with stable performance throughout the 8-h monitoring period. Cytotoxicity assays indicated no obvious cytotoxicity under the tested conditions, and comfort evaluations supported good wearability. This hydrogel electrode provides a promising approach for non-invasive and reliable overnight EEG acquisition and shows potential for BCI-assisted nighttime care.

Steady-state visual evoked potential-based brain-computer interfaces (SSVEP-BCIs) hold significant promise for enabling high-speed human-computer interaction in real-world scenarios. However, existing frequency-domain decoding methods treat frequency spectrum features (the real and imaginary spectrum features) as a single feature without considering their unique spatial and spectral characteristics, resulting in insufficient generalizable features and limited classification accuracy in cross-subject scenarios. To address this issue, we propose a Dual-Branch Attention-Based Frequency Domain Network (DB-AFDNet) to independently decode real and imaginary spectral components, aiming to acquire more discriminative and generalizable features for cross-subject applications. Specifically, we construct inter-branch attention similarity constraints to encourage the two branches to have similar attention properties, promoting to learn the consensus characteristics in the dual branches. Furthermore, we propose intra-branch orthogonality constraints to explore branch-specific discriminative features to learn generalizable features. Experimental studies on two public datasets, the Benchmark and Beta datasets, demonstrate that DB-AFDNet outperforms state-of-the-art methods in cross-subject classification, achieving a relative improvement of 1.36$\%$ and 1.45$\%$, respectively.

Increasing the number of optional instructions has emerged as a significant issue for brain-computer interfaces (BCIs) based on Visual Evoked Potential (VEP). A large instruction set contributes to enhancing the system performance and facilitating more complex system control. Nevertheless, expanding the instruction set typically raises the encoding complexity, and calibration session becomes necessary for most encoding methods. This study proposed a code-modulated visual evoked potential (c-VEP) based BCI with 504 targets employing narrowband random sequences. The proposed BCI system is calibration-free and achieves a large instruction set of 504 targets. To further enhance the system performance, brain signals were acquired using a high-density electroencephalogram cap. The online experiments demonstrated an average accuracy of 81.55 +/- 3.66 % and an average Information Transfer Rate (ITR) of 66.95 +/- 4.09 bits/min. Offline simulation presented that the proposed system could achieve an average ITR of 98.24 +/- 7.88 bits/min after the optimization of the visual pathway latency. The proposed encoding method expanded the number of targets for VEP-based BCIs to over 500, demonstrating the high encoding efficiency and enabling more complex applications of VEP-based BCIs.

In recent years, the popularity of brainwave verification has been steadily increasing, with a plethora of paradigms and algorithmic research emerging in this field. However, most works suffer from issues such as low performance, lack of online systems, irrevocability, and are far from meeting practical demands. In response to these challenges, we propose a dual-factor brainwave verification system that integrates passwords with electroencephalogram (EEG) templates. Our approach involved conducting single-target experiments to select sequences, multi-target offline experiments to optimize parameters, and ultimately performed online experiments for validation. Using a 240 Hz refresh rate, we constructed a keyboard input for EEG signals using a 127-bit m-sequence code, and combined passwords with EEG signals to achieve online identification and authentication functionalities. We collected cross-day data from 30 participants, with a time interval of approximately six months. In the online system, the recognition accuracy was 99.2 % with a 4-digit password (4 s), 99.93 % with a 6-digit password (6 s), and 100 % with an 8-digit password (8 s). The equal error rates for online authentication were 1.79 % (4-digit password, 8 s), 1.03 % (6-digit password, 12 s), and 0.8 % (8-digit password, 16 s). The proposed system introduces dual-factor verification, significantly improving online performance in cross-day recognition and authentication. Simultaneously, it allows template replacement without the need for repeated registration, thus providing revocability and enhancing the system's resilience against attacks. These advantages position the proposed system favorably in the field of biometric recognition, contributing to the practical application of EEG-based identity verification systems.

Abstract Background Ear-EEG-based brain-computer interfaces (BCIs) provide improved wearability and comfort compared to traditional scalp-EEG systems. However, their performance is constrained by low signal-to-noise ratios (SNRs) and high rates of BCI illiteracy under conventional luminance-modulated steady-state visual evoked potential (SSVEP) paradigms. Methods This study introduces a text-sequence stimulation paradigm to address these limitations by leveraging ventral visual pathway responses that are more accessible to electrodes near the ear. Using offline frequency-sweeping experiments across 4–8 Hz, we identified optimal stimulus parameters (4.6–6.8 Hz with 0.25π phase shifts) and integrated them into a 12-target BCI system. We further conducted online experiments to compare the response characteristics and real-time spelling performance between the proposed text-sequence paradigm and conventional luminance stimulation. Results Comparative experiments with 14 participants demonstrate that text sequence stimuli achieve an average information transfer rate (ITR) of 44.59 ± 10.50 bits/min, outperforming luminance modulation by 76.18% in ITR. Notably, text sequence stimulation effectively mitigated BCI illiteracy, with all participants achieving near or above 70% accuracy (mean: 86.37 ± 9.61%). This represents a significant improvement over luminance modulation, where 50% of users fell below 70% accuracy. Conclusions By reducing the flicker area by 14% and mimicking the natural luminance variations that occur during reading, the proposed method enhanced visual comfort. The online results further validate text-sequence stimulation as a high-performance and user-friendly paradigm for ear-EEG BCIs, supporting their practicality for assistive applications.

Abstract The objective assessment of patients with disorders of consciousness (DOC) remains a significant clinical challenge. Behavioral scales like the Coma Recovery Scale-Revised (CRS-R) are susceptible to rater subjectivity and have difficulty in detecting patients with cognitive-motor dissociation (CMD), while existing electrophysiological paradigms typically evaluate isolated processing levels, especially in visual functions. To address these limitations, we developed a novel, hierarchical visual EEG framework that evaluates three progressive tiers of visual processing—sensory input, selective attention, and object discrimination—within a single, unified paradigm. This framework uses steady-state and event-related potentials, analyzed with statistical testing and machine learning, to provide objective detection. In a cohort of 85 participants, the framework demonstrated a robust alignment with behavioral CRS-R levels and successfully identified CMD patients missed by bedside behavioral examinations. Notably, model predictions derived from this framework showed a significant correlation with 3-month clinical outcomes. This prognostic utility generalized effectively and remained consistent across distinct EEG acquisition systems in an independent validation cohort of 17 patients. In summary, this work offers electrophysiological validation for the hierarchical design of the CRS-R and provides a practical tool for bedside objective assessment of DOC.

Objective.Steady-state visual evoked potential (SSVEP) based brain-computer interfaces (BCIs) predominantly employ frequency, phase, or spatial coding. This study proposes a color-dimension SSVEP encoding scheme and evaluates its feasibility and elicited response characteristics.Approach.Seven isoluminant colors were paired to form 21 combinations, and four stimulation paradigms (sliding checkerboard, reversing checkerboard, flickering checkerboard, and solid-color flicker) were used to investigate the modulatory effects of color on SSVEP. Offline simulations and an online four-target SSVEP-BCI were conducted for validation purposes.Main results.Under identical frequencies and initial phases, different color combinations produced separable SSVEP patterns in amplitude, topography, and phase, enabling reliable classification. At a stimulation frequency of 10 Hz, the four-target online system using the solid-color flicker paradigm achieved an average information transfer rate (ITR) of 80 ± 0 bits/min, and its user experience was favorable compared with the classic black-white stimulation. Additionally, the system employing the flickering checkerboard paradigm obtained an average ITR of 68.76 ± 1.79 bits/min while delivering even better user experience, with a comfort score of 3.10 ± 0.21, a flicker perception score of 3.07 ± 0.21, and a preference score of 2.78 ± 0.21.Significance.The proposed approach introduces an additional encoding dimension for SSVEP-BCI, expanding stimulus design options and supporting broader applications.

Brain–computer interface (BCI) technology establishes a direct communication pathway between the brain and external devices. Current visual BCI systems suffer from insufficient information transfer rates (ITRs) for practical use. Spatial information, a critical component of visual perception, remains underexploited in existing systems because the limited spatial resolution of recording methods hinders the capture of the rich spatiotemporal dynamics of brain signals. This study proposed a hybrid frequency–phase–space encoding method, integrated with high-density electroencephalogram (EEG) recordings, to develop high-speed BCI systems. EEG data were recorded using a 256-channel standard cap, and 4 electrode configurations comprising 66, 32, 21, and 9 parieto-occipital electrodes, extracted from 256-, 128-, and 64-channel caps (abbreviated as 66/256, 32/128, 21/64, and 9/64), were systematically compared. In the classical frequency–phase encoding the 40-target BCI paradigm, the 66/256, 32/128, and 21/64 electrode configurations brought theoretical ITR increases of 83.66%, 79.99%, and 55.50% over the traditional 9/64 setup. In the proposed frequency–phase–space encoding 200-target BCI paradigm, these increases climbed to 195.56%, 153.08%, and 103.07%, respectively. The online BCI system achieved an average actual ITR of (472.72 ± 15.06) bits per minute. Taken together, these findings clarify how the spatiotemporal encoding strategy and electrode density jointly determine achievable ITRs and provide quantitative design guidelines for future high-speed visual BCIs.

We present a lateral ventricular brain-computer interface (LV-BCI) that deploys an expandable, flexible electrode into the lateral ventricle through a minimally invasive external ventricular drainage pathway. Inspired by the framework of traditional Chinese lanterns, the electrode expands uniformly within the ventricle and conforms to the ependymal wall. Compared with conventional subdural ECoG electrodes, the LV-BCI shows superior signal stability and immunocompatibility. Resting-state spectral analyses revealed a maximum effective bandwidth comparable to subdural ECoG. In evoked potential tests, the LV-BCI maintained a consistently higher signal-to-noise ratio over 112 days without the decline typically associated with scarring or other immune responses. Immunohistochemistry showed only a transient, early microglial activation after implantation, returning to control levels and remaining stable through 168 days. We further designed an "action-memory T-maze" task and developed a microstate sequence classifier (MSSC) to predict rats' turn decisions. The LV-BCI achieved prediction accuracy up to 98

Steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs) are widely recognized for their high information transfer rates (ITRs), yet their practical application remains constrained by the challenge of balancing performance and user experience. To address this issue, this study proposes a design framework for individualized spatial-phase-amplitude modulation stimuli aimed at enhancing response strength at frequencies above the critical flicker fusion threshold. In this framework, phase parameters are allocated to subregional stimuli to alleviate cancellation effects induced by the cortical cruciform organization, while amplitude parameters selectively determine whether a given subregion should be stimulated, thereby avoiding redundant use of visual resources. A single-target experiment was first conducted to record SSVEP responses to 60 Hz stimuli from 16 subregions across two eccentricities and eight polar angle positions. Pronounced differences were observed in both amplitude and phase distributions across subregions, and substantial inter-individual variability further underscored the necessity of subject-specific parameter optimization. A first-choice hill climbing strategy was subsequently introduced to enable efficient optimization of spatial-phase-amplitude parameters within the large space generated by the numerous possible combinations. Finally, the superiority of individualized stimuli was validated in a four-target BCI experiment. Within the 0-7 degrees eccentricity range, individualized stimuli (offline peak: 156.58 f 8.53 bpm, online: 77.83 f 1.23 bpm) achieved a significantly higher ITR than unmodulated stimuli (offline peak: 123.16 f 7.45 bpm, online: 69.29 f 2.97 bpm) while maintaining high subjective userexperience ratings. These results demonstrate the feasibility of individualized spatial-phase-amplitude modulation as a pathway toward high-performance, visually comfortable, flicker-free SSVEP-BCIs.

Steady-state visual evoked potential brain–computer interfaces offer a high-speed communication channel. However, traditional steady-state visual evoked potential paradigms often rely on strong flickering visual stimulation, which can lead to substantial visual fatigue. Moreover, the electroencephalography responses evoked by brightness flicker are spatially constrained and are primarily associated with occipital visual processing. This study presents a novel text sequence stimulation paradigm that combines periodic visual stimulation with orthographic information and elicits distinct occipital and occipitotemporal scalp response patterns relative to conventional brightness flicker. Frequency-sweep experiments were conducted to investigate the temporal, spatial, and spectral characteristics of the evoked responses. A comparison experiment further showed that text sequence stimulation is less sensitive to variations in stimulus size and luminance than conventional brightness flicker. Based on these findings, a 40-target speller was developed and validated through online experiments. The proposed paradigm achieved an information transfer rate of 235.12 ± 30.12 bits/min while significantly improving user comfort, as confirmed by questionnaire evaluations. These results suggest that text sequence stimulation offers a practical design direction for high-speed and more comfortable visual brain–computer interface.

Objective: Rapid Serial Visual Presentation (RSVP) had been applied to human-computer interactions such as spelling, device control and so on. Traditional single RSVP paradigm detects targets in only one image stream, which may lead to missed or false detection. Dual RSVP paradigm can enhance the classification robustness through specific encoding to increase the number of targets. Methods: To enhance the classification performance of dual RSVP, an EEG-EM Self Attention and Cross Attention Network (EESCAN) is proposed. EEG and EM signals undergo a symmetric two-stream backbone. Each stream consists a convolution module and a self-attention module to extract local and global features. Subsequently, an inter-modal bidirectional interaction module is proposed to provide complementary information between EEG and EM modality. Finally, dynamic reweighting and fusion module is employed to dynamically adjust the sample-level weights according to the contribution of EEG and EM features. Moreover, a VR-based dual RSVP virtual robotic arm control system using the proposed algorithm is then designed to achieve gaze-independent device control. Results: EEG and EM data from 21 subjects were collected and analyzed. The proposed network and system achieved better performance than existing decoding methods and single-modal baselines. Ablation experiments and visualization results further verified the effectiveness of each proposed module. Conclulsion: EESCAN network is proposed for the dual RSVP paradigm that integrates intra-modal self-attention, inter-modal bidirectional interaction, and dynamic fusion to achieve EEG-EM fusion. Significance: EESCAN markedly improves classification performance of dual RSVP-based BCIs. This gaze-independent control system is suitable for patients with restricted gaze shifts.

Abstract The visual receptive field (RF) characterizes the spatiotemporal properties of the visual pathway and serves as a fundamental unit for information encoding. While RFs have been extensively studied across various neural modalities, such as functional Magnetic Resonance Imaging (fMRI), Electrocorticography (ECoG), and Magnetoencephalography (MEG), their investigation via Electroencephalography (EEG) remains limited. In this study, we introduce a stimulation paradigm that combines white noise image sequences with a letter detection task to elicit central visual field EEG responses. Using the aligned/shuffled reverse correlation, we estimate RFs across different resolutions and demonstrate that the resulting RFs exhibit rich spatiotemporal characteristics. To validate the reliability of the estimated RFs, we constructed a visual EEG reconstruction model, which achieved good performance in a classification task. The same RF estimation method was subsequently applied to high-density EEG recordings to investigate the information gain afforded by high-density configurations in visual space. This work fills a gap in the study of visual RFs regarding the EEG modality and may inform the paradigm design of visual brain–computer interfaces.

Steady-State Visually Evoked Potential (SSVEP) signals can be decoded by either a traditional machine learning algorithm or a deep learning network. Combining the two methods is expected to enhance the performance of an SSVEP-based brain-computer interface (BCI) by exploiting their advantages. However, an efficient strategy for integrating the two methods has not yet been established. To address this issue, we propose a classification framework named eTRCA + sbCNN that combines an ensemble task-related component analysis (eTRCA) algorithm and a sub-band convolutional neural network (sbCNN) for recognizing the frequency of SSVEP signals. The two models are first trained separately, then their classification score vectors are added together, and finally the frequency corresponding to the maximal summed score is decided as the frequency of SSVEP signals. The proposed framework can effectively exploit the complementarity between the two kinds of feature signals and significantly improve the classification performance of SSVEP-based BCIs. The performance of the proposed method is validated on two SSVEP BCI datasets and compared with that of eTRCA, sbCNN and other state-of-the-art models. Experimental results indicate that the proposed method significantly outperform the compared algorithms, and thus helps to promote the practical application of SSVEP- BCI systems.

In this study, we introduce a brain-computer interface (BCI) framework incorporating MXene microneedle EEG electrodes, tailored for versatile deployment. The dry electrodes, configured as 1 mm2 microneedles, underwent meticulous processing to establish a cohesive integration with the MXene conductive material. The microneedle architecture facilitates epidermal penetration, yielding low contact impedance, enabling the recording of spontaneous EEG and induced brain activity, and ensuring high precision in steady-state visual evoked potential (SSVEP) speller. Simultaneously, the microneedle electrode demonstrates commendable biological compatibility and superior nuclear magnetic resonance compatibility. It exhibits minimal artifact generation and manifests no heating-related adaptations in nuclear magnetic environments. The inherent microneedle electrode structure endows it with robust anti-interference capabilities. In vibrational environments, the SSVEP text input accuracy of the microneedle electrode remains comparable to that of gel electrodes, maintaining consistent impedance and delivering high-fidelity EEG acquisition during real-motion scenarios. The microneedle electrode devised in this study serves as a reliable signal acquisition tool, thereby advancing the development of BCI systems tailored for practical usage scenarios.

Objective: Auditory Evoked Potentials (AEP), particularly the N100 component and the auditory steady-state response (ASSR), have been utilized in the clinical assessment of patients with Disorders of Consciousness (DOC). However, the specific utility of these measures remains debated across studies. Methods: To clarify the roles of N100 and ASSR in evaluating auditory function and levels of consciousness in DOC patients, we recorded N100 and ASSR responses in 30 DOC patients and assessed their significance at the individual level through statistical analyses. Results: Our findings indicate that, compared to N100, the significance of the ASSR response appears to be a more reliable marker of auditory function. However, neither N100 nor ASSR, at both response and microstate levels, could effectively distinguish between patients diagnosed with unresponsive wakefulness syndrome (UWS) and minimally conscious state (MCS). Additionally, we validated the role of ASSR using a portable EEG device in an independent cohort of 30 patients. Conclusion: In summary, our results suggest that ASSR holds promise for assessing auditory function in DOC patients, but its utility in differentiating levels of consciousness may require further consideration. Significance: These findings offer valuable insights for clinicians and neuroscientists in selecting and designing objective tools for DOC assessment.

Electroencephalography (EEG), known for its convenient noninvasive acquisition but moderate signal-to-noise ratio, has recently gained much attention due to the potential to decode image information. However, previous works have not delivered sufficient evidence of this task, primarily limited by performance and biological plausibility. In this work, we first introduce a self-supervised framework to demonstrate the feasibility of recognizing images from EEG signals. Contrastive learning is leveraged to align the representations of EEG responses with image stimuli. Then, language descriptions of the stimuli generated by large language models (LLMs) help guide learning core semantic information. With the framework, we attain significantly above-chance results on the THINGS-EEG2 dataset, achieving a top-1 accuracy of 19.7% and a top-5 accuracy of 51.5% in challenging 200-way zero-shot tasks. Furthermore, we conduct thorough experiments to resolve the human visual responses with EEG from temporal, spatial, spectral, and semantic perspectives. These results provide evidence of feasibility and plausibility regarding EEG-based image recognition, substantiated by comparative studies with the THINGS-Magnetoencephalography (MEG) dataset. The findings offer valuable insights for neural decoding and real-world applications of brain-computer interfaces (BCIs), such as health care and robot control. The code is available at https://github.com/eeyhsong/NICE-LLM.

Major Depressive Disorder is a leading cause of disability worldwide. An accurate assessment of depression severity is critical for diagnosis, treatment planning, and monitoring, yet current clinical tools are largely subjective, relying on self-report and clinician judgment via traditional assessment scales. EEG has emerged as a promising, non-invasive modality for capturing neural correlates of depression. However, most EEG-based machine learning diagnostic studies focus on boosting classification accuracy through complex algorithms and small, homogenous datasets. These black-box approaches often yield results that are difficult to interpret and poorly generalizable, making clinical translation impractical. Therefore there remains a critical need for models that are not only accurate but also transparent, robust, and grounded in the physiological properties of the data itself. We proposed a data-centric, interpretable framework for EEG-based depression severity grading. A hybrid feature selection method was used, combining p-value and SHapley Additive exPlanations (SHAP) methods to select features that are both independently significant and jointly informative. The system was trained and evaluated on a large-scale, multi-site resting-state EEG dataset, using random forest for both classification and regression tasks. The SHAP method, an explainable artificial intelligence technique, is also used post-hoc to infer the key electrophysiological features and key brain regions associated with MDD mechanism to further increase interpretability. The proposed system achieved 74.5

Most existing studies analyzed the resting-state electroencephalogram (EEG) of DOC patients, and recent research demonstrated that the passive auditory paradigm was helpful for bedside detection of DOC and better captured sensory and cognitive responses. However, further studies of classification algorithms were needed for consciousness assessment in DOC based on task-state EEG data. In this study, EEG data from minimally conscious state (MCS) patients, vegetative state (VS) patients, and a healthy control group (HC) were collected using an auditory oddball paradigm. First, compared to the fragmented features adopted by most studies, multiple effective biomarkers for consciousness assessment in the time-frequency domains, connectivity and nonlinear dynamics were identified. Event-related potentials (ERP) results showed that MCS and VS patients exhibited lower N100 and MMN amplitudes than the HC group. Spectral analysis results indicated that VS patients had higher Delta power, and lower Alpha and Beta power than the MCS and HC groups. Second, different from insufficient classifiers in previous studies, this study systematically compared the performance of multiple machine learning and deep learning (DL) classifiers, including support vector machine (SVM), linear discriminant analysis (LDA), random forest (RF), eXtreme Gradient Boosting (XGBoost), decision tree (DT), EEGNet and ShallowConvNet. For machine learning methods, SVM and RF had an advantage in binary classification, and SVM had better performance in three-class classification. Among all individual classifiers, Shallow ConvNet had the best performance for binary and three-class classification. Moreover, an ensemble model incorporating all seven classifiers was proposed using a voting strategy, and further improved classification performance that was superior to existing studies. In addition, the importance of each feature was analyzed, identifying N100, MMN, Delta, Alpha, and Beta power as significant biomarkers of consciousness assessment.