@article {pmid41206890,
year = {2025},
author = {Zapata-Catzin, GA and Vargas-Coronado, RF and Ceballos-Gongora, E and Arana-Argáez, VE and Rodríguez-Velázquez, E and Alatorre-Meda, M and Molina-Salinas, GM and Uc-Cachon, A and Gallardo, A and Copes, F and Mantovani, D and Cauich-Rodríguez, JV},
title = {Effect of Polyurethane Structure on the Physicochemical, Mechanical, and Biological Properties on their Copper Complexes Composites.},
journal = {Macromolecular bioscience},
volume = {},
number = {},
pages = {e00419},
doi = {10.1002/mabi.202500419},
pmid = {41206890},
issn = {1616-5195},
support = {AtencionaProblemasNacionales(248378)//Consejo Nacional de Humanidades, Ciencia y Tecnología/ ; FronterasdelaCiencia(1360)//Consejo Nacional de Humanidades, Ciencia y Tecnología/ ; 1360//Atención a Problemas Nacionales/ ; //Fronteras de la Ciencia/ ; },
abstract = {Polyurethanes and their composites are versatile materials widely used in numerous medical applications. However, limited information is available regarding their copper composites. Copper is a trace element in the human body that functions as an enzyme cofactor in both normal and pathological angiogenesis, as well as in muscle and brain formation. Considering this, copper complexes of D-penicillamine (DP), L-cysteine (LC), and dopamine (DOP) were incorporated into segmented polyurethanes (SPU) synthesized with either a semi-crystalline (poly-ε-caprolactone, PCL) or an amorphous (polytetramethylene ether glycol, PTMEG) soft segment. FTIR and Raman revealed new absorptions and peak shifts, confirming the presence of the complexes within the matrix of all composites. XPS further corroborated the presence of copper and sulfur. The crystallinity of the PCL-based polyurethanes was influenced by the addition of the filler, as observed through DSC and DRX. Furthermore, TGA analysis indicated the emergence of new decomposition temperatures following the incorporation of copper complexes. In general, no significant reduction in Young's modulus was observed, except for certain composites containing DPENCUII as filler, which exhibited a slight increase compared to pristine SPU´s. Finally, the composites demonstrated neither hemolytic nor procoagulating behavior (hemolysis < 5% and BCI > 20), although they exhibited some degree of impairment in cytocompatibility compared to their respective pristine SPUs. Collectively, these findings suggest that some composites possess promising properties for potential cardiovascular applications.},
}

@article {pmid41206869,
year = {2025},
author = {Tian, Y and Jiang, R and Guo, F},
title = {Protocol for in vivo two-photon calcium imaging of the Drosophila brain.},
journal = {STAR protocols},
volume = {6},
number = {4},
pages = {104194},
doi = {10.1016/j.xpro.2025.104194},
pmid = {41206869},
issn = {2666-1667},
abstract = {Two-photon calcium imaging facilitates the real-time observation of neuronal activity. Here, we present a protocol for conducting in vivo two-photon calcium imaging of the Drosophila melanogaster brain. We describe steps for fly preparation, recording chamber construction, and preparation of the buffer solution. We then detail procedures for fly brain surgery, execution of the recording, and data analysis. This protocol enables the monitoring and assessment of neuronal responses to external stimuli and the mapping of functional connectivity coupled with optogenetics. For complete details on the use and execution of this protocol, please refer to Jiang et al.[1].},
}

@article {pmid41205898,
year = {2025},
author = {Taranath, JR},
title = {On questions of predictability and control of an intelligent system using probabilistic state-transitions.},
journal = {Neuroscience},
volume = {},
number = {},
pages = {},
doi = {10.1016/j.neuroscience.2025.10.062},
pmid = {41205898},
issn = {1873-7544},
abstract = {One of the central aims of neuroscience is to reliably predict the behavioral response of an organism using its neural activity. If possible, this implies we can causally manipulate the neural response and design brain-computer-interface systems to alter behavior, and vice-versa. Hence, predictions play an important role in both fundamental neuroscience and its applications. Can we predict the neural and behavioral states of an organism at any given time? Can we predict behavioral states using neural states, and vice-versa, and is there a memory-component required to reliably predict such states? Are the predictions computable within a given timescale to meaningfully stimulate and make the system reach the desired states? Through a series of mathematical treatments, such conjectures and questions are discussed. Answering them might be key for future developments in understanding intelligence and designing brain-computer-interfaces.},
}

@article {pmid41205562,
year = {2025},
author = {Sun, Z and Sun, Y and Zeng, Y},
title = {BACNet: A multi-attention network for cross-subject and cross-task EEG-based pilot operational intent recognition.},
journal = {Computer methods and programs in biomedicine},
volume = {274},
number = {},
pages = {109134},
doi = {10.1016/j.cmpb.2025.109134},
pmid = {41205562},
issn = {1872-7565},
abstract = {BACKGROUND AND OBJECTIVE: Recognizing pilot operational intent is crucial for enhancing flight safety and improving the efficiency of human-machine interaction. Electroencephalography (EEG), known for its high temporal resolution and non-invasive acquisition, has become a prominent modality for this task. However, current approaches often suffer from high model complexity and limited accuracy in EEG feature extraction. This study aims to address these limitations by proposing efficient and accurate neural network architecture for pilot intent recognition based on EEG signals.

METHODS: We introduce a novel framework, the Balanced Attention Convolutional Network (BACNet), designed to enhance EEG-based intent recognition through collaborative optimization in both channel and spatial dimensions. BACNet features: (1) a three-branch parallel convolutional structure that extracts multi-scale time-frequency features; and (2) dynamic feature modulation mechanisms to adaptively highlight salient channels and spatial locations. EEG data were collected from 15 participants across various simulated flight phases, forming a labeled dataset for model training and evaluation. Five-fold cross-validation was conducted to ensure the robustness of the performance assessment.

RESULTS: BACNet achieved an average classification accuracy of 96.07 % in a three-class EEG-based intent recognition task, outperforming five state-of-the-art baseline methods. The model also demonstrated a significant reduction in computational complexity. Ablation experiments validated the individual and combined contributions of the multi-scale attention modules, highlighting the effectiveness of the collaborative attention design.

CONCLUSION: With its lightweight architecture and high accuracy, BACNet not only provides a novel solution for pilot operational intent recognition but also demonstrates broad applicability in brain-computer interface (BCI) systems.},
}

@article {pmid41205408,
year = {2025},
author = {Bao, C and Ma, Y and Li, M and Li, Y and Zhang, C and Liu, X and Fan, R and Cui, W and Fan, X and Zheng, F and Duan, F and Liu, J},
title = {Assessment of glymphatic dysfunction in ulcerative colitis using DKI-ALPS: An innovative imaging biomarker.},
journal = {Journal of neuroradiology = Journal de neuroradiologie},
volume = {53},
number = {1},
pages = {101402},
doi = {10.1016/j.neurad.2025.101402},
pmid = {41205408},
issn = {0150-9861},
abstract = {PURPOSE: Ulcerative colitis (UC) is associated with higher anxiety, depression, and cognitive disorders linked to brain glymphatic dysfunction. In this study, we used along-the-perivascular-space (ALPS) index (based on DTI and DKI) to determine if UC relates to glymphatic dysfunction and explore how microbiota dysbiosis and inflammation affect brain glymphatic function.

MATERIALS AND METHODS: In this study, 63 patients with UC and 68 healthy controls underwent 3-Tesla MRI scans to evaluate DTI-ALPS and DKI-ALPS index. The protocol included diffusion-weighted imaging (DWI) and diffusion kurtosis imaging (DKI) sequences to calculate the ALPS index, which quantifies glymphatic system function. All participants completed cognitive (MMSE) and depression (SAS/SDS) assessments (SAS/SDS). Patients with UC also underwent assessment for inflammation and gut microbiota (based on metagenomic analysis). Data analysis was performed using correlation analysis and linear regression.

RESULTS: Patients with UC showed lower DTI-ALPS index (1.25) and DKI-ALPS index (1.40) compared to controls (1.40 vs. 1.69; P < 0.001). In multi-adjusted linear regression models, UC was associated with lower DTI-ALPS index and DKI-ALPS index (β =-0.142 vs.-0.284), with DKI-ALPS showing higher sensitivity. The results remained significant even after stratification by age and sex. The Mayo score correlated negatively with DTI and DKI-ALPS index. The ALPS index correlates with gut microbiota, particularly those involved in butyrate and short-chain fatty acid (SCFA) production. DTI-ALPS index was significantly correlated with ESR (β =-0.003), CRP (β =-0.035), SII (β =-0.062), INFLA (β =-0.010), and SIRI (β =-0.058). We also observed significant correlations between DKI ALPS index and ESR (β =-0.006), CRP (β =-0.051), SII (β =-0.130), INFLA (β =-0.017), SIRI (β =-0.095), IL-6 (β =-0.081) and NLR (β =-0.108).

CONCLUSIONS: UC is associated with brain glymphatic dysfunction, correlating with inflammation level. DKI-ALPS serves as a more sensitive method than DTI-ALPS, offering a new approach for managing ulcerative colitis through glymphatic dysfunction.},
}

@article {pmid41204711,
year = {2025},
author = {Scherer, J and Finke, A and Everding, V and Lindenbaum, L and Kayser, C and Kissler, J},
title = {NeuroCommTrainer: Toward an Adaptive and Wearable Multimodal Brain-Computer Interface.},
journal = {Brain connectivity},
volume = {},
number = {},
pages = {},
doi = {10.1177/21580014251393151},
pmid = {41204711},
issn = {2158-0022},
abstract = {Introduction: To date, brain-computer interfaces (BCIs) have not achieved reliable real-time communication through auditory or tactile modalities. Such interfaces would be crucial for brain-injured patients with severe motor impairments who are also blind or deaf. This study validates the functionality of the NeuroCommTrainer, a mobile and easy-to-use multimodal BCI with flex-printed electrode strips that does not require vision and adapts to users' attentiveness levels to initiate stimulation. Methods: In a study of 20 healthy participants, we evaluated auditory and vibrotactile oddball paradigms to train the system to differentiate rare and frequent event-related potentials (ERPs). In real-time online sessions, the system detected participants' mental focus to adaptively initiate stimulation through attentiveness monitoring. Results: The NeuroCommTrainer successfully captured auditory and tactile ERPs, achieving a classification accuracy of 75% for stimuli in the calibration session, which is not yet reflected in the online session with 34% of found targets (chance level = 16.7%). Discussion: The presented early-stage prototype of the NeuroCommTrainer requires several improvements before clinical application in brain-damaged patients, which include refined algorithms to reduce classification variance across participants, and enhanced attentiveness detection specifically tuned to brain activity of the targeted patient group. The present study makes a critical step in this direction and shows that a transition into a practicable communication system for brain-damaged patients may be achievable in the future.},
}

@article {pmid41204680,
year = {2025},
author = {Ehrlich, SK and Tougas, G and Bernstein, J and Buie, N and Rumbach, AF and Simonyan, K},
title = {Brain-Computer Interface Improves Symptoms of Isolated Focal Laryngeal Dystonia: A Single-Blind Study.},
journal = {Movement disorders : official journal of the Movement Disorder Society},
volume = {},
number = {},
pages = {},
doi = {10.1002/mds.70114},
pmid = {41204680},
issn = {1531-8257},
support = {R01DC019353/DC/NIDCD NIH HHS/United States ; },
abstract = {BACKGROUND AND OBJECTIVE: Laryngeal dystonia (LD) is a focal task-specific dystonia, affecting speaking but not whispering or emotional vocalizations. Therapeutic options for LD are limited. We developed and tested a non-invasive, closed-loop, neurofeedback, brain-computer interface (BCI) intervention for LD treatment.

METHODS: Ten patients with isolated focal LD participated in the study. The personalized BCI system included visual neurofeedback of individual real-time electroencephalographic (EEG) activity during symptomatic speaking compared to asymptomatic whispering, presented in the virtual reality (VR) environment of real-life scenarios. During five consecutive days of intervention, patients used the BCI to learn to modulate their abnormally increased brain activity during speaking and match it to near-normal activity of asymptomatic whispering. Changes in voice symptoms and EEG activity were quantified for the evaluation of BCI effects.

RESULTS: Compared to baseline, LD patients had a statistically significant reduction of their voice symptoms on Days 1-5 of BCI intervention. Thi was paralleled by improved controllability of the visual neurofeedback and a significant reduction of left frontal delta power, including superior and middle frontal gyri, on Day 1 and left central gamma power, including premotor, primary sensorimotor, and inferior parietal areas, on Days 3 and 5. The majority of patients (70%) reported sustained positive effects of the BCI intervention on their voice quality 1 week after the study participation.

CONCLUSION: The closed-loop BCI neurofeedback intervention specifically targeting disorder pathophysiology shows significant potential as a novel treatment option for patients with LD and likely other forms of task-specific focal dystonia. © 2025 International Parkinson and Movement Disorder Society.},
}

@article {pmid41204389,
year = {2025},
author = {Zhang, H and Chen, WJ and Chao, YG and Su, N and Robba, C and Czosnyka, M and Smielewski, P and Czosnyka, Z and He, W and Hu, X and Yao, DZ and Hu, CG and Zhou, M and Wang, YJ and Ma, XC and Liu, XY and Ming, D},
title = {Neurogenic organ dysfunction syndrome after acute brain injury.},
journal = {Military Medical Research},
volume = {12},
number = {1},
pages = {77},
pmid = {41204389},
issn = {2054-9369},
support = {ZYGXQNJSKYCXNLZCXM-H15//Scientific Research Innovation Capability Support Project for Young Faculty/ ; 0401260011//National Science Fund for Excellent Overseas Scholars/ ; 82472098//Innovative Research Group Project of the National Natural Science Foundation of China/ ; 32300704//National Natural Science Foundation of China/ ; 24JCJQJC00250//National Outstanding Youth Science Fund Project of National Natural Science Foundation of China/ ; 24ZXZSSS00510//Major Science and Technology Special Projects and Engineering - Major Project of National Key Laboratories/ ; 2021YFF1200602//National Key Technologies Research and Development Program/ ; 2024-JKCS-16//the Non-profit Central Research Institute Fund of Chinese Academy of Medical Sciences/ ; },
mesh = {Humans ; *Brain Injuries/complications/physiopathology ; *Multiple Organ Failure/etiology/physiopathology ; },
abstract = {Systemic complications are common after acute brain injury (ABI) and may trigger coagulation cascades, systemic inflammation, as well as dysfunction of the cardiovascular, respiratory, and gastrointestinal systems, etc. The pathogenesis of these systemic manifestations is multifactorial but not yet fully elucidated. This paper introduces the novel term neurogenic organ dysfunction syndrome (NODS) to characterize systemic instability arising from internal and external perturbations of the neuronal center following ABI. Elucidating the central neurogenic mechanisms of NODS is critical for early detection and prevention of complications, thereby reducing mortality and improving patient outcomes following ABI. In this paper, we explore the potential central neurogenic mechanisms of NODS from the perspective of complex brain network theory, focusing on the structural network of the central autonomic system (CAS) that maintains systemic stability, and the functional network governed by the central stress system (CSS). The CAS can be divided into the cortical autonomic network, which involves higher cortical regions, and the subcortical autonomic network, which is relatively conserved, with its main connections located in deep brain structures. The CSS is a large-scale complex network characterized by hierarchy, hubs, and modularity, which together enable the competitive optimization of functional segregation and integration. Under physiological conditions, modules (mediating functional segregation) and hubs (functional integration) within the CSS dynamically trade-off with each other to maintain the overall homeostasis. However, this balance is disrupted following pathological insults or injury, resulting in weakened functional integrity of the CSS following ABI, impaired module activity, and disturbed hub integration. This paper also demonstrates the distinct pathological manifestations arising from disturbances at different levels of the homeostatic system. Finally, this study proposes potential clinical interventions, including analgesia and sedation, neuromodulation, and receptor regulation, for early interventions and potential treatment of NODS, aiming to improve patient outcomes.},
}

Humans
*Brain Injuries/complications/physiopathology
*Multiple Organ Failure/etiology/physiopathology

@article {pmid41203630,
year = {2025},
author = {Thapa, BR and Boggess, J and Bae, J},
title = {A large electroencephalogram database of freewill reaching and grasping tasks for brain machine interfaces.},
journal = {Scientific data},
volume = {12},
number = {1},
pages = {1760},
pmid = {41203630},
issn = {2052-4463},
mesh = {Humans ; *Electroencephalography ; *Brain-Computer Interfaces ; Male ; Female ; Young Adult ; Adolescent ; *Hand Strength ; },
abstract = {Brain machine interfaces (BMIs) offer great potential to improve the quality of life for individuals with neurological disorders or severe motor impairments. Among various neural recording modalities, electroencephalogram (EEG) is particularly favorable for BMIs due to its noninvasive nature, portability, and high temporal resolution. Existing EEG datasets for BMIs are often limited to experimental settings that fail to address subjects' freewill in decision making. We present a large EEG dataset, containing a total of 6808 trials, recorded from 23 healthy young adults (eight females and 15 males with an age range from 18 to 24 years) while performing reaching and grasping tasks, where the target object is freely chosen at their desired pace according to their own will. This EEG dataset provides a realistic representation of reaching and grasping movement, making it useful for developing practical BMIs.},
}

Humans
*Electroencephalography
*Brain-Computer Interfaces
Male
Female
Young Adult
Adolescent
*Hand Strength

@article {pmid41201930,
year = {2025},
author = {Yang, Y and Wang, Z and Jia, Z and Wang, B and Zhang, S and Wong, CM and Gao, X and Jung, TP and Wan, F},
title = {Dual-Branch Attention-based Frequency Domain Network for Cross-subject SSVEP-BCIs.},
journal = {IEEE journal of biomedical and health informatics},
volume = {PP},
number = {},
pages = {},
doi = {10.1109/JBHI.2025.3630249},
pmid = {41201930},
issn = {2168-2208},
abstract = {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. The code is available at https://github.com/YYingDL/DBAFDNet.},
}

@article {pmid41199758,
year = {2025},
author = {Wang, Y and Yu, H and Zhao, X and Yin, X and Li, H and Wang, C},
title = {A dual-branch neural network and attention mechanism for decoding EEG-based motor imagery.},
journal = {Cognitive neurodynamics},
volume = {19},
number = {1},
pages = {177},
pmid = {41199758},
issn = {1871-4080},
abstract = {Motor imagery (MI) is a fundamental paradigm in brain-computer interfaces (BCIs), extensively employed to assist individuals with disabilities to operate external devices. Accurate decoding of MI signals is essential for effective interaction. However, robust decoding remains a challenge due to the inherent complexity and variability of MI-EEG signals. To address this issue, we propose an innovative Dual-Branch Multi-Attention Temporal Convolutional Network (DBMATCN) to improve the performance of MI-EEG signal classification. First, the dual-branch structure extracts rich spatial-temporal features. Then, the channel attention enhances local channel feature extraction and calibrate feature mapping. Next, by combining a sliding window technique and multi-head locality self-attention improves the feature representation of MI-EEG signals by emphasizing the most relevant features. Finally, the temporal convolution fusion network decoding module is used to extensively capture comprehensive temporal features from MI data and carry out the classification task. DBMATCN achieves average accuracies of 88.08%, 96.83%, and 89.71% in inter-session validation on the BCI-IV-2a, HGD, and BCI-IV-2b datasets, respectively. In cross-validation, the model reaches an accuracy of 85.14%, and in the subject-independent scenario, it attains 71.78%. DBMATCN outperforms all baseline models in these cases. These results suggest that our model is effective in decoding MI signals.},
}

@article {pmid41199757,
year = {2025},
author = {Zhu, L and Ding, Y and Hung, A and Tan, X and Zhang, J},
title = {SFT-HN: a novel spatial-frequency-temporal hybrid network for EEG-based emotion recognition.},
journal = {Cognitive neurodynamics},
volume = {19},
number = {1},
pages = {176},
pmid = {41199757},
issn = {1871-4080},
abstract = {Electroencephalograph (EEG) emotion recognition is a key task in the brain-computer interface(BCI) field. A mounting quantity of studies have shown that deep learning methods for emotion recognition exhibit superior performance compared to traditional techniques. However, it is still challenging to fuse the EEG's Spatial, Frequency and Temporal information, as well as how to make full use of discriminative local patterns among the features for different emotions. To address these issues, a novel hybrid model called Spatial-Frequency-Temporal Hybrid Network(SFT-HN) is proposed. This model includes three Spatial Frequency Residual Modules (SFRM) and an attention-based Bidirectional Long Short-Term Memory (ATBI-LSTM). The former module extracts spatial-frequency features, while the latter learns temporal contexts. SFT-HN is trained to seize the complementarity among the spatial-frequency-temporal information and adaptively explore discriminative local patterns. Specifically, 4D representations are created from raw EEG signals to preserve spatial, frequency, and temporal information. The SFRM module then adopts split-convert-merge techniques, residual and attention mechanisms to enhance its spatial-frequency feature extraction ability for each input 4D representation tensor time slice. Moreover, an attention-enhanced mechanism is incorporated into a bidirectional LSTM module to capture the crucial temporal dependencies among the extracted features, thereby enhancing the discriminative power of the EEG features. The proposed method attains average accuracies of 97.61% and 97.57% for arousal-based and valence-based classification on the DEAP dataset, respectively. On SEED dataset, the method achieves average accuracy of 97.44%. Furthermore, we validate the robust generalization of our proposed model on a novel dataset, FACED, achieving an average accuracy of 96.24%. The model code is available at: https://github.com/AllGGI/SFT-HN-model.},
}

@article {pmid41199756,
year = {2025},
author = {Wu, X and Long, D and Yang, J},
title = {Generative motor imagery dynamic networks: EEG-controlled grasping via individualized model training.},
journal = {Cognitive neurodynamics},
volume = {19},
number = {1},
pages = {174},
pmid = {41199756},
issn = {1871-4080},
abstract = {Improving the accuracy of non-invasive brain-computer interface (BCI) and promoting their daily use can be achieved by developing an individualized model training framework, where individual training means that the model is based on small-sample learning from individual data. In the process of data augmentation through synthetic data, the criteria for data generation needs to be further specified according to the requirements. Therefore, in this study, the proposed BCI model utilizes dynamic networks to describe electroencephalogram (EEG) activity during the motor imagery (MI) task, innovatively generates individualized dynamic networks from individual data, and ultimately achieves EEG-controlled grasping through model training. Specifically, this study involves the EEG signals of the right-hand grasping movements of eight subjects and proposes using morphological pattern spectrum (MPS) to encode EEG potentials during MI processes. The MI condition representation was achieved by combining the dynamic networks with MPS encoding, and more dynamic network EEG encoding samples were synthesized through generative adversarial network (GAN) or variational autoencoder (VAE). The AUCs based on the long short-term memory (LSTM) architecture for generating and classifying can be improved by 0.003-0.07. The optimal BCI model based on the Wasserstein GAN and Granger causality (GC) dynamic network encoded by MPS achieved a mean true/false positive rate (TPR/FPR) of 90.0%/0.0%, far better than the 52.9%/4.4% achieved without individualized modeling. Moreover, the BCI establishment of handling multi-task and complex command outputs further demonstrates the reliability of MPS encoding of the GC dynamic network in BCI modeling. The advantage of this "generative-individual" approach is that it not only reduces the sample size requirement while ensuring accuracy but also avoids building models that are applicable to all individuals, which leads to difficult convergence.},
}

@article {pmid41199478,
year = {2025},
author = {Bobby J, S and V Francis, S and Ramya V, S and C L, A},
title = {Preliminary Findings on a Deep Learning Model Using Electroencephalogram for Multi-Level Neuropathic Pain Detection in Post-Stroke Patients.},
journal = {The International journal of neuroscience},
volume = {},
number = {},
pages = {1-10},
doi = {10.1080/00207454.2025.2584081},
pmid = {41199478},
issn = {1563-5279},
abstract = {AIM: Neuropathic pain occurs commonly after stroke and represents a major source of disability for affected patients. This study aims to develop an accurate and computationally efficient framework for multi-level neuropathic pain detection using electroencephalography signals.

METHODS: A Quantum-Inspired Pyramid Depthwise Separable Residual Network is proposed, which integrates three innovations: a depthwise separable Residual Network to reduce computational complexity, a pyramid attention mechanism to capture multi-scale patterns, and a quantum-inspired transformation layer to model complex nonlinear dependencies among Electroencephalogram features.

RESULTS: Experiments conducted on benchmark electroencephalography datasets confirm that the proposed model gains a accuracy of 99.65%, with a recall of 98.00%.

CONCLUSION: The proposed model provides a reliable solution for objective neuropathic pain detection in post-stroke patients. The framework demonstrates potential for integration into intelligent clinical decision-support and brain-computer interface-based rehabilitation systems.},
}

@article {pmid41199005,
year = {2025},
author = {Zhang, H and Wang, X and Xi, K and Shen, Q and Xue, J and Zhu, Y and Zang, SK and Yu, T and Shen, DD and Guo, J and Chen, LN and Ji, SY and Qin, J and Dong, Y and Zhao, M and Yang, M and Wu, H and Yang, G and Zhang, Y},
title = {The molecular basis of μ-opioid receptor signaling plasticity.},
journal = {Cell research},
volume = {},
number = {},
pages = {},
pmid = {41199005},
issn = {1748-7838},
support = {2019YFA0508800//Chinese Ministry of Science and Technology | Department of S and T for Social Development (Department of S&T for Social Development)/ ; 32430051//National Natural Science Foundation of China (National Science Foundation of China)/ ; 92353303//National Natural Science Foundation of China (National Science Foundation of China)/ ; 32141004//National Natural Science Foundation of China (National Science Foundation of China)/ ; 82271001//National Natural Science Foundation of China (National Science Foundation of China)/ ; 2024M752856//China Postdoctoral Science Foundation/ ; },
abstract = {Activation of the μ-opioid receptor (μOR) alleviates pain but also elicits adverse effects through diverse G proteins and β-arrestins. The structural details of μOR complexes with Gz and β-arrestins have not been determined, impeding a comprehensive understanding of μOR signaling plasticity. Here, we present the cryo-EM structures of the μOR-Gz and μOR-βarr1 complexes, revealing selective conformational preferences of μOR when engaged with specific downstream signaling transducers. Integrated receptor pharmacology, including high-resolution structural analysis, cell signaling assays, and molecular dynamics simulations, demonstrated that transmembrane helix 1 (TM1) acts as an allosteric regulator of μOR signaling bias through differential stabilization of the Gi-, Gz-, and βarr1-bound states. Mechanistically, outward TM1 displacement confers structural flexibility that promotes G protein recruitment, whereas inward TM1 retraction facilitates βarr1 recruitment by stabilizing the intracellular binding pocket through coordinated interactions with TM2, TM7, and helix8. Structural comparisons between the Gi-, Gz-, and βarr1-bound complexes identified a TM1-fusion pocket with significant implications for downstream signaling regulation. Overall, we demonstrate that the conformational and thermodynamic heterogeneity of TM1 allosterically drives the downstream signaling specificity and plasticity of μOR, thereby expanding the understanding of μOR signal transduction mechanisms and providing new avenues for the rational design of analgesics.},
}

@article {pmid41198855,
year = {2025},
author = {Heerspink, HJL and Collier, WH and Chaudhari, J and Miao, S and Tighiouart, H and Appel, GB and Caravaca-Fontán, F and Floege, J and Hannedouche, T and Imai, E and Jafar, TH and Lewis, JB and Li, PKT and Locatelli, F and Maes, BD and Neuen, BL and Perkovic, V and Perrone, RD and Remuzzi, G and Schena, FP and Wanner, C and Greene, T and Inker, LA},
title = {A meta-analysis of albuminuria as a surrogate endpoint for kidney failure.},
journal = {Nature medicine},
volume = {},
number = {},
pages = {},
pmid = {41198855},
issn = {1546-170X},
abstract = {Albuminuria is a central biomarker in chronic kidney disease (CKD), used for the detection and prognosis of the disease. In clinical trials assessing CKD progression, change in the level of albuminuria is a candidate surrogate endpoint for kidney failure. Evaluation of the validity of this surrogate endpoint across a diverse range of interventions and populations is required to support its further acceptance. Here, in an individual participant data analysis of 48 randomized controlled trials (studies) involving 85,681 participants, we assessed the association between treatment effects on 6-month urinary albumin:creatinine ratio (UACR) change and the established clinical endpoint of kidney failure or doubling of serum creatinine concentrations. Across all trials, each 30% reduction in the geometric mean of the UACR in the treatment group relative to the control group was associated with an average of 19% lower hazard for the clinical endpoint (95% Bayesian credible interval (BCI): 5-30%); median coefficient of determination (R[2]) = 0.66 (95% BCI: 0.06-0.98). There was no clear evidence that this association varied by CKD etiology. These results provide further support for use of albuminuria change as a surrogate endpoint in CKD clinical trials.},
}

@article {pmid41197164,
year = {2025},
author = {Cinquetti, E and Menegaz, G and Storti, SF},
title = {Toward in-silico data assessment for passive BCIs: Generating EEG rhythms with GANs.},
journal = {Journal of neural engineering},
volume = {},
number = {},
pages = {},
doi = {10.1088/1741-2552/ae1c6f},
pmid = {41197164},
issn = {1741-2552},
abstract = {Passive brain-computer interface based on electroencephalography (EEG) has gained traction as reliable method for monitoring human vigilance in attention-demanding critical contexts. Unfortunately, the lack of extensive public datasets compromises artificial intelligence (AI) research. Proposing a solution to this issue, we augmented two EEG datasets using generative adversarial networks (GAN) and defined a quality-assessment pipeline to overcome the absence of a univocal method to test synthetic data. Approach. Using GAN, we augmented a publicly resting-state EEG dataset (SPIS) and a custom one simulating activity during repetitive tasks. After extracting relevant time-variant rhythms via the continuous wavelet transform, we quantitatively compared synthetic data with the real one using L2 distance and cross-correlation function. To evaluate the impact of data augmentation, we trained six forecasting models, three on the original and three on the augmented datasets, over the whole, half and a quarter of total available data, and compared improvements in MAE and SMAPE. To study the forecaster's embeddings, we computed a metric inspired by the Fréchet Inception Distance (FID) between latent values of real and synthetic data. Finally, to offer a baseline comparison, we extended the performance and embeddings analysis to data generated by a simple linear interpolation method. Main Results. The integration of GAN-produced synthetic data improved signal prediction, as evidenced by a 29.0%, 46.4%, 37.4% reduction in mean absolute error (MAE) for splits of the resting-state dataset, and an average MAE reduction of 15.4%, 21.2% for 100% and 50% splits, and a ∽-2.5\% increase for the 25% split). Conversely, training on interpolated data manifest worse performance and denotes extremely small FID distances w.r.t real signals, a sign of overspecialization. Significance. This study contributes a reproducible and complete framework for EEG signal generation and evaluation, addressing one of the main barriers to scalable AI application in BCI.},
}

@article {pmid41196745,
year = {2025},
author = {Zhang, S and Chen, W and Chang, S and Zhou, LF and Ding, X},
title = {How visual imagery representations are formed: Through suppression, not activation.},
journal = {Journal of experimental psychology. General},
volume = {},
number = {},
pages = {},
doi = {10.1037/xge0001863},
pmid = {41196745},
issn = {1939-2222},
support = {//National Natural Science Foundation of China/ ; //Natural Science Foundation of Guangdong Province/ ; //Guangzhou Science and Technology Plan Project-Leading Elite Program/ ; //Fundamental Research Funds for the Central Universities of China/ ; //Major Project Cultivation and Emerging Interdisciplinary Cultivation Plan/ ; //Shanghai Science and Technology Development Foundation/ ; },
abstract = {Voluntary imagery is described as "weak perception" and is thought to be represented through activating the neurons corresponding to imagined features, that is, activation hypothesis. However, direct evidence for this hypothesis is lacking. Inspired by Pace et al. (2023), we examine an alternative suppression hypothesis, which states imagery involves suppression of neurons favoring nearby nonimagined features. While the activation hypothesis predicts a bell-shaped tuning curve of the neural representation for the imagined feature, the suppression hypothesis predicts a W-shaped tuning curve. To test these two hypotheses, we combined an imagery task with a discrimination task following the logic that different imagery-induced tuning curves would differently bias the perceived difference in the discrimination task. We probed the bias pattern by systematically manipulating the physical orientation difference and the discrimination-imagery relation condition. A series of psychophysical experiments were conducted. Results showed that after an imagery prior, bias pattern in the discrimination task followed the prediction of suppression hypothesis (Experiment 1a). By contrast, when substituting the imagery prior with a strong/weak perceptual prior, bias pattern was consistent with the prediction of activation hypothesis (Experiments 2a and 2b). Confounding effects of visual attention and perceptual imagery cue were excluded (Experiments 1b and 1c). We further constructed mathematical models and again validated our findings. In conclusion, behavioral and modeling results coherently suggested that the suppression hypothesis was a better explanation for imagery than the activation hypothesis. Our study challenges the traditional activation theory and provides novel empirical evidence for the suppressive representation of voluntary visual imagery. (PsycInfo Database Record (c) 2025 APA, all rights reserved).},
}

@article {pmid41193512,
year = {2025},
author = {Kim, H and Won, K and Ahn, M and Jun, SC},
title = {A 40-Class SSVEP Speller Dataset: Beta Range Stimulation for Low-Fatigue BCI Applications.},
journal = {Scientific data},
volume = {12},
number = {1},
pages = {1751},
pmid = {41193512},
issn = {2052-4463},
support = {RS-2024-00361688//National Research Foundation of Korea (NRF)/ ; },
mesh = {Humans ; *Brain-Computer Interfaces ; Electroencephalography ; *Evoked Potentials, Visual ; *Fatigue ; Adult ; Photic Stimulation ; Male ; Female ; },
abstract = {The inherent non-stationarity of electroencephalography (EEG) signals necessitates large, consistent datasets for reliable brain-computer interface (BCI) research. In steady-state visual evoked potential (SSVEP) paradigms, prolonged exposure to visual stimuli can induce visual fatigue, leading to alterations in EEG patterns that degrade BCI performance. To mitigate fatigue-induced variability, this study employs visual stimulation in the beta frequency range (14-22 Hz), a range that appears less susceptible to the effects of fatigue. We present a comprehensive 40-class SSVEP speller dataset acquired from 40 participants, with EEG data recorded from 31 central-to-occipital channels. Each subject completed six sessions of the SSVEP speller task in addition to pre- and post-experiment resting-state recordings under both eyes-open and eyes-closed conditions. Subjective fatigue ratings combined with EEG band power analyses confirm that beta-range stimulation minimizes fatigue effects. Moreover, the high classification accuracy achieved by calibration-based algorithms indicates that the dataset is well-suited for training advanced SSVEP-based BCI systems.},
}

Humans
*Brain-Computer Interfaces
Electroencephalography
*Evoked Potentials, Visual
*Fatigue
Adult
Photic Stimulation
Male
Female

@article {pmid41193466,
year = {2025},
author = {Chiti, E and Micera, S and Palmerini, E},
title = {Making the case for sandboxes in implantable neurotechnologies.},
journal = {Nature communications},
volume = {16},
number = {1},
pages = {9783},
pmid = {41193466},
issn = {2041-1723},
abstract = {Regulatory sandboxes could be fruitfully used to boost Invasive Brain-Computer Interfaces, but they should be carefully designed. We highlight five elements are essential: they concern the entry criteria, the participated, adaptive and supervised design of decision-making process, and long-term risk management.},
}

@article {pmid41192731,
year = {2025},
author = {Dai, C and Lin, M and Xu, N and Fu, Y and Li, X and Shi, Y and Wu, M and Li, Y and Xie, J and Hu, S and Zhao, Q},
title = {The impact of CYP3A4 rs2242480 on oral lurasidone: A population pharmacokinetic model and exposure-efficacy analysis in Chinese bipolar depression patients.},
journal = {Journal of affective disorders},
volume = {},
number = {},
pages = {120588},
doi = {10.1016/j.jad.2025.120588},
pmid = {41192731},
issn = {1573-2517},
abstract = {OBJECTIVE: This study aims to develop a population pharmacokinetic (PPK) model and perform an exposure-efficacy analysis for lurasidone in patients with bipolar depression, thus interpretating the inter-individual variability in its pharmacokinetics and optimizing dosing regimens.

METHODS: A PPK model and exposure-efficacy analysis were established in Chinese patients with bipolar depression. 241 lurasidone concentration measurments from 133 patients were included. Demographic information was collected and genotypes for CYP3A4 and HTR1A alleles were determined. Treatment efficacy was defined as the reduction in the Montgomery-Asberg Depression Rating Scale (MADRS) score at week 4.

RESULTS: A one-compartment model with first-order kinetics for lurasidone was fitted. The apparent clearance (CL/F) of lurasidone was significantly lower in CYP3A4 rs2242480 CC carriers (330 L/h) than in TC (385 L/h) and TT (441 L/h) carriers, representing reductions of 14.3 % and 25.2 %, respectively. Additionally, CL/F was positively correlated with ideal body weight (IBW). Incorporating these covariates reduced the interindividual variability in CL/F from 40.5 % to 37.1 %. The exposure-efficacy analysis demonstrated a dose-denpedent increase in area under the curve (AUC), and MADRS score improved with an increasing AUC and reached a plateau at an AUC of approximately 167 mg·h·L[-1], corresponding to an optimal daily dose range of 45-55 mg.

CONCLUSION: The pharmacokinetics of lurasidone in patients with bipolar depression are significantly influenced by IBW and the rs2242480 genotype, enabling a practical framework for precision dosing.},
}

@article {pmid41192010,
year = {2025},
author = {Yang, M and Wang, Z and Zhou, Q and Zhang, Q and Li, Y and Wang, Z},
title = {The adjunctive efficacy of repetitive transcranial magnetic stimulation with non-pharmacological interventions in cognitive disorders: A meta-analysis of randomized sham-controlled trials.},
journal = {Asian journal of psychiatry},
volume = {114},
number = {},
pages = {104758},
doi = {10.1016/j.ajp.2025.104758},
pmid = {41192010},
issn = {1876-2026},
abstract = {OBJECTIVE: This meta-analysis aimed to systematically evaluate the specific, adjunctive efficacy of repetitive transcranial magnetic stimulation (rTMS) when combined with non-pharmacological interventions-namely, transcranial direct current stimulation (tDCS), Tai Chi, or cognitive training (CT)-in patients with Alzheimer's disease (AD) or mild cognitive impairment (MCI). The goal is to isolate the net therapeutic contribution of rTMS beyond the effects of the base interventions alone.

METHODS: A comprehensive search of Chinese and English databases was conducted from their inception until April 26, 2025. Randomized controlled trials (RCTs) that compared "a non-pharmacological intervention plus active rTMS" versus "the same non-pharmacological intervention plus sham rTMS".This "add-on" study design was selected to precisely isolate the effect of rTMS. The risk of bias was assessed using the PEDro scale and Cochrane tools. Statistical analyses were performed using Review Manager 5.4 software.

RESULTS: 9 studies involving 391 participants were included. The pooled analysis revealed that the adjunctive use of rTMS was significantly superior to the sham control in improving global cognitive function at the immediate post-treatment assessment (SMD=0.38, 95 %CI[0.20,0.56], P < .001, n = 9). This benefit was consistent across the MMSE (SMD=0.38, n = 6), MoCA (SMD=0.37, n = 2), and ADAS-cog (SMD=0.39, n = 3) scores. Subgroup analysis suggested that the rTMS-tDCS combination might offer a short-term advantage in improving MMSE scores (MD=4.67, P = .008). Furthermore, the adjunctive effect of rTMS was sustained, as particularly evidenced by the ADAS-cog at follow-up (SMD=0.74, P = .02). The pooled analysis indicated that rTMS combined with non-pharmacological therapy demonstrated a short-term, sustained (4-8weeks) improvement in global cognitive function (SMD=0.34, 95 % CI[0.07, 0.60], P = .01). Subgroup analysis revealed that this sustained benefit reached statistical significance on the ADAS-cog scale (SMD = 0.41, 95 %CI[0.01, 0.81], P = .04) but showed a non-significant positive trend on the MMSE (SMD=0.26, 95 %CI[-0.19, 0.72], P = .26). However, a key limitation was that most studies did not systematically report outcomes related to activities of daily living or behavioral function.

CONCLUSION: The evidence indicates that rTMS as an adjunct to non-pharmacological interventions provides a significant specific effect on global cognitive function in patients with AD and MCI shortly after treatment, which may be sustained in the short-term. However, long-term follow-up data are extremely limited, and the effect on activities of daily living remains to be validated. The combination of rTMS and tDCS shows promise,but conclusions are constrained by the small number of studies,limited sample sizes,and heterogeneity in intervention protocols. Future large-scale studies incorporating long-term, standardized follow-up and assessments of daily living abilities are warranted to confirm the specific clinical value of rTMS as an augmentative therapy.},
}

@article {pmid41191990,
year = {2025},
author = {Xing, Y and He, Y and Gong, Z and Zhou, J and Sun, Y and Zhong, Z},
title = {A study on the microstructure and micromechanical properties of Drosophila larval cuticle using scanning probe microscopy and viscoelastic modeling.},
journal = {Journal of biomechanics},
volume = {194},
number = {},
pages = {113051},
doi = {10.1016/j.jbiomech.2025.113051},
pmid = {41191990},
issn = {1873-2380},
abstract = {The Drosophila larval cuticle exhibits compliant yet resilient viscoelasticity, serving as a soft exoskeleton that enables effective locomotion while maintaining structural integrity. Investigating its microstructure and micromechanical properties not only advances our understanding of soft-bodied biomechanics but also guides the design of biomimetic materials and soft robotic systems. In this study, we employed scanning probe microscopy (SPM)-based stress relaxation tests to characterize viscoelastic properties across the denticle and smooth skin bands in three larval instars. Four viscoelastic models were evaluated, and the five-element Maxwell (MX5) model provided the best fit, enabling the extraction of mechanical parameters and plotting of relaxation modulus functions. Results showed that the larval instar stage had minimal influence on viscoelasticity, while the denticle and smooth skin bands exhibited distinct mechanical behaviors. Across all instars, the denticle bands showed higher moduli throughout the relaxation process, and notably, exhibited a greater degree and faster rate of relaxation compared to the smooth skin bands. These findings reveal region-specific viscoelastic adaptations that enable rapid stress dissipation while maintaining stiffness, supporting effective deformation during locomotion. This study provides essential quantitative foundations for bioinspired stretchable electronics, soft robotic materials, and broader understanding of soft exoskeleton mechanics.},
}

@article {pmid41191976,
year = {2025},
author = {Forrest, A and Kunigk, NG and Collinger, J and Gaunt, RA and Vande Geest, JP and Chen, X and Kozai, TDY},
title = {Finite element model predicts micromotion-induced strain profiles that correlate with the functional performance of Utah arrays in humans and non-human primates.},
journal = {Journal of neural engineering},
volume = {},
number = {},
pages = {},
doi = {10.1088/1741-2552/ae1bda},
pmid = {41191976},
issn = {1741-2552},
abstract = {OBJECTIVE: Utah arrays are widely used in both humans and non-human primates (NHPs) for intracortical brain-computer interfaces (BCIs), primarily for detecting electrical signals from cortical tissue to decode motor commands. Recently, these arrays have also been applied to deliver electrical stimulation aimed at restoring sensory functions. A key challenge limiting their longevity is the micromotion between the array and cortical tissue, which may induce mechanical strain in surrounding tissue and contribute to performance decline. This strain, due to mechanical mismatch, can exacerbate glial scarring around the implant, reducing the efficacy of Utah arrays in recording neuronal activity and delivering electrical stimulation.

APPROACH: To investigate this, we employed a finite element model (FEM) to predict tissue strains resulting from micromotion.

MAIN RESULTS: Our findings indicated that strain profiles around edge and corner electrodes were greater than those around interior shanks, affecting both maximum and average strains within 50 µm of the electrode tip. We then correlated these predicted tissue strains with in-vivo electrode performance metrics. We found negative correlations between 1 kHz impedance and tissue strains in human motor arrays and NHP area V4 arrays at 1-mo, 1-yr, and 2-yrs post-implantation. In human motor arrays, the peak-to-peak waveform voltage (PTPV) and signal-to-noise ratio (SNR) of spontaneous activity were also negatively correlated with strain. Conversely, we observed a positive correlation between the evoked SNR of multi-unit activity and strain in NHP area V4 arrays.

SIGNIFICANCE: This study establishes a spatial dependence of electrode performance in Utah arrays that correlates with tissue strain.},
}

@article {pmid41191971,
year = {2025},
author = {Bjanes, D and Bashford, L and Pejsa, K and Lee, B and Liu, CY and Andersen, RA},
title = {Charge density of multi-channel intra-cortical micro-stimulation modulates intensity and naturalness of evoked somatosensations.},
journal = {Journal of neural engineering},
volume = {},
number = {},
pages = {},
doi = {10.1088/1741-2552/ae1bd8},
pmid = {41191971},
issn = {1741-2552},
abstract = {Human patients with somatosensory loss often experience severe motor deficits, causing profound challenges to independently accomplish typical tasks of daily life. Brain-machine Interfaces (BMIs) offer the potential to restore lost functionality through direct electrical stimulation of the somatosensory cortex via intra-cortical micro-stimulation (ICMS). By modulating temporal patterns of stimulation, our group has previously shown single-channel ICMS can evoke both naturalistic cutaneous and proprioceptive sensory feedback. However, accurate modulation of the sensory feedback's qualia (somatotopic location, intensity and description) will be critical for fluid, dexterous motor control. In nonhuman primate studies, multi-channel ICMS has shown promise in improving quantifiable metrics such as reaction time. In recent human work, multi-channel ICMS has improved discrimination performance; however, evoked qualia has not been well characterized. We hypothesized multi-channel ICMS could evoke unique qualia compared to single-channel. A human participant with tetraplegia and chronically implanted microelectrode arrays in primary somatosensory cortex, reported perceptual thresholds, sensation descriptions, intensity and somatotopic locations of single- and multi-channel ICMS patterns. We found multi-channel ICMS patterns evoked unique qualia compared to single-channel ICMS. To investigate the role of charge in producing these unique evoked sensory percepts, we delivered equal amounts of charge with differing spatial patterns across multiple electrodes. Multi-channel ICMS substantially reduced the minimum stimulation amplitude required to evoked somatosensations, lowering the charge per electrode detection threshold, while increasing the total charge injected. Delivered charge across multiple electrodes, positively modulated the sensation's perceived intensity; providing early evidence of spatial integration of ICMS in the target network. Multi-channel ICMS resulted in more frequent verbal reports of "natural" sensation descriptors (100% vs 85% for single-channel ICMS, p-val<0.05) and robustly evoked sensations with high repeatability in stable somatotopic locations. Multi-channel ICMS patterns demonstrated improvements in reliability, somatotopic coverage and "natural-ness" of the evoked sensations, marking significant advances towards state-of-the-art somatosensory brain-machine-interfaces (BMIs). By better understanding of the input/output relationship for somatosensory feedback BMIs, we can expect to improve movement accuracy and increase embodiment for human users. .},
}

@article {pmid41191851,
year = {2025},
author = {Williams, C and Anik, FI and Hasan, MM and Rodriguez-Cardenas, J and Chowdhury, A and Tian, S and He, S and Sakib, N},
title = {Advancing Brain-Computer Interface Closed-Loop Systems for Neurorehabilitation: Systematic Review of AI and Machine Learning Innovations in Biomedical Engineering.},
journal = {JMIR biomedical engineering},
volume = {10},
number = {},
pages = {e72218},
doi = {10.2196/72218},
pmid = {41191851},
issn = {2561-3278},
abstract = {BACKGROUND: Brain-computer interface (BCI) closed-loop systems have emerged as a promising tool in health care and wellness monitoring, particularly in neurorehabilitation and cognitive assessment. With the increasing burden of neurological disorders, including Alzheimer disease and related dementias (AD/ADRD), there is a critical need for real-time, noninvasive monitoring technologies. BCIs enable direct communication between the brain and external devices, leveraging artificial intelligence (AI) and machine learning (ML) to interpret neural signals. However, challenges such as signal noise, data processing limitations, and privacy concerns hinder widespread implementation.

OBJECTIVE: The primary objective of this study is to investigate the role of ML and AI in enhancing BCI closed-loop systems for health care applications. Specifically, we aim to analyze the methods and parameters used in these systems, assess the effectiveness of different AI and ML techniques, identify key challenges in their development and implementation, and propose a framework for using BCIs in the longitudinal monitoring of AD/ADRD patients. By addressing these aspects, this study seeks to provide a comprehensive overview of the potential and limitations of AI-driven BCIs in neurological health care.

METHODS: A systematic literature review was conducted following PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines, focusing on studies published between 2019 and 2024. We sourced research articles from PubMed, IEEE, ACM, and Scopus using predefined keywords related to BCIs, AI, and AD/ADRD. A total of 220 papers were initially identified, with 18 meeting the final inclusion criteria. Data extraction followed a structured matrix approach, categorizing studies based on methods, ML algorithms, limitations, and proposed solutions. A comparative analysis was performed to synthesize key findings and trends in AI-enhanced BCI systems for neurorehabilitation and cognitive monitoring.

RESULTS: The review identified several ML techniques, including transfer learning (TL), support vector machines (SVMs), and convolutional neural networks (CNNs), that enhance BCI closed-loop performance. These methods improve signal classification, feature extraction, and real-time adaptability, enabling accurate monitoring of cognitive states. However, challenges such as long calibration sessions, computational costs, data security risks, and variability in neural signals were also highlighted. To address these issues, emerging solutions such as improved sensor technology, efficient calibration protocols, and advanced AI-driven decoding models are being explored. In addition, BCIs show potential for real-time alert systems that support caregivers in managing AD/ADRD patients.

CONCLUSIONS: BCI closed-loop systems, when integrated with AI and ML, offer significant advancements in neurological health care, particularly in AD/ADRD monitoring and neurorehabilitation. Despite their potential, challenges related to data accuracy, security, and scalability must be addressed for widespread clinical adoption. Future research should focus on refining AI models, improving real-time data processing, and enhancing user accessibility. With continued advancements, AI-powered BCIs can revolutionize personalized health care by providing continuous, adaptive monitoring and intervention for patients with neurological disorders.},
}

@article {pmid41191764,
year = {2025},
author = {Qian, Y and Liu, C and Yu, P and Ran, X and Li, S and Yang, Q and Liu, Y and Xia, L and Wang, Y and Qi, J and Zhou, E and Lu, J and Li, Y and Tao, TH and Zhou, Z and Wu, J},
title = {Real-time decoding of full-spectrum Chinese using brain-computer interface.},
journal = {Science advances},
volume = {11},
number = {45},
pages = {eadz9968},
doi = {10.1126/sciadv.adz9968},
pmid = {41191764},
issn = {2375-2548},
mesh = {*Brain-Computer Interfaces ; Humans ; Language ; Male ; *Speech/physiology ; Female ; Adult ; Electroencephalography ; China ; East Asian People ; },
abstract = {Speech brain-computer interfaces (BCIs) offer a promising means to provide functional communication capacity for patients with anarthria caused by neurological conditions such as amyotrophic lateral sclerosis (ALS) or brainstem stroke. Current speech decoding research has predominantly focused on English using phoneme-driven architectures, whereas real-time decoding of tonal monosyllabic languages such as Mandarin Chinese remains a major challenge. This study demonstrates a real-time Mandarin speech BCI that decodes monosyllabic units directly from neural signals. Using the 256-channel microelectrocorticographic BCI, we achieved robust decoding of a comprehensive set of 394 distinct syllables based purely on neural signals, yielding median syllable identification accuracy of 71.2% in a single-character reading task. Leveraging this high-performing syllable decoder, we further demonstrated real-time sentence decoding. Our findings demonstrate the efficacy of a tonally integrated, direct syllable neural decoding approach for Mandarin Chinese, paving the way for full-coverage systems in tonal monosyllabic languages.},
}

*Brain-Computer Interfaces
Humans
Language
Male
*Speech/physiology
Female
Adult
Electroencephalography
China
East Asian People

@article {pmid41187598,
year = {2025},
author = {Jui, JJ and Hettiarachchi, IT and Bhatti, A and Creighton, D},
title = {PLVNet: EEG-based trust classification using Phase Locking Value connectivity and deep neural networks.},
journal = {Computers in biology and medicine},
volume = {198},
number = {Pt B},
pages = {111269},
doi = {10.1016/j.compbiomed.2025.111269},
pmid = {41187598},
issn = {1879-0534},
abstract = {Trust in automation is critical for effective human-automation interaction, yet traditional subjective measures are limited in capturing rapid and dynamic changes in user trust. This study introduces PLVNet, a novel deep neural network architecture designed to classify trust versus distrust states from EEG functional connectivity features derived using Phase Locking Value (PLV). PLV features were extracted across six canonical EEG frequency bands (Delta, Theta, Alpha, Beta, Low Gamma, High Gamma) from 30-channel EEG recordings. The PLVNet model was evaluated using three complementary approaches: aggregated analysis (5× 5 stratified cross-validation), participant-wise analysis, and leave-one-subject-out (LOSO) cross-validation. PLVNet significantly outperformed convolutional neural network (CNN), support vector machine (SVM) and k-nearest neighbours (KNN) classifiers across all evaluation schemes. Beta and Low Gamma bands provided the highest discriminative power, while functional connectivity analysis revealed that trust is associated with enhanced fronto-parietal and fronto-occipital synchronisation, reflecting global network integration, whereas distrust shows fragmented connectivity patterns. PLVNet's ability to capture non-linear inter-dependencies in connectivity patterns highlights its advantages over conventional methods. These findings demonstrate that PLV-based connectivity robustly reflects trust-related neural dynamics, underscoring the potential of PLVNet for real-time, objective monitoring of trust in human-automation systems, which paves the way for adaptive and neuro-aware interfaces.},
}

@article {pmid41187597,
year = {2025},
author = {Liu, J and Deng, X and Li, H and Kazemi, A and Grashei, C and Wilkens, G and You, X and Groll, T and Navab, N and Mogler, C and Schüffler, PJ},
title = {From pixels to pathology: Restoration diffusion for diagnostic-consistent virtual IHC.},
journal = {Computers in biology and medicine},
volume = {198},
number = {Pt B},
pages = {111264},
doi = {10.1016/j.compbiomed.2025.111264},
pmid = {41187597},
issn = {1879-0534},
abstract = {Hematoxylin and eosin (H&E) staining is the clinical standard for assessing tissue morphology, but it lacks molecular-level diagnostic information. In contrast, immunohistochemistry (IHC) provides crucial insights into biomarker expression, such as HER2 status for breast cancer grading, but remains costly and time-consuming, limiting its use in time-sensitive clinical workflows. To address this gap, virtual staining from H&E to IHC has emerged as a promising alternative, yet faces two core challenges: (1) Lack of fair evaluation of synthetic images against misaligned IHC ground truths, and (2) preserving structural integrity and biological variability during translation. To this end, we present an end-to-end framework encompassing both generation and evaluation in this work. We introduce Star-Diff, a structure-aware staining restoration diffusion model that reformulates virtual staining as an image restoration task. By combining residual and noise-based generation pathways, Star-Diff maintains tissue structure while modeling realistic biomarker variability. To evaluate the diagnostic consistency of the generated IHC patches, we propose the Semantic Fidelity Score (SFS), a clinical-grading-task-driven metric that quantifies class-wise semantic degradation based on biomarker classification accuracy. Unlike pixel-level metrics such as SSIM and PSNR, SFS remains robust under spatial misalignment and classifier uncertainty. Experiments on the BCI dataset demonstrate that Star-Diff achieves state-of-the-art (SOTA) performance in both visual fidelity and diagnostic relevance. With rapid inference and strong clinical alignment, it presents a practical solution for applications such as intraoperative virtual IHC synthesis.},
}

@article {pmid41187327,
year = {2025},
author = {Bialostocki, LS and Adhia, DB and Mudiyanselage, DR and Smith, ML and Cakmak, YO and De Ridder, D and Mani, R and Mathew, J},
title = {Neurofeedback Training for Managing Neuropathic Pain-Like Features in Chronic Musculoskeletal Pain: Protocol for an Open-Label Pilot Feasibility Clinical Trial.},
journal = {JMIR research protocols},
volume = {14},
number = {},
pages = {e78806},
doi = {10.2196/78806},
pmid = {41187327},
issn = {1929-0748},
mesh = {Humans ; *Neurofeedback/methods ; Pilot Projects ; *Neuralgia/therapy/physiopathology ; Feasibility Studies ; *Musculoskeletal Pain/therapy/physiopathology ; Electroencephalography/methods ; *Chronic Pain/therapy ; Adult ; Male ; Female ; Middle Aged ; Pain Measurement ; *Pain Management/methods ; },
abstract = {BACKGROUND: Neuropathic pain (NP) is characterized as pain arising from lesions of the somatosensory nervous system. However, NP-like features have been found in several chronic secondary musculoskeletal (MSK) pain conditions in the absence of detectable lesion or damage to the somatosensory pathways. Emerging evidence has demonstrated associations between NP-like symptoms and altered neural activity within brain regions implicated in sensory perception and affective-emotional processing of pain with consistent findings of abnormal activity in the right insula (RIns) cortex and dorsal anterior cingulate cortex (dACC). Electroencephalography neurofeedback (EEG-NF) is a brain-computer interface biofeedback technique that allows individuals to self-regulate the real-time cortical brain activities of the regions of interest.

OBJECTIVE: The primary objective of this study is to investigate the feasibility and safety of a novel EEG-NF intervention designed to simultaneously downtrain activity in the RIns and dACC in individuals with a chronic secondary MSK pain condition exhibiting NP-like features. In addition, this study will conduct secondary exploratory analyses to investigate EEG-derived neuronal changes and their associations with clinical and experimental pain outcomes following the EEG-NF training.

METHODS: We will design a single-arm, open-label, pilot-feasibility trial. We will recruit adults aged 35-75 years with a score of ≥19 using the PainDETECT questionnaire and an average pain score of ≥4 on the 11-point Numeric Pain Rating Scale over the last 3 months, with a minimum pain duration of 3 months, to receive active EEG-NF training. Participants will receive auditory feedback as a reward for achieving a predetermined activity threshold of the RIns and dACC. Primary outcomes will evaluate feasibility, acceptability, and safety using both self-reported questionnaires and monitoring data. Collected data will be summarized descriptively, with mean (SD) reported where appropriate. Secondary outcomes will include EEG parameters, self-reported measures, heart rate variability, and quantitative sensory testing. An exploratory within-group pre-post statistical comparison will be conducted for all secondary outcome measures, and correlation analysis will be performed to explore relationships between EEG measures, self-reported outcomes, heart rate variability, and quantitative sensory testing measures.

RESULTS: This study has received approval from the Health and Disability Ethics Committee and is registered with the Australian New Zealand Clinical Trials Registry. Participant recruitment began in April 2025 and is ongoing. As of October 2025, data collection has been completed, with a total of 5 participants enrolled, all of whom have completed the study to date. We expect to complete the study in March 2026. This study will generate data on feasibility, safety, acceptability, and preliminary data to inform a fully powered effectiveness clinical trial.

CONCLUSIONS: The results and data generated will inform the design and sample size calculation for a fully powered randomized controlled trial aimed at evaluating the effectiveness of EEG-NF in targeting NP-like features in individuals with chronic MSK pain.

TRIAL REGISTRATION: Australian New Zealand Clinical Trials Registry ACTRN12625000706471; https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=389568&isReview=true.

DERR1-10.2196/78806.},
}

Humans
*Neurofeedback/methods
Pilot Projects
*Neuralgia/therapy/physiopathology
Feasibility Studies
*Musculoskeletal Pain/therapy/physiopathology
Electroencephalography/methods
*Chronic Pain/therapy
Adult
Male
Female
Middle Aged
Pain Measurement
*Pain Management/methods

@article {pmid41183389,
year = {2025},
author = {Yang, T and Cai, S and Xu, D and Hu, N},
title = {End-to-End EEG Artifact Removal Method via Nested Generative Adversarial Network.},
journal = {Biomedical physics & engineering express},
volume = {},
number = {},
pages = {},
doi = {10.1088/2057-1976/ae1a8c},
pmid = {41183389},
issn = {2057-1976},
abstract = {As physiological artifacts commonly overlap with EEG signals in both time and frequency domains, developing an effective end-to-end EEG artifact removal method is essential for a brain-computer interface (BCI) system. Approach. An end-to-end artifact removal method based on nested generative adversarial network (GAN) is proposed, to recover the EEG signals from artifact-contaminated ones. The nested GAN consists of two components: an inner GAN operating in time-frequency domain and an outer GAN functioning in time domain. A light-weighted complex-valued restormer, designed in time-frequency domain, is employed as the generator to reconstruct the denoised EEG signal. Two metric discriminators in the inner GAN and two multi-resolution discriminators in the outer GAN are used, and gradient balance is used to address the partial learning issue during training. Main results. The performance of the nested GAN has been evaluated in the realistic EEG dataset and semi-synthetic dataset. Compared to the benchmark methods, the proposed one achieved best average performance evaluation metrics, including mean square error (MSE) = 0.098, Pearson correlation coefficient (PCC) = 0.892, relative root MSE (RRMSE) = 0.065, the percentage reduction of time domain artifacts () = 71.6%, and the percentage reduction of frequency domain artifacts () = 76.9%. The performance of artifact removal also showed robustness across a wide range of signal-to-noise ratio (SNR) levels. Significance. The superior performance of the proposed end-to-end artifact removal method is expected to contribute to the advancement of BCI system development. .},
}

@article {pmid41183383,
year = {2025},
author = {Sharafkhani, N and Zhang, H},
title = {Deployable electrode arrays for brain interfaces: structural reconfiguration strategies for long-term stability and high-fidelity recording - a review.},
journal = {Journal of neural engineering},
volume = {},
number = {},
pages = {},
doi = {10.1088/1741-2552/ae1ab3},
pmid = {41183383},
issn = {1741-2552},
abstract = {Neural electrodes, as essential tools for recording and stimulating neural tissues, significantly impact therapeutic strategies for neurological disorders through deep brain stimulation, responsive neurostimulation, and brain-computer interfaces. Despite considerable advancements, the efficiency and longevity of neural electrodes are compromised by brain micromotion, induced by physiological activities such as cardiac pulsation and respiration. The mechanical mismatch between rigid electrodes and soft neural tissue generates persistent stresses at the electrode-tissue interface, triggering tissue damage, inflammatory responses, encapsulation, and ultimately electrode failure. Deployable neural electrodes, characterized by structural reconfiguration after implantation, have emerged to address these challenges. Deployment mechanisms, including unfolding, expanding, unrolling, or ejecting electrode arms from an initially compact configuration, reduce insertion trauma, maximize spatial coverage, and mitigate brain micromotion effects, thereby enhancing long-term stability and recording fidelity. Approach. This review provides the first comprehensive analysis of deployable intracortical and electrocorticography electrodes, emphasizing their design principles, deployment mechanisms, mechanical performance, advantages, and limitations. This review fills a critical gap in the existing neural electrode literature by transitioning the focus from traditional geometric and material considerations to advanced structural reconfiguration strategies. Significance. An understanding of the advantages and disadvantages of these deployment strategies provides essential insights and future directions for optimizing neural electrode technologies. .},
}

@article {pmid41182766,
year = {2025},
author = {O'Regan, RM and Ren, Y and Zhang, Y and Siuliukina, N and Schnabel, CA and Kammler, R and Viale, G and Dell'Orto, P and Munzone, E and Láng, I and Tondini, C and Gomez, HL and Chini, C and Nicoletti, SVL and Puglisi, F and Zaman, K and Goetz, MP and Stearns, V and Martino, S and Salim, M and Loibl, S and Geyer, CE and Bonnefoi, HR and Ciruelos, EM and Loi, S and Colleoni, M and Fleming, GF and Francis, PA and Walley, BA and Pagani, O and Treuner, K and Regan, MM},
title = {Assessment of Adjuvant Endocrine Therapy With Ovarian Function Suppression by Breast Cancer Index.},
journal = {JAMA network open},
volume = {8},
number = {11},
pages = {e2540931},
doi = {10.1001/jamanetworkopen.2025.40931},
pmid = {41182766},
issn = {2574-3805},
mesh = {Humans ; Female ; *Breast Neoplasms/drug therapy ; *Tamoxifen/therapeutic use ; Chemotherapy, Adjuvant/methods ; Adult ; Middle Aged ; Androstadienes/therapeutic use ; Prospective Studies ; *Antineoplastic Agents, Hormonal/therapeutic use ; Prognosis ; Retrospective Studies ; *Ovary/drug effects ; Premenopause ; Biomarkers, Tumor ; Homeodomain Proteins ; Receptors, Interleukin-17 ; },
abstract = {IMPORTANCE: The Breast Cancer Index (BCI) previously identified premenopausal patients with tumors in which the ratio of expression of HOXB13 relative to IL17BR (hereafter, BCI [H/I]-low tumors) as likely to derive greatest benefit from ovarian function suppression (OFS)-containing adjuvant therapy in the Suppression of Ovarian Function Trial (SOFT) trial.

OBJECTIVES: To assess BCI as a predictive biomarker of benefit from exemestane plus OFS vs tamoxifen plus OFS and to validate BCI as a prognostic biomarker for premenopausal patients.

This prognostic study used a prospective-retrospective translational design within the Tamoxifen and Exemestane (TEXT) and SOFT trials (enrolled November 2003 to April 2011). Blinded BCI testing in all available tumor samples was completed in March 2024. Premenopausal women with hormone receptor-positive breast cancer randomized to tamoxifen plus OFS or exemestane plus OFS who had BCI assessed were included. Analysis occurred from March to August 2024.

EXPOSURE: 5 years of adjuvant tamoxifen plus OFS or exemestane plus OFS.

MAIN OUTCOMES AND MEASURES: The primary outcomes were breast cancer-free interval (BCFI) for predictive analyses and distant recurrence-free interval (DRFI) for prognostic analyses after a median follow-up of 13 years in the TEXT cohort. Secondary objectives examined the predictive performance of BCI (H/I) in the combined TEXT and SOFT cohort overall and in prespecified clinical subgroups.

RESULTS: Of 1782 patients in the TEXT study, 1034 (58.0%) had BCI (H/I)-low tumors; 915 (51.3%) of patients had N0 disease and 1077 (60.4%) were younger than 45 years. Patients with BCI (H/I)-low tumors had a 6.6% absolute benefit in 12-year BCFI (HR, 0.61; 95% CI, 0.44-0.85) for exemestane plus OFS vs tamoxifen plus OFS, while those with BCI (H/I)-high tumors had a 6.3% absolute benefit (HR, 0.78; 95% CI, 0.57-1.07; P for interaction = .29). Results were consistent in the combined TEXT plus SOFT cohort (2896 patients) and adjusting for clinicopathological variables. Clinical subgroup analyses consistently showed benefit of exemestane plus OFS vs tamoxifen plus OFS for BCI (H/I)-low tumors, and more variable relative treatment effects among BCI (H/I)-high tumors, including by age. Post hoc exploratory time-varying estimates suggested the treatment × BCI associations may differ in years 0 to 5 vs greater than 5 years. BCI and BCI N+ as continuous indices were prognostic for distant recurrence in N0 (HR, 1.27; 95% CI, 1.11-1.44; P < .001) and N1 (HR, 1.58; 95% CI, 1.21-2.05; P < .001) cancers. The 12-year DRFI was 96.3%, 90.3%, and 84.9% for BCI low-, intermediate-, and high-risk N0 cancers, respectively.

CONCLUSIONS AND RELEVANCE: In this study of premenopausal women with hormone receptor-positive breast cancer, BCI (H/I) status did not clearly predict greater benefit of adjuvant exemestane plus OFS vs tamoxifen plus OFS for women with BCI (H/I)-low tumors than for those with BCI (H/I)-high tumors; BCI continuous indices were reconfirmed as prognostic for premenopausal women. These findings support prior results of SOFT, which compared tamoxifen-alone vs OFS with either exemestane or tamoxifen, indicating premenopausal patients with BCI (H/I)-low tumors may benefit from more intensive endocrine therapy.},
}

Humans
Female
*Breast Neoplasms/drug therapy
*Tamoxifen/therapeutic use
Chemotherapy, Adjuvant/methods
Adult
Middle Aged
Androstadienes/therapeutic use
Prospective Studies
*Antineoplastic Agents, Hormonal/therapeutic use
Prognosis
Retrospective Studies
*Ovary/drug effects
Premenopause
Biomarkers, Tumor
Homeodomain Proteins
Receptors, Interleukin-17

@article {pmid41134945,
year = {2025},
author = {Kim, DS and Lee, SH and Yin, K and Lee, SW},
title = {Reconstructing Unseen Sentences From Speech-Related Biosignals for Open-Vocabulary Neural Communication.},
journal = {IEEE transactions on neural systems and rehabilitation engineering : a publication of the IEEE Engineering in Medicine and Biology Society},
volume = {33},
number = {},
pages = {4338-4348},
doi = {10.1109/TNSRE.2025.3625219},
pmid = {41134945},
issn = {1558-0210},
mesh = {Humans ; Electroencephalography ; *Speech/physiology ; Male ; Female ; Adult ; Electromyography ; Young Adult ; Algorithms ; Brain-Computer Interfaces ; Phonetics ; Communication ; Brain/physiology ; Signal Processing, Computer-Assisted ; Speech Perception/physiology ; },
abstract = {Brain-to-speech (BTS) systems represent a groundbreaking approach to human communication by enabling the direct transformation of neural activity into linguistic expressions. While recent non-invasive BTS studies have largely focused on decoding predefined words or sentences, achieving open-vocabulary neural communication comparable to natural human interaction requires decoding unconstrained speech. Additionally, effectively integrating diverse signals derived from speech is crucial for developing personalized and adaptive neural communication and rehabilitation solutions for patients. This study investigates the potential of speech synthesis for previously unseen sentences across various speech modes by leveraging phoneme-level information extracted from high-density electroencephalography (EEG) signals, both independently and in conjunction with electromyography (EMG) signals. Furthermore, we examine the properties affecting phoneme decoding accuracy during sentence reconstruction and offer neurophysiological insights to further enhance EEG decoding for more effective neural communication solutions. Our findings underscore the feasibility of biosignal-based sentence-level speech synthesis for reconstructing unseen sentences, highlighting a significant step toward developing open-vocabulary neural communication systems adapted to diverse patient needs and conditions. Additionally, this study provides meaningful insights into the development of communication and rehabilitation solutions utilizing EEG-based decoding technologies.},
}

Humans
Electroencephalography
*Speech/physiology
Male
Female
Adult
Electromyography
Young Adult
Algorithms
Brain-Computer Interfaces
Phonetics
Communication
Brain/physiology
Signal Processing, Computer-Assisted
Speech Perception/physiology

@article {pmid41182326,
year = {2025},
author = {Yuan, Z and Chen, F and Huang, X and Huang, K and Song, Z and Ding, Y and Gong, Z and Gu, G},
title = {Soft Tubular-Surface Rolling Robots.},
journal = {Soft robotics},
volume = {},
number = {},
pages = {},
doi = {10.1177/21695172251387190},
pmid = {41182326},
issn = {2169-5180},
abstract = {Soft creatures like Drosophila larvae can quickly ascend tubular surfaces via rolling, a capability not yet replicated by soft robots. Here, we present a single-piece soft robot capable of rolling along tubular structures by sequentially actuating its built-in axial muscles. We reveal that the sequential actuation generates distributed spinning torques along the robot's curved axis, enabling continuous non-coaxial rolling-distinct from current gravity-dependent rolling solutions. This non-coaxial rolling mechanism allows the robot to swiftly navigate tubular surfaces while conforming to their shapes and maintaining a stable grip. The robot's deformation and gripping force are actively adjusted to enhance its adaptability to various surfaces. We demonstrate that our robot can ascend pipes with varying geometries (e.g., varying-diameter, spiral-shaped, or non-cylindrical), traverse diverse terrains, pass through confined tunnels, and transition smoothly between planar rolling and pipe climbing. The robot's great adaptability and rapid movement underscore its potential for navigating scenarios with intricate surface geometries.},
}

@article {pmid41180699,
year = {2025},
author = {Ali, E and Kamran, S and Cheema, AAA},
title = {Brain-computer interfaces in post-stroke rehabilitation: a neurotechnological leap toward functional recovery.},
journal = {Annals of medicine and surgery (2012)},
volume = {87},
number = {11},
pages = {7784-7785},
pmid = {41180699},
issn = {2049-0801},
}

@article {pmid41180117,
year = {2025},
author = {Hall, R and Jackson, M and Maleki, M and Crogman, HT},
title = {Modeling cognition through adaptive neural synchronization: a multimodal framework using EEG, fMRI, and reinforcement learning.},
journal = {Frontiers in computational neuroscience},
volume = {19},
number = {},
pages = {1616472},
pmid = {41180117},
issn = {1662-5188},
abstract = {INTRODUCTION: Understanding the cognitive process of thinking as a neural phenomenon remains a central challenge in neuroscience and computational modeling. This study addresses this challenge by presenting a biologically grounded framework that simulates adaptive decision making across cognitive states.

METHODS: The model integrates neuronal synchronization, metabolic energy consumption, and reinforcement learning. Neural synchronization is simulated using Kuramoto oscillators, while energy dynamics are constrained by multimodal activity profiles. Reinforcement learning agents-Q-learning and Deep Q-Network (DQN)-modulate external inputs to maintain optimal synchrony with minimal energy cost. The model is validated using real EEG and fMRI data, comparing simulated and empirical outputs across spectral power, phase synchrony, and BOLD activity.

RESULTS: The DQN agent achieved rapid convergence, stabilizing cumulative rewards within 200 episodes and reducing mean synchronization error by over 40%, outperforming Q-learning in speed and generalization. The model successfully reproduced canonical brain states-focused attention, multitasking, and rest. Simulated EEG showed dominant alpha-band power (3.2 × 10[-4] a.u.), while real EEG exhibited beta-dominance (3.2 × 10[-4] a.u.), indicating accurate modeling of resting states and tunability for active tasks. Phase Locking Value (PLV) ranged from 0.9806 to 0.9926, with the focused condition yielding the lowest circular variance (0.0456) and a near significant phase shift compared to rest (t = -2.15, p = 0.075). Cross-modal validation revealed moderate correlation between simulated and real BOLD signals (r = 0.30, resting condition), with delayed inputs improving temporal alignment. General Linear Model (GLM) analysis of simulated BOLD data showed high region-specific prediction accuracy (R [2] = 0.973-0.993, p < 0.001), particularly in prefrontal, parietal, and anterior cingulate cortices. Voxel-wise correlation and ICA decomposition confirmed structured network dynamics.

DISCUSSION: These findings demonstrate that the framework captures both electrophysiological and spatial aspects of brain activity, respects neuroenergetic constraints, and adaptively regulates brain-like states through reinforcement learning. The model offers a scalable platform for simulating cognition and developing biologically inspired neuroadaptive systems.

CONCLUSION: This work provides a novel and testable approach to modeling thinking as a biologically constrained control problem and lays the groundwork for future applications in cognitive modeling and brain-computer interfaces.},
}

@article {pmid41179991,
year = {2025},
author = {Yuan, L and Wei, J and Liu, Y},
title = {Spiking neural networks for EEG signal analysis using wavelet transform.},
journal = {Frontiers in neuroscience},
volume = {19},
number = {},
pages = {1652274},
pmid = {41179991},
issn = {1662-4548},
abstract = {INTRODUCTION: Brain-computer interfaces (BCIs) leverage EEG signal processing to enable human-machine communication and have broad application potential. However, existing deep learning-based BCI methods face two critical limitations that hinder their practical deployment: reliance on manual EEG feature extraction, which constrains their ability to adaptively capture complex neural patterns, and high energy consumption characteristics that make them unsuitable for resource-constrained portable BCI devices requiring edge deployment.

METHODS: To address these limitations, this work combines wavelet transform for automatic feature extraction with spiking neural networks for energy-efficient computation. Specifically, we present a novel spiking transformer that integrates a spiking self-attention mechanism with discrete wavelet transform, termed SpikeWavformer. SpikeWavformer enables automatic EEG signal time-frequency decomposition, eliminates manual feature extraction, and provides energy-efficient classification decision-making, thereby enhancing the model's cross-scene generalization while meeting the constraints of portable BCI applications.

RESULTS: Experimental results demonstrate the effectiveness and efficiency of SpikeWavformer in emotion recognition and auditory attention decoding tasks.

DISCUSSION: These findings indicate that SpikeWavformer can address the key limitations of existing BCI methods and holds promise for practical deployment in portable, resource-constrained scenarios.},
}

@article {pmid41179694,
year = {2025},
author = {Fernández-Rodríguez, Á and Velasco-Álvarez, F and Vizcaíno-Martín, FJ and Ron-Angevin, R},
title = {Impact of stimulus presentation speed in a visual ERP-based BCI under RSVP.},
journal = {Cognitive neurodynamics},
volume = {19},
number = {1},
pages = {171},
pmid = {41179694},
issn = {1871-4080},
abstract = {Rapid serial visual presentation (RSVP) is one of the most effective gaze-independent paradigms for event-related potential (ERP)-based brain-computer interfaces (BCIs), particularly for individuals with limited muscle and eye movement control. The speed of visual stimulus presentation is a critical factor influencing system performance and warrants thorough investigation. This study evaluates the impact of different stimulus presentation speeds on the performance of an ERP-BCI used for pictogram selection under RSVP. Thirteen participants tested the ERP-BCI across three experimental conditions, each with a different stimulus onset asynchrony (SOA): 80 ms (C080), 160 ms (C160), and 320 ms (C320). In addition to performance metrics such as accuracy, information transfer rate (ITR), and pictograms per minute (PPM), a subjective evaluation of the user experience was conducted for each condition. The results indicate that C160 outperformed both C080 and C320 across all performance metrics, achieving an ITR of 26.49 bit/min (81.28% accuracy in 4.8 s). Subjective evaluations also revealed a preference for C160 and C320 over C080. Therefore, among the SOAs evaluated, 160 ms appears to be the most suitable for enhancing system usability. These findings underscore the crucial role of stimulus presentation speed in the usability of ERP-BCIs for pictogram selection under RSVP, emphasizing its importance in future gaze-independent ERP-BCI designs for communication purposes.},
}

@article {pmid41178032,
year = {2025},
author = {Li, Q and Choi, EPH and Gou, M and Tian, Y and Baptiste, D},
title = {Brain-Computer Interface: Bring Care Into a Future Phase? Challenges and Opportunities for Nursing in the Era of Emerging Technologies.},
journal = {Nursing open},
volume = {12},
number = {11},
pages = {e70345},
doi = {10.1002/nop2.70345},
pmid = {41178032},
issn = {2054-1058},
}

@article {pmid41177816,
year = {2025},
author = {Fu, R and Liu, Y and Wang, Z and Liang, Z},
title = {Virtual Reality (VR) Paradigm-Agnostic Motor Imagery Decoding Using Lightweight Network With Adaptive Attention Mechanism.},
journal = {Journal of medical systems},
volume = {49},
number = {1},
pages = {152},
pmid = {41177816},
issn = {1573-689X},
support = {62073282//National Natural Science Foundation of China/ ; F2022203092//Natural Science Foundation of Hebei Province/ ; 202302B015//the S&T Program of Qinhuangdao City/ ; KFKT2025B88//Project of State Key Laboratory for Novel Software Technology/ ; },
mesh = {Humans ; *Brain-Computer Interfaces ; *Virtual Reality ; Algorithms ; *Imagination/physiology ; Electroencephalography ; Movement/physiology ; Attention ; Male ; Adult ; },
abstract = {Motor imagery (MI) is widely used in brain-computer interfaces (BCIs) due to its simplicity and reproducibility, enabling individuals with motor impairments to perform non-muscular limb training for the rehabilitation of motor-related neurons. While MI-based BCIs have shown promise for neurorehabilitation, current 2D paradigms fail to engage critical sensorimotor networks. To address this limitation, we designed an immersive MI paradigm in a virtual reality (VR) environment, where participants imagined limb movements in response to continuous three-dimensional (3D) palm motion stimuli. Furthermore, we proposed a novel decoding algorithm that integrates depthwise separable convolution with multi-head self-attention mechanisms. The proposed method was evaluated against existing approaches, demonstrating superior classification accuracy while reducing the temporal and spatial complexity associated with attention mechanisms. To assess the generalizability and robustness of the algorithm across different scenarios, we conducted experiments on two publicly available datasets: BCI Competition IV-2a and the PhysioNet MI dataset. Results showed that our method achieved an average increase of nearly 8% in kappa score over EEGNet in decoding four-class MI tasks in 2D paradigms. Consistent performance across both VR and 2D paradigms confirmed the algorithm's effectiveness and applicability in multi-scenario MI decoding. This study introduces a novel immersive MI paradigm and decoding framework, offering a promising approach for enhancing user engagement in neurorehabilitation and advancing EEG-based intention recognition in VR environments.},
}

Humans
*Brain-Computer Interfaces
*Virtual Reality
Algorithms
*Imagination/physiology
Electroencephalography
Movement/physiology
Attention
Male
Adult

@article {pmid41177674,
year = {2025},
author = {Almufareh, MF and Kausar, S and Humayun, M and Tehsin, S and Farooq, A},
title = {Inner Speech Decoding: A Comprehensive Review.},
journal = {Wiley interdisciplinary reviews. Cognitive science},
volume = {16},
number = {6},
pages = {e70016},
doi = {10.1002/wcs.70016},
pmid = {41177674},
issn = {1939-5086},
support = {KSRG-2024-063//King Salman Center for Disability Research/ ; },
mesh = {Humans ; *Speech/physiology ; Electroencephalography ; Machine Learning ; Magnetic Resonance Imaging ; *Brain/physiology ; *Speech Perception/physiology ; Brain-Computer Interfaces ; },
abstract = {Inner speech decoding is the process of identifying silently generated speech from neural signals. In recent years, this candidate technology has gained momentum as a possible way to support communication in severely impaired populations. Specifically, this approach promises hope for people with a variety of physical or neurological disabilities who need alternative means of verbal expression. This review covers recording modalities that range from the noninvasive EEG to the high-density electrocorticography and discusses how linear discriminant analysis, deep convolutional networks, and hybrid fusion of EEG with fMRI are integrated into machine learning strategies to infer covert speech. This review synthesizes evidence to suggest that small vocabularies, under controlled conditions, can yield relatively reasonable accuracy while further refining the decoding outcome via context-based approaches. The impact of sensor quality, training data size, and domain adaptation is illustrated by focusing on public datasets of imagined or articulated speech. Throughout the article, the methodological standards emerging across laboratories will be discussed, emphasizing that effective inner speech recognition involves high-quality preprocessing, subject calibration, and informed modeling choices balanced against computational power for interpretability. In addition to technical advancements, this review also examines the ethical, societal, and regulatory challenges surrounding inner speech decoding, including brain data privacy, neural rights, informed consent, and user trust. Addressing these interdisciplinary issues is critical for the responsible development and real-world adoption of such technologies. This article is categorized under: Neuroscience > Computation Computer Science and Robotics > Machine Learning.},
}

Humans
*Speech/physiology
Electroencephalography
Machine Learning
Magnetic Resonance Imaging
*Brain/physiology
*Speech Perception/physiology
Brain-Computer Interfaces

@article {pmid41177306,
year = {2025},
author = {Zhong, X and Li, G and Xu, C and Luo, R and Meng, J and Schalk, G},
title = {Detection of eye movements and eye blinks using a portable two-channel EEG platform.},
journal = {Journal of neuroscience methods},
volume = {},
number = {},
pages = {110616},
doi = {10.1016/j.jneumeth.2025.110616},
pmid = {41177306},
issn = {1872-678X},
abstract = {BACKGROUND: The ability to detect eye movements can facilitate human-computer interaction (HCI) and may complement brain-computer interfaces (BCIs). Recent studies have shown that multi-channel EEG systems can provide information about eye movements, but these systems can be bulky and/or require complex setup.

NEW METHOD: We introduce a portable, two-channel EEG platform that can be placed in seconds and detect eye blinks/movements and gaze trajectories. Forty adults performed cued blinks and horizontal/vertical gaze shifts; 21 EEG features were extracted, and machine learning models were evaluated with leave-one-subject-out validation.

RESULTS: Our system effectively identified eye blinks (avg. detection accuracy of 95%, 50% chance) and horizontal eye movements (avg. accuracy of 94%, 33% chance), and showed decreased performance detecting vertical eye movements (avg. accuracy of 60%, 33% chance). It was also able to predict horizontal and vertical eye movement trajectories (r=0.79 and r=0.14, respectively).

Classification accuracies for eye blinks and horizontal eye movements using our system with only two electrodes are comparable to those previously reported only for complex multi-channel EEG/EOG setups.

CONCLUSION: This study provides evidence, for the first time, that a wearable EEG device can give substantial information about eye blinks and eye movements. With further refinements, this approach may enable portable solutions for real-world HCI and BCI applications.},
}

@article {pmid41176895,
year = {2025},
author = {Cavallé Garrido, L and de Paula Vernetta, C and Guzmán Calvete, A and Álvarez Arocas, J and Gonçalves, C and Armengot Carceller, M},
title = {Bonebridge active transcutaneous bone conduction hearing implant: Results in the pediatric population.},
journal = {International journal of pediatric otorhinolaryngology},
volume = {199},
number = {},
pages = {112610},
doi = {10.1016/j.ijporl.2025.112610},
pmid = {41176895},
issn = {1872-8464},
abstract = {PURPOSE: This study provides prospective and retrospective data on safety and performance results with the Bonebridge BCI 602 (MED-EL) active transcutaneous bone conduction implant in children.

METHODS: Audiological data were collected at 3 intervals (preoperative, initial activation and 3 months postoperative). Quality of life was assessed with the Speech, Spatial, and Qualities of Hearing (SSQ12/P), KID KINDL and Audio Processor Satisfaction Questionnaire (APSQ) as well as a postoperative questionnaire specifically designed for this study.

RESULTS: 22 pediatric patients (20 conductive/mixed hearing loss (CHL/MHL) and 2 single-sided deafness (SSD)) aged 4-17 received a BCI 602. Three-month post-op pure-tone average (PTA4) functional gain (FG) was 31.9 dB HL for the CHL/MHL group and 11.3 dB HL in the SSD patients. CHL/MHL patients had a mean word recognition score (WRS) improvement of 80.6 ± 23.9 % at initial activation and 83 ± 20.3 % at 3 months post-op. Speech recognition in noise at +5 dB SNR in the CHL/MHL group improved from 24.6 ± 28.3 % unaided to 74.9 ± 26 % aided at 3 months post-op. The mean post-op total scores were 5.5 ± 1.8 on the SSQ12/P and 8.87 ± 0.93 on the APSQ questionnaires. No major complications were noted on the postoperative questionnaire; minor complications were resolved by the end of the study. Stable bone and air conduction thresholds confirmed device safety.

CONCLUSION: The Bonebridge BCI 602 is safe and effective for use in the pediatric population.},
}

@article {pmid41174212,
year = {2025},
author = {Zhou, H and Iramina, K},
title = {Discovery of EEG effective connectivity during visual motor imagery with multi-scale symbolic transfer entropy.},
journal = {Scientific reports},
volume = {15},
number = {1},
pages = {38200},
pmid = {41174212},
issn = {2045-2322},
mesh = {Humans ; *Electroencephalography/methods ; Male ; Female ; Adult ; *Imagination/physiology ; Brain-Computer Interfaces ; Entropy ; Young Adult ; Parietal Lobe/physiology ; *Brain/physiology ; Brain Mapping ; Occipital Lobe/physiology ; },
abstract = {Visual motor imagery (VMI) is an important component of motor imagery, with potential applications in brain-computer interfaces and motor rehabilitation due to its lower training cost compared to kinesthetic motor imagery (KMI). However, the neural mechanisms underlying VMI, particularly the effects of imagery hand and imagery perspective (first-person perspective, 1pp, vs. third-person perspective, 3pp) remain unclear. This study examines the effective connectivity of VMI EEG using multi-scale symbolic transfer entropy. Time-frequency analysis revealed prominent event-related synchronization (ERS) in the alpha and high-beta bands, while connectivity analysis emphasized strong information flow within the parieto-occipital network. Notably, hand effect dominant information flows were found between the motor and posterior parietal-occipital regions, while perspective suggested a more remarkable effect. 1pp imagery significantly enhanced top-down modulation of the occipital cortex, whereas 3pp imagery engaged the right posterior parietal region, suggesting stronger spatial localization processing. These findings provide novel insights into the distinct neural mechanisms of VMI and its potential applications in cognitive neuroscience and brain-machine engineering.},
}

Humans
*Electroencephalography/methods
Male
Female
Adult
*Imagination/physiology
Brain-Computer Interfaces
Entropy
Young Adult
Parietal Lobe/physiology
*Brain/physiology
Brain Mapping
Occipital Lobe/physiology

@article {pmid41173359,
year = {2025},
author = {Chen, Z and Lu, Y and Xu, X},
title = {EEG-SGENet: A lightweight convolutional network integrating SGE for motor imagery brain-computer interfaces.},
journal = {Neuroscience},
volume = {},
number = {},
pages = {},
doi = {10.1016/j.neuroscience.2025.09.040},
pmid = {41173359},
issn = {1873-7544},
abstract = {In recent years, there has been a significant increase in research activity on electroencephalography (EEG)-based motor imagery brain-computer interfaces (MI-BCI) in the field of deep learning. However, despite achieving high accuracy, the size of models is increasing, requiring significant memory and computational resources. Therefore, finding a balance between accuracy and computational cost has always been a challenge in MI classification research. Convolutional Neural Networks (CNNs) generate feature representations of objects by collecting semantic sub-features. The activation of subfeatures is susceptible to noisy backgrounds. The Spatial Group-wise Enhance (SGE) module adjusts the importance of each sub-feature by generating an attention factor for the spatial location of each semantic group, thus enhancing useful features and suppressing noise. The design of the SGE module is lightweight, with few parameters and computations. Therefore, we introduce the SGE module to improve accuracy and minimize model parameters. In this paper, we propose EEG-SGENet, a novel end-to-end convolutional neural network model that considers both the lightweight model and accuracy. Experimental results on the BCI IV 2a dataset show that EEG-SGENet achieves an accuracy of 80.98% in the four categories of MI. The average classification accuracy for the two-category task of BCI IV 2b is 76.17%. Comparisons with other lightweight models in terms of classification accuracy and other aspects have shown that this model achieves a good balance between decoding performance and computational cost. Overall, experimental results demonstrate that the proposed model is expected to become a new method for decoding EEG signals.},
}

@article {pmid41171945,
year = {2025},
author = {Khanam, T and Siuly, S and Ahmad, K and Wang, H},
title = {A novel channel reduction concept to enhance the classification of motor imagery tasks in brain-computer interface systems.},
journal = {PloS one},
volume = {20},
number = {10},
pages = {e0335511},
pmid = {41171945},
issn = {1932-6203},
mesh = {*Brain-Computer Interfaces ; Humans ; *Electroencephalography/methods ; Algorithms ; Neural Networks, Computer ; Male ; Adult ; *Imagination/physiology ; Female ; Movement/physiology ; Brain/physiology ; Signal Processing, Computer-Assisted ; },
abstract = {Electroencephalogram (EEG) signals play a critical role in advancing brain-computer interface (BCI) systems, particularly for detecting motor imagery (MI) movements. However, analysing large volume of EEG datasets faces some challenges due to redundant information, and performance degradation. Irrelevant channels introduce noise, which reduces accuracy and slows system performance. To address these issues, this study aims to develop a novel channel selection method to enhance EEG-based MI task performance in BCI applications. Our proposed hybrid approach combines statistical t-tests with a Bonferroni correction-based channel reduction technique, followed by the application of a Deep Learning Regularized Common Spatial Pattern with Neural Network (DLRCSPNN) framework. This framework employs DLRCSP for feature extraction and neural network (NN) algorithm for classification. Our developed method excluded channels with correlation coefficients below 0.5, retaining only significant, non-redundant channels and tested on three real-time EEG-based BCI datasets. This study produces the highest accuracy score in the case of every subjects above 90% for all the applied datasets. In the first dataset, our method achieved the highest accuracy, improving by 3.27% to 42.53% in terms of individual subject compared to seven existing machine learning algorithms. In the second and third dataset, it outperformed existing approaches, with accuracy gains of 5% to 45% and 1% to 17.47% respectively. Comparisons with a CSP and NN framework confirmed DLRCSPNN's algorithms superior performance. These results demonstrate the effectiveness of the approach, offering a new perspective on the identification of MI task performance in EEG based BCI technology. This proposed technique will enable rapid identification of motor-disabled individuals' intentions, supporting patient rehabilitation and improving daily living.},
}

*Brain-Computer Interfaces
Humans
*Electroencephalography/methods
Algorithms
Neural Networks, Computer
Male
Adult
*Imagination/physiology
Female
Movement/physiology
Brain/physiology
Signal Processing, Computer-Assisted

@article {pmid41171651,
year = {2025},
author = {Liu, H and Wang, Z and Li, R and Zhao, X and Xu, T and Zhou, T and Hu, H},
title = {A Novel Binocular-Encoded SSVEP Framework for Efficient VR-Based Brain-Computer Interface.},
journal = {IEEE journal of biomedical and health informatics},
volume = {PP},
number = {},
pages = {},
doi = {10.1109/JBHI.2025.3626332},
pmid = {41171651},
issn = {2168-2208},
abstract = {This paper presents a novel binocular-encoded SSVEP (beSSVEP) method, leveraging binocular vision in virtual reality (VR) to enhance brain-computer interface (BCI) applications. We introduce the Binocular Periodically Repeated Component Analysis (bPRCA) algorithm, designed to address the unique characteristics of binocular-encoded targets, which include combinations of monocular single-frequency SSVEP units or void units, with frequency units being reused multiple times in the encoded interface. To further optimize performance, we propose the Fusion Component Analysis (FusionCA) framework, which integrates bPRCA with Task-related Component Analysis (TRCA), effectively utilizing both steady-state periodic components and cross-trial aperiodic components. Experimental results demonstrate that ensemble-FusionCA achieves the highest information transfer rate (ITR) with an average accuracy of $71.39\%$ and an ITR of 138.50 bits/min at 0.4 seconds, among the comparison with ensemble-bPRCA and ensemble-TRCA. Compared to traditional SSVEP approaches, beSSVEP significantly enhances frequency utilization, making VR-BCI systems more efficient and practical. This study highlights the application of physiological mechanisms of binocular vision to improve BCI systems, offering a new perspective for developing fast and scalable brain-computer interactions in VR environments.},
}

@article {pmid41170547,
year = {2025},
author = {Ren, J and Mo, WY and Wang, L and Ni, GJ and Yang, JJ},
title = {[Research progress on the role of dopamine system in regulating hippocampal related brain functions].},
journal = {Sheng li xue bao : [Acta physiologica Sinica]},
volume = {77},
number = {5},
pages = {893-904},
doi = {10.13294/j.aps.2025.0055},
pmid = {41170547},
issn = {0371-0874},
mesh = {*Hippocampus/physiology ; *Dopamine/physiology ; Humans ; Animals ; Receptors, Dopamine D2/physiology ; Memory/physiology ; Signal Transduction/physiology ; Neurodegenerative Diseases/physiopathology ; },
abstract = {Dopamine, as a catecholamine neurotransmitter widely distributed in the central nervous system, is involved in physiological functions such as motivation, arousal, reinforcement, and movement through various dopamine signaling pathways. The hippocampus receives dopaminergic neuron projections from regions such as the ventral tegmental area, locus coeruleus, and substantia nigra. Through D1-like and D2-like receptors, dopamine exerts significant regulatory effects such as spatial navigation, episodic memory, fear, anxiety, and reward. This review mainly summarizes the research progress on the functions of dopamine in the hippocampus from aspects including the sources of dopamine, receptor distribution and function, and the association of hippocampal dopamine system dysregulation with neurodegenerative diseases. The aim is to provide insights into the involvement of the dopamine system in hippocampal functions and the diagnosis and treatment of related diseases.},
}

*Hippocampus/physiology
*Dopamine/physiology
Humans
Animals
Receptors, Dopamine D2/physiology
Memory/physiology
Signal Transduction/physiology
Neurodegenerative Diseases/physiopathology

@article {pmid41170533,
year = {2025},
author = {Gherman, DE and Zander, TO},
title = {Towards neuroadaptive chatbots: a feasibility study.},
journal = {Frontiers in neuroergonomics},
volume = {6},
number = {},
pages = {1589734},
pmid = {41170533},
issn = {2673-6195},
abstract = {INTRODUCTION: Large-language models (LLMs) are transforming most industries today and are set to become a cornerstone of the human digital experience. While integrating explicit human feedback into the training and development of LLM-based chatbots has been integral to the progress we see nowadays, more work is needed to understand how to best align them with human values. Implicit human feedback enabled by passive brain-computer interfaces (pBCIs) could potentially help unlock the hidden nuance of users' cognitive and affective states during interaction with chatbots. This study proposes an investigation on the feasibility of using pBCIs to decode mental states in reaction to text stimuli, to lay the groundwork for neuroadaptive chatbots.

METHODS: Two paradigms were created to elicit moral judgment and error-processing with text stimuli. Electroencephalography (EEG) data was recorded with 64 gel electrodes while participants completed reading tasks. Mental state classifiers were obtained in an offline manner with a windowed-means approach and linear discriminant analysis (LDA) for full-component and brain-component data. The corresponding event-related potentials (ERPs) were visually inspected.

RESULTS: Moral salience was successfully decoded at a single-trial level, with an average calibration accuracy of 78% on the basis of a data window of 600 ms. Subsequent classifiers were not able to distinguish moral judgment congruence (i.e., moral agreement) and incongruence (i.e., moral disagreement). Error processing in reaction to factual inaccuracy was decoded with an average calibration accuracy of 66%. The identified ERPs for the investigated mental states partly aligned with other findings.

DISCUSSION: With this study, we demonstrate the feasibility of using pBCIs to distinguish mental states from readers' brain data at a single-trial level. More work is needed to transition from offline to online investigations and to understand if reliable pBCI classifiers can also be obtained in less controlled language tasks and more realistic chatbot interactions. Our work marks preliminary steps for understanding and making use of neural-based implicit human feedback for LLM alignment.},
}

@article {pmid41169537,
year = {2025},
author = {Li, X and Ji, X and Wang, Y and Chen, X},
title = {The influence of different visual eccentricity on SSVEPs elicited by ultra-low frequency visual stimulation in the lower peripheral visual field.},
journal = {Cognitive neurodynamics},
volume = {19},
number = {1},
pages = {170},
pmid = {41169537},
issn = {1871-4080},
abstract = {Steady-state visual evoked potential (SSVEP)-based brain-computer interfaces (BCIs) have been widely explored due to their high information transfer rate (ITR) and minimal training requirements. Traditional SSVEP-based BCIs typically use low- and medium-frequency visual stimuli from the central visual field to induce SSVEPs, but these can easily lead to visual fatigue. In order to improve system's comfort, some studies have attempted to use visual stimuli from the peripheral visual field to elicit SSVEPs. However, few studies have investigated the effects of different visual eccentricities on induced SSVEPs. In this study, we used ultra-low frequency (i.e., 2.00-3.32 Hz) visual stimulation in the lower peripheral visual field to induce SSVEPs. Furthermore, we further explored the effects of different visual eccentricities (i.e., 2.1°, 3.1°, and 4.1°) on induced SSVEPs. Experimental results obtained from twelve participants revealed that all three eccentricity conditions were capable of eliciting SSVEP responses. Moreover, SSVEP amplitude gradually decreased as eccentricity increased. These results provide new parametric references for optimizing the spatial layout of visual stimuli in peripheral SSVEP-based BCI systems.},
}

@article {pmid41167038,
year = {2025},
author = {Schippers, A and Berezutskaya, J and Vansteensel, MJ and Freudenburg, ZV and Crone, NE and Ramsey, NF},
title = {The effect of perceived auditory feedback on speech Brain-Computer Interface decoding performance.},
journal = {Clinical neurophysiology : official journal of the International Federation of Clinical Neurophysiology},
volume = {180},
number = {},
pages = {2111403},
doi = {10.1016/j.clinph.2025.2111403},
pmid = {41167038},
issn = {1872-8952},
abstract = {OBJECTIVE: Brain-Computer Interfaces (BCI) provide alternative means of communication for individuals with severe motor impairment. Implantable speech BCIs have shown great potential, particularly in individuals who could still produce some speech-related movements and/or sounds. As perception of auditory feedback is important for correct speech sound production in able-bodied people, it is conceivable that a complete absence of such feedback in individuals who lost all ability to produce audible speech affects BCI performance. The current study therefore set out to investigate to what extent perception of auditory feedback of self-produced speech contributes to speech decoding performance.

METHODS: In three able-bodied participants, patterns of 65-95 Hz power over sensorimotor cortex were compared between normal speech and speech in which auditory feedback was masked by noise. In addition, decoding accuracy was compared between feedback situations.

RESULTS & CONCLUSIONS: We found subtle differences in brain activity patterns associated with speech production between situations in which participants could versus could not perceive their produced speech. Importantly, absence of such auditory feedback led to lower speech decoding performance in all participants.

SIGNIFICANCE: These results underline the need to validate speech BCI efficacy with fully paralyzed individuals, as perceived feedback can influence the attainable speech decoding accuracy.},
}

@article {pmid41165717,
year = {2025},
author = {Lydiatt, WB},
title = {Mind, Machine, and Medicine-Challenges and Opportunities.},
journal = {JAMA otolaryngology-- head & neck surgery},
volume = {},
number = {},
pages = {},
doi = {10.1001/jamaoto.2025.4108},
pmid = {41165717},
issn = {2168-619X},
}

@article {pmid41163348,
year = {2025},
author = {Song, K and Liu, Y and Xu, P},
title = {Acute Effects of Portable Dry-EEG Neurofeedback on Classical Chinese Learning: A Three-Arm Repeated-Measures Study.},
journal = {Brain and behavior},
volume = {15},
number = {11},
pages = {e70977},
pmid = {41163348},
issn = {2162-3279},
support = {2025ZSD017//Shanghai Municipal Education Science Research Project "Special Program for Philosophy and Social Sciences Research in Shanghai Higher Education Institutions"/ ; },
mesh = {Humans ; *Neurofeedback/methods ; Male ; Female ; Young Adult ; *Electroencephalography/methods ; Adult ; *Learning/physiology ; Attention/physiology ; Cognition/physiology ; Comprehension/physiology ; Language ; Adolescent ; East Asian People ; },
abstract = {OBJECTIVE: Dry-electrode electroencephalography (dry-EEG) systems offer promising opportunities for real-time neurofeedback in naturalistic educational settings, yet their effectiveness in supporting complex language learning remains underexplored. This study investigated the acute effects of portable dry-EEG neurofeedback on students' cognitive performance and attentional states during classical Chinese learning, using a repeated-measures design to compare neurofeedback, sham feedback, and device control conditions.

METHODS: A total of 20 undergraduate participants completed three sessions involving a customized semantic disambiguation task after passive reading. EEG signals were acquired using a dry-sensor OpenBCI system from four frontal sites (Fp1, Fp2, F3, F4). Real-time attention indices were computed based on the beta/(alpha+theta) ratio and fed back visually in the neurofeedback condition. Cognitive outcomes included comprehension test scores and semantic conflict resolution performance (RT, accuracy, cognitive load).

RESULTS: Compared to sham and control conditions, neurofeedback significantly improved comprehension accuracy (p < 0.001), reduced reaction times in the interference task (p < 0.05), and lowered subjective cognitive load (p = 0.002). EEG indices of attention were significantly elevated during neurofeedback (p < 0.001) and positively correlated with behavioral gains (r = 0.63, p < 0.05).

CONCLUSIONS: Portable dry-electrode EEG systems can reliably support real-time neurofeedback to enhance attention and cognitive control in complex language learning contexts. This study provides empirical validation for deploying dry-EEG sensors in adaptive educational technologies and contributes to the broader integration of wearable brain-computer interfaces in cognitive augmentation applications.},
}

Humans
*Neurofeedback/methods
Male
Female
Young Adult
*Electroencephalography/methods
Adult
*Learning/physiology
Attention/physiology
Cognition/physiology
Comprehension/physiology
Language
Adolescent
East Asian People

@article {pmid40586414,
year = {2025},
author = {Tian, Y and Wallace, DM and Cederna, PS and Chestek, CA and Kemp, SWP},
title = {Toward Natural Limb Function: A New Era in Prosthetic Innovation.},
journal = {Annals of neurology},
volume = {98},
number = {5},
pages = {913-928},
doi = {10.1002/ana.27287},
pmid = {40586414},
issn = {1531-8249},
mesh = {Humans ; *Artificial Limbs/trends ; *Brain-Computer Interfaces/trends ; *Extremities/physiology ; Electroencephalography ; },
abstract = {The past decade has witnessed groundbreaking clinical implementation of neuroprosthetic limbs driven by signals from peripheral targets (eg, nerves and muscle) and the brain to restore limb function for individuals with limb loss or impairment. In this review, we highlight recent key clinical trials in peripheral neuroprosthetic interfaces directly with nerve, residual muscle, and reinnervated muscle. We then highlight the key advances in brain interfaces, including clinical trials using electroencephalography, electrocorticography, and intracortical electrodes to control neuroprosthetics. Finally, we explore the future of neuroprosthetic control where both peripheral and brain interfaces can be combined to improve neuroprosthetic performance. ANN NEUROL 2025;98:913-928.},
}

Humans
*Artificial Limbs/trends
*Brain-Computer Interfaces/trends
*Extremities/physiology
Electroencephalography

@article {pmid41161815,
year = {2025},
author = {Yang, Y and Liu, C and Liu, S and Ding, P and Bai, R and Chen, G and Li, S and Song, X and Cheng, Y and Xu, J},
title = {Role of combination immunotherapy in restoring brain synergistic functional connectivity in patients with systemic lupus erythematosus without overt neuropsychiatric manifestations.},
journal = {Lupus science & medicine},
volume = {12},
number = {2},
pages = {},
doi = {10.1136/lupus-2025-001771},
pmid = {41161815},
issn = {2053-8790},
mesh = {Humans ; Female ; *Lupus Erythematosus, Systemic/drug therapy/physiopathology ; Adult ; Male ; Cyclophosphamide/therapeutic use/administration & dosage ; Magnetic Resonance Imaging ; *Immunosuppressive Agents/therapeutic use/administration & dosage ; Hydroxychloroquine/therapeutic use/administration & dosage ; *Brain/physiopathology/diagnostic imaging/drug effects ; Glucocorticoids/therapeutic use/administration & dosage ; Middle Aged ; Case-Control Studies ; Drug Therapy, Combination ; Young Adult ; *Immunotherapy/methods ; },
abstract = {OBJECTIVE: To determine whether subclinical brain dysfunction in SLE can be detected by disrupted interhemispheric connectivity and assess its modulation by immunosuppressive regimens.

METHODS: 234 subjects (140 patients with SLE and 94 healthy controls (HCs)) were included. Through stratified analysis, patients with SLE were divided into treatment-naïve group (n=22), glucocorticoid monotherapy group (GC group, n=30) and GC combined with cyclophosphamide (CTX) and/or hydroxychloroquine (HCQ) treatment group (n=50) to assess the differences in voxel-mirrored homotopic connectivity (VMHC) between groups.

RESULTS: SLE group showed lower VMHC than the HC group in bilateral superior temporal gyrus, medial superior frontal gyrus, calcarine fissure and surrounding cortex and middle occipital cortices (Gaussian random field corrected: voxel p<0.005, cluster p<0.01). The VMHC in the bilateral superior temporal gyrus (rs=-0.250, p=0.024) and medial superior frontal gyrus (rs=-0.246, p=0.026) was negatively correlated with the depression score, while the VMHC in the medial superior frontal gyrus was negatively correlated with the anxiety score (rs=-0.239, p=0.031). Three SLE subgroups and HCs had different VMHC in the postcentral/precentral gyrus (F=8.942) and anterior cingulate/paracingulate gyrus (F=9.868). Post hoc analysis found that compared with the HC group, VMHC in the treatment-naïve group was decreased in the bilateral posterior central gyrus (t=-2.953), while in the GC monotherapy group, it decreased in the posterior central gyrus (t=-2.999) and anterior cingulate/paracingulate gyrus (t=-2.999). Compared with GC combined with CTX and/or HCQ group, VMHC in GC monotherapy group was decreased in the postcentral gyrus (t=-2.999).

CONCLUSION: Even without overt neuropsychiatric symptoms, patients with SLE exhibit impaired interhemispheric functional synergy that is partially reversed by combination immunosuppression, suggesting an early targetable brain pathway.},
}

Humans
Female
*Lupus Erythematosus, Systemic/drug therapy/physiopathology
Adult
Male
Cyclophosphamide/therapeutic use/administration & dosage
Magnetic Resonance Imaging
*Immunosuppressive Agents/therapeutic use/administration & dosage
Hydroxychloroquine/therapeutic use/administration & dosage
*Brain/physiopathology/diagnostic imaging/drug effects
Glucocorticoids/therapeutic use/administration & dosage
Middle Aged
Case-Control Studies
Drug Therapy, Combination
Young Adult
*Immunotherapy/methods

@article {pmid41160913,
year = {2025},
author = {Russo, JS and Colebatch, JG and Lin, CS and John, SE and Grayden, DB and Todd, NPM},
title = {Feasibility of decoding cerebellar movement-related potentials for brain-computer interface applications.},
journal = {Journal of neural engineering},
volume = {},
number = {},
pages = {},
doi = {10.1088/1741-2552/ae18fa},
pmid = {41160913},
issn = {1741-2552},
abstract = {In Brain-Computer Interface (BCI) applications, signals are conventionally acquired from the cerebrum, and only a subset of the complex interactions that occur in several areas of the brain are collected. One area that has not been investigated for BCI application is the cerebellum, despite its involvement in movement and executive function. The present study aimed to determine the features of movement-related cerebellar electrocerebellography (ECeG) that are most useful for decoding, and how performance compares with conventional electroencephalography (EEG) recordings from the cerebrum. Approach. ECeG and EEG data were collected from six healthy adults to identify useful movement-related features from both cerebrum and cerebellum. Electromyography was used to capture the movements from the muscles. Decoding was conducted in binary movement vs. rest and movement vs. movement systems using support vector machines. Decoding performance was compared between cerebral, cerebellar, a combination of both, and temporal groups. Re-referencing techniques were applied to compensate for possible common reference artefacts or volume conduction effects. Main Results. Movement-related features were decoded from over the cerebellum and the cerebrum. Classification accuracies were similar in both the cerebrum and cerebellum, when classifying movement vs. rest (cerebrum: 0.78 ± 0.02, cerebellum: 0.70 ± 0.01) and movement vs. movement states (cerebrum: 0.76 ± 0.02, cerebellum: 0.71 ± 0.02). The delta band (1-3 Hz) was the most useful feature for decoding. Significance. This study demonstrated, for the first time, that ECeG is a feasible source of movement related signals for implementing a BCI. The present study also demonstrated that the ECeG closely resembled the EEG signals and represents an alternate approach for BCI where the signal from the cerebrum is unreliable either due to disease or injury. .},
}

@article {pmid41160812,
year = {2025},
author = {Li, J and Lu, Y and Li, Z and Jin, L and Zhou, L and Ding, K and Liu, J and Hu, B and Liu, P and An, D and Liang, F and Hu, Y and Shao, Y and Ding, Y and Ma, L and Li, R and Mei, Y and Zhang, R and Song, E},
title = {An Active, Multimodal Neural Interface for Real-Time Monitoring of Cortical Electrical, Thermal, and Optical Dynamics.},
journal = {Advanced science (Weinheim, Baden-Wurttemberg, Germany)},
volume = {},
number = {},
pages = {e12114},
doi = {10.1002/advs.202512114},
pmid = {41160812},
issn = {2198-3844},
support = {2022ZD0209900//STI 2030-Major Project/ ; 62204057//National Natural Science Foundation of China/ ; 62304044//National Natural Science Foundation of China/ ; 82304124//National Natural Science Foundation of China/ ; U2230108//National Natural Science Foundation of China/ ; 22ZR1406400//Science and Technology Commission of Shanghai Municipality/ ; LGL-8998-09//Lingang Laboratory/ ; SKLICS-K202515//State Key Laboratory of Integrated Chips and Systems/ ; },
abstract = {Chronic neurophysiological monitoring devices facilitate the timely diagnosis and treatment of episodic or recurrent neurological disorders. Compared with passive electrodes, silicon-based active transistors provide intrinsic signal amplification and, when combined with capacitive-coupling measurement mechanisms, enable high-density, high-fidelity recordings. However, most existing systems remain limited to single-modality electrical sensing and fail to address the growing demands of contemporary neurodynamic research. Here, a chronically implantable, large-area cortical interface capable of real-time multimodal monitoring of electrical, thermal, and photodynamic signals is presented. Building upon a silicon-transistor array for neural electrical detection, the device integrates thin-film metal resistors for temperature sensing while preserving mechanical flexibility sufficient for stable, long-term tissue contact. By leveraging the photoelectric effect of silicon transistors and functional multiplexing of active elements, the interface also achieves precise photodynamic measurement. In vitro experiments confirm long-term stability and channel isolation. In vivo evaluation in Sprague-Dawley rats, together with biocompatibility assessments, demonstrates reliable performance under physiological conditions. The technology used in this multifunctional platform has universal applicability in neural interfaces, offering continuous multimodal neurodynamic data acquisition with potential utility in monitoring, diagnosing, and treating chronic neurological conditions such as epilepsy and brain tumors.},
}

@article {pmid41160441,
year = {2025},
author = {Jeong, SY and Lee, JW and Kim, TG},
title = {Comparative analysis across diverse plant species reveals superior antibiofilm efficacy and dose-dependency of root extracts compared to leaf extracts.},
journal = {FEMS microbiology letters},
volume = {},
number = {},
pages = {},
doi = {10.1093/femsle/fnaf116},
pmid = {41160441},
issn = {1574-6968},
abstract = {Although both root- and leaf-derived plant extracts hold potential as antibiofilm agents, research has predominantly focused on leaf tissues. In this study, we systematically compared the antibiofilm efficacy of 158 root and 248 leaf extracts from 360 plant species across five concentrations (0.1, 0.25, 0.5, 1.0, and 2.0 g/L). As concentration increased, the biological control incidence (BCI) of root extracts rose from 68.4% to 94.3%, while leaf extracts showed a smaller increase, from 52.2% to 71.7%. Similarly, the biological control efficacy (BCE) of root extracts increased from 27.6% to 54.2%, whereas leaf extracts ranged from -2.7% to 16.2%. Bootstrapping analysis (10 000 iterations) confirmed significantly higher antibiofilm activity of root extracts at concentrations ≥ 0.5 g/L (P < 0.05). Paired comparisons of species with both extract types further demonstrated the consistent superiority of root extracts across all concentrations (bootstrapped, P < 0.05), despite interspecific variation at higher doses. Linear regression revealed a significantly steeper dose-response slope for root extracts (29.2 ± 2.4) than for leaf extracts (8.1 ± 2.8) (bootstrapped, P < 0.05), indicating a stronger concentration-dependent effect of root extracts. These results suggest that plant roots typically harbor more potent and/or diverse antibiofilm compounds than leaves, underscoring their untapped potential for biofilm control applications.},
}

@article {pmid41160433,
year = {2025},
author = {Chen, Y and Liu, T and Jia, K and Theeuwes, J and Gong, M},
title = {Dual-format attentional template during preparation in human visual cortex.},
journal = {eLife},
volume = {13},
number = {},
pages = {},
doi = {10.7554/eLife.103425},
pmid = {41160433},
issn = {2050-084X},
support = {Major Project 2021ZD0200409//National Science and Technology Innovation 2030/ ; 32371087//National Natural Science Foundation of China/ ; 32300855//National Natural Science Foundation of China/ ; 3200784//National Natural Science Foundation of China/ ; 226-2024-00118//Fundamental Research Funds for the Central University/ ; Non-profit Central Research Institute Fund 2023-PT310-01//Chinese Academy of Medical Sciences/ ; },
mesh = {Humans ; *Attention/physiology ; *Visual Cortex/physiology ; Magnetic Resonance Imaging ; Male ; Female ; Adult ; Young Adult ; *Visual Perception/physiology ; Brain Mapping ; Cues ; Photic Stimulation ; },
abstract = {Goal-directed attention relies on forming internal templates of key information relevant for guiding behavior, particularly when preparing for upcoming sensory inputs. However, evidence on how these attentional templates are represented during preparation remains controversial. Here, we combine functional magnetic resonance imaging with an orientation cueing task to isolate preparatory activity from stimulus-evoked responses. Using multivariate pattern analysis, we found decodable information about the to-be-attended orientation during preparation; yet preparatory activity patterns were different from those evoked when actual orientations were perceived. When perturbing the neural activity by means of a visual impulse ('pinging' technique), the preparatory activity patterns in visual cortex resembled those associated with perceiving these orientations. The observed differential patterns with and without the impulse perturbation suggest a predominantly non-sensory format and a latent, sensory-like format of representation during preparation. Furthermore, the emergence of the sensory-like template coincided with enhanced information connectivity between V1 and frontoparietal areas and was associated with improved behavioral performance. By engaging this dual-format mechanism during preparation, the brain is able to encode both abstract, non-sensory information and more detailed, sensory information, potentially providing advantages for adaptive attentional control. For example, consistent with recent theories of visual search, a predominantly non-sensory template can support the initial guidance and a latent sensory-like format can support prospective stimulus processing.},
}

Humans
*Attention/physiology
*Visual Cortex/physiology
Magnetic Resonance Imaging
Male
Female
Adult
Young Adult
*Visual Perception/physiology
Brain Mapping
Cues
Photic Stimulation

@article {pmid41159356,
year = {2025},
author = {Atan, Y and Doğan, M and Karayel, F and Üzün, İ},
title = {Fatal Isolated Right Ventricular Rupture Without External Chest Injury in a Young Driver: Forensic Autopsy Findings After a One-Sided Vehicle Collision.},
journal = {Archives of Iranian medicine},
volume = {28},
number = {9},
pages = {530-535},
doi = {10.34172/aim.34401},
pmid = {41159356},
issn = {1735-3947},
mesh = {Humans ; Male ; *Accidents, Traffic ; Young Adult ; *Heart Ventricles/injuries/pathology ; Fatal Outcome ; Autopsy ; *Heart Injuries/pathology/etiology ; *Wounds, Nonpenetrating/pathology ; Forensic Pathology ; },
abstract = {Traumatic deaths are common, with cardiac trauma affecting 7‒12% of patients with thoracic injuries. Blunt cardiac injury (BCI), although rare, is associated with a high mortality rate. This report presents a case of blunt cardiac rupture (BCR) observed at autopsy despite the absence of external chest trauma, suggesting the presence of severe internal injuries. A 19-year-old male was found dead in his vehicle which had collided with a wall. At the crime scene investigation, external examination revealed no substantial chest wall injuries in the individual despite significant damage to the vehicle. Autopsy revealed a 2-cm rupture of the right ventricle (heart), accompanied by 400 cc of partially coagulated blood in the pericardial cavity, consistent with cardiac tamponade. Pregabalin was detected in the toxicology analysis, but not in lethal concentrations. Traffic accidents are a major cause of BCI, typically resulting from compression of the heart between the thoracic structures during high-energy impacts. BCR is particularly fatal and often results in rapid death before arrival to the hospital. The absence of external trauma in the current case underscores the need for thorough internal examination in trauma-related deaths.},
}

Humans
Male
*Accidents, Traffic
Young Adult
*Heart Ventricles/injuries/pathology
Fatal Outcome
Autopsy
*Heart Injuries/pathology/etiology
*Wounds, Nonpenetrating/pathology
Forensic Pathology

@article {pmid41157441,
year = {2025},
author = {Moreno-Castelblanco, SR and Vélez-Guerrero, MA and Callejas-Cuervo, M},
title = {Lower-Limb Motor Imagery Recognition Prototype Based on EEG Acquisition, Filtering, and Machine Learning-Based Pattern Detection.},
journal = {Sensors (Basel, Switzerland)},
volume = {25},
number = {20},
pages = {},
doi = {10.3390/s25206387},
pmid = {41157441},
issn = {1424-8220},
mesh = {Humans ; *Electroencephalography/methods ; *Machine Learning ; Brain-Computer Interfaces ; *Lower Extremity/physiology ; Signal Processing, Computer-Assisted ; Male ; Algorithms ; Adult ; *Imagination/physiology ; Movement/physiology ; *Pattern Recognition, Automated/methods ; Female ; Young Adult ; },
abstract = {Advances in brain-computer interface (BCI) research have explored various strategies for acquiring and processing electroencephalographic (EEG) signals to detect motor imagery (MI) activities. However, the complexity of multichannel clinical systems and processing techniques can limit their accessibility outside specialized centers, where complex setups are not feasible. This paper presents a proof-of-concept prototype of a single-channel EEG acquisition and processing system designed to identify lower-limb motor imagery. The proposed proof-of-concept prototype enables the wireless acquisition of raw EEG values, signal processing using digital filters, and the detection of MI patterns using machine learning algorithms. Experimental validation in a controlled laboratory with participants performing resting, MI, and movement tasks showed that the best performance was obtained by combining Savitzky-Golay filtering with a Random Forest classifier, reaching 87.36% ± 4% accuracy and an F1-score of 87.18% ± 3.8% under five-fold cross-validation. These findings confirm that, despite limited spatial resolution, MI patterns can be detected using appropriate AI-based filtering and classification. The novelty of this work lies in demonstrating that a single-channel, portable EEG prototype can be effectively used for lower-limb MI recognition. The portability and noise resilience achieved with the prototype highlight its potential for research, clinical rehabilitation, and assistive device control in non-specialized environments.},
}

Humans
*Electroencephalography/methods
*Machine Learning
Brain-Computer Interfaces
*Lower Extremity/physiology
Signal Processing, Computer-Assisted
Male
Algorithms
Adult
*Imagination/physiology
Movement/physiology
*Pattern Recognition, Automated/methods
Female
Young Adult

@article {pmid41157340,
year = {2025},
author = {Iadarola, G and Mengarelli, A and Iarlori, S and Monteriù, A and Spinsante, S},
title = {RGB-D Cameras and Brain-Computer Interfaces for Human Activity Recognition: An Overview.},
journal = {Sensors (Basel, Switzerland)},
volume = {25},
number = {20},
pages = {},
doi = {10.3390/s25206286},
pmid = {41157340},
issn = {1424-8220},
support = {CUP I33C2200133000//Vitality Project/ ; },
mesh = {Humans ; *Brain-Computer Interfaces ; *Human Activities ; Electroencephalography ; Brain/physiology ; },
abstract = {This paper provides a perspective on the use of RGB-D cameras and non-invasive brain-computer interfaces (BCIs) for human activity recognition (HAR). Then, it explores the potential of integrating both the technologies for active and assisted living. RGB-D cameras can offer monitoring of users in their living environments, preserving their privacy in human activity recognition through depth images and skeleton tracking. Concurrently, non-invasive BCIs can provide access to intent and control of users by decoding neural signals. The synergy between these technologies may allow holistic understanding of both physical context and cognitive state of users, to enhance personalized assistance inside smart homes. The successful deployment in integrating the two technologies needs addressing critical technical hurdles, including computational demands for real-time multi-modal data processing, and user acceptance challenges related to data privacy, security, and BCI illiteracy. Continued interdisciplinary research is essential to realize the full potential of RGB-D cameras and BCIs as AAL solutions, in order to improve the quality of life for independent or impaired people.},
}

Humans
*Brain-Computer Interfaces
*Human Activities
Electroencephalography
Brain/physiology

@article {pmid41156422,
year = {2025},
author = {He, J and Xu, J and Wang, Y},
title = {Non-Linear Modeling and Precision Analysis Approach for Implantable Multi-Channel Neural Recording Systems.},
journal = {Micromachines},
volume = {16},
number = {10},
pages = {},
doi = {10.3390/mi16101176},
pmid = {41156422},
issn = {2072-666X},
support = {2021ZD0200401//STI 2030-Major Project/ ; 2025C01187,2024C03001//Pioneer R&D Program of Zhejiang/ ; 62176232,62336007//National Natural Science Foundation of China/ ; SNZJU-SIAS-002//Starry Night Science Fund of Zhe- 406 jiang University Shanghai Institute for Advanced Study/ ; 2025ZFJH01,226-2024-00127//Fundamental Research Funds for the Central Universities/ ; },
abstract = {High-precision implantable multi-channel neural recording systems are considered as having a crucial role in the diagnosis and treatment of neurological disorders. However, it is a significant design challenge to achieve an optimal trade-off among linear parameters, signal fidelity, power consumption, and circuit area. To address this challenge, a Simulink-based modeling approach has been proposed to incorporate adjustable non-linear parameters across the front-end circuits and analog-to-digital converter (ADC) stages. The model evaluates non-linearity impacts on system performance through both quantitative spike detection accuracy analysis and a neural decoding paradigm based on Chinese handwriting reconstruction. Simulated results show that total harmonic distortion (THD) can be set to -34.32 dB for the low-noise amplifier (LNA), -33.73 dB for the programmable gain amplifier (PGA), and -57.95 dB for the ADC in order to achieve reliable detection accuracy with minimal design cost. Moreover, ADC non-linearity has a greater influence on system performance than that of the LNA and PGA. The proposed approach offers quantitative and systematic hardware design guidance to balance signal fidelity and resource efficiency for future low-power, high-accuracy neural recording systems.},
}

@article {pmid41155027,
year = {2025},
author = {Yao, Y and Wang, X and Hao, X and Sun, H and Dong, R and Li, Y},
title = {Trans-cVAE-GAN: Transformer-Based cVAE-GAN for High-Fidelity EEG Signal Generation.},
journal = {Bioengineering (Basel, Switzerland)},
volume = {12},
number = {10},
pages = {},
doi = {10.3390/bioengineering12101028},
pmid = {41155027},
issn = {2306-5354},
abstract = {Electroencephalography signal generation remains a challenging task due to its non-stationarity, multi-scale oscillations, and strong spatiotemporal coupling. Conventional generative models, including VAEs and GAN variants such as DCGAN, WGAN, and WGAN-GP, often yield blurred waveforms, unstable spectral distributions, or lack semantic controllability, limiting their effectiveness in emotion-related applications. To address these challenges, this research proposes a Transformer-based conditional variational autoencoder-generative adversarial network (Trans-cVAE-GAN) that combines Transformer-driven temporal modeling, label-conditioned latent inference, and adversarial learning. A multi-dimensional structural loss further constrains generation by preserving temporal correlation, frequency-domain consistency, and statistical distribution. Experiments on three SEED-family datasets-SEED, SEED-FRA, and SEED-GER-demonstrate high similarity to real EEG, with representative mean ± SD correlations of Pearson ≈ 0.84 ± 0.08/0.74 ± 0.12/0.84 ± 0.07 and Spearman ≈ 0.82 ± 0.07/0.72 ± 0.12/0.83 ± 0.08, together with low spectral divergence (KL ≈ 0.39 ± 0.15/0.41 ± 0.20/0.37 ± 0.18). Comparative analyses show consistent gains over classical GAN baselines, while ablations verify the indispensable roles of the Transformer encoder, label conditioning, and cVAE module. In downstream emotion recognition, augmentation with generated EEG raises accuracy from 86.9% to 91.8% on SEED (with analogous gains on SEED-FRA and SEED-GER), underscoring enhanced generalization and robustness. These results confirm that the proposed approach simultaneously ensures fidelity, stability, and controllability across cohorts, offering a scalable solution for affective computing and brain-computer interface applications.},
}

@article {pmid41154635,
year = {2025},
author = {Tabish, M and Malik, I and Akhtar, A and Afzal, M},
title = {A Review on Low-Dimensional Nanoarchitectonics for Neurochemical Sensing and Modulation in Responsive Neurological Outcomes.},
journal = {Biomolecules},
volume = {15},
number = {10},
pages = {},
doi = {10.3390/biom15101405},
pmid = {41154635},
issn = {2218-273X},
support = {KSRG-2024-41//King Salman Center for Disability Research/ ; },
mesh = {Humans ; Brain-Computer Interfaces ; Animals ; Artificial Intelligence ; *Biosensing Techniques/methods ; *Nanostructures/chemistry ; Nanotechnology/methods ; Brain ; },
abstract = {Low-Dimensional Nanohybrids (LDNHs) have emerged as potent multifunctional platforms for neurosensing and neuromodulation, providing elevated spatial-temporal precision, versatility, and biocompatibility. This review examines the intersection of LDNHs with artificial intelligence, brain-computer interfaces (BCIs), and closed-loop neurotechnologies, highlighting their transformative potential in personalized neuro-nano-medicine. Utilizing stimuli-responsive characteristics, optical, thermal, magnetic, and electrochemical LDNHs provide real-time feedback-controlled manipulation of brain circuits. Their pliable and adaptable structures surpass the constraints of inflexible bioelectronics, improving the neuronal interface and reducing tissue damage. We also examined their use in less invasive neurological diagnostics, targeted therapy, and adaptive intervention systems. This review delineates recent breakthroughs, integration methodologies, and fundamental mechanisms, while addressing significant challenges such as long-term biocompatibility, deep-tissue accessibility, and scalable manufacturing. A strategic plan is provided to direct future research toward clinical use. Ultimately, LDNHs signify a transformative advancement in intelligent, tailored, and closed-loop neurotechnologies, integrating materials science, neurology, and artificial intelligence to facilitate the next era of precision medicine.},
}

Humans
Brain-Computer Interfaces
Animals
Artificial Intelligence
*Biosensing Techniques/methods
*Nanostructures/chemistry
Nanotechnology/methods
Brain

@article {pmid41154223,
year = {2025},
author = {Du, A and Huang, M and Wang, Z and Zhou, H and Duan, H and Hu, S and Zheng, Y},
title = {Using Low-Intensity Focused Ultrasound to Treat Depression and Anxiety Disorders: A Review of Current Evidence.},
journal = {Brain sciences},
volume = {15},
number = {10},
pages = {},
doi = {10.3390/brainsci15101129},
pmid = {41154223},
issn = {2076-3425},
support = {2023YFC2506200//National Key Research and Development Program of China/ ; 2023YFC2506203//National Key Research and Development Program of China/ ; 2022YFB3204300//National Key Research and Development Program of China/ ; 2022C01002//Zhejiang Provincial Key Research and Development Program of China/ ; },
abstract = {Background: Depression and anxiety disorders impact millions globally. In recent years, low-intensity focused ultrasound (LIFU), characterized by its high precision, deep penetration, and non-invasive nature, has garnered significant interest in neuroscience and clinical practice. To enhance understanding of its effects on mood, therapeutic availability in treatment of depression/anxiety disorders, and potential mechanisms, a systematic review of studies investigating the emotional impact of LIFU on depressive/anxious-like animal models, healthy volunteers, and patients with depression or anxiety disorders has been undertaken. Methods: Relevant papers published before 15 July 2025 were searched across four databases: Web of Science, PubMed, Science Direct, and Embase. A total of 28 papers which met the inclusion and exclusion criteria are included in this review. Results: Our findings indicate that LIFU reversed the depressive/anxious-like behaviors in the animal models and showed antidepressant/anti-anxiety effects among the state-of-art clinical studies. For example, immobility time in FST or TST is reduced in depressive animal models, and HRSD/BAI scales are improved in human studies. Key molecules such as BDNF/5-HT are found restored in animal models, and FC between key brain areas related to depression/anxiety is modulated after LIFU treatment. Notably, no brain tissue damage was observed in animal studies, and only mild adverse effects (such as dizziness and vomiting) were noted in a few human studies. Conclusions: The studies using LIFU to treat depression and anxiety remain in the preliminary stage. The mechanisms underlying LIFU's mood effects-such as activation or inhibition of specific brain regions or neural circuits, anti-inflammatory effects, alterations in functional connectivity, synaptic plasticity, neurotransmitter levels, and BDNF-remain incompletely understood and warrant further investigation. Nevertheless, the LIFU technique holds promise for regulating both cortical and subcortical brain areas implicated in depression/anxiety disorders as a precise neuromodulation tool.},
}

@article {pmid41154218,
year = {2025},
author = {Tan, L and Fang, H and Ding, P and Wang, F and Wei, Y and Fu, Y},
title = {P300 Spatiotemporal Prior-Based Transformer-CNN for Auxiliary Diagnosis of PTSD.},
journal = {Brain sciences},
volume = {15},
number = {10},
pages = {},
doi = {10.3390/brainsci15101124},
pmid = {41154218},
issn = {2076-3425},
support = {82172058, 62376112, 81771926, 61763022, 62366026, 62006246//The National Natural Science Foundation of China under Grant Nos/ ; },
abstract = {Objectives: To address the challenges of subjectivity, misdiagnosis and underdiagnosis in post-traumatic stress disorder (PTSD), this study proposes an objective auxiliary diagnostic method based on P300 signals. Existing studies largely rely on conventional P300 features, lacking the systematic integration of event-related potential (ERP) priors and facing limitations in spatiotemporal feature modeling. Methods: Using common spatiotemporal pattern (CSTP) analysis and quantitative evaluation, we revealed significant spatiotemporal differences in P300 signals between PTSD patients and healthy controls. ERP prior information was then extracted and integrated into a hybrid architecture combining transformer encoders and a convolutional neural network (CNN), enabling joint modeling of long-range temporal dependencies and local spatial patterns. Results: The proposed P300 spatiotemporal transformer-CNN (P300-STTCNet) achieved a classification accuracy of 93.37% in distinguishing PTSD from healthy controls, markedly outperforming traditional approaches. Conclusions: Significant spatiotemporal differences in P300 signals exist between PTSD and healthy control groups. The P300-STTCNet model effectively captures PTSD-related spatiotemporal features, demonstrating strong potential for electroencephalogram-based objective auxiliary diagnosis.},
}

@article {pmid41152182,
year = {2025},
author = {Cao, Y and Xue, Y and Yang, H and Wang, F and Li, T and Zhao, L and Fu, Y},
title = {[Ethical considerations for artificial intelligence-enhanced brain-computer interface].},
journal = {Sheng wu yi xue gong cheng xue za zhi = Journal of biomedical engineering = Shengwu yixue gongchengxue zazhi},
volume = {42},
number = {5},
pages = {1085-1091},
doi = {10.7507/1001-5515.202507024},
pmid = {41152182},
issn = {1001-5515},
mesh = {*Brain-Computer Interfaces/ethics ; *Artificial Intelligence/ethics ; Humans ; Deep Learning ; User-Computer Interface ; Electroencephalography ; },
abstract = {Artificial intelligence-enhanced brain-computer interfaces (BCI) are expected to significantly improve the performance of traditional BCIs in multiple aspects, including usability, user experience, and user satisfaction, particularly in terms of intelligence. However, such AI-integrated or AI-based BCI systems may introduce new ethical issues. This paper first evaluated the potential of AI technology, especially deep learning, in enhancing the performance of BCI systems, including improving decoding accuracy, information transfer rate, real-time performance, and adaptability. Building on this, it was considered that AI-enhanced BCI systems might introduce new or more severe ethical issues compared to traditional BCI systems. These include the possibility of making users' intentions and behaviors more predictable and manipulable, as well as the increased likelihood of technological abuse. The discussion also addressed measures to mitigate the ethical risks associated with these issues. It is hoped that this paper will promote a deeper understanding and reflection on the ethical risks and corresponding regulations of AI-enhanced BCIs.},
}

*Brain-Computer Interfaces/ethics
*Artificial Intelligence/ethics
Humans
Deep Learning
User-Computer Interface
Electroencephalography

@article {pmid41152174,
year = {2025},
author = {Wang, P and Ji, X and Wang, J and Yu, X},
title = {[Brain computer interface nursing bed control system based on deep learning and dual visual feedback].},
journal = {Sheng wu yi xue gong cheng xue za zhi = Journal of biomedical engineering = Shengwu yixue gongchengxue zazhi},
volume = {42},
number = {5},
pages = {1021-1028},
doi = {10.7507/1001-5515.202504047},
pmid = {41152174},
issn = {1001-5515},
mesh = {Humans ; *Brain-Computer Interfaces ; *Deep Learning ; Electroencephalography ; *Feedback, Sensory ; Neural Networks, Computer ; Beds ; },
abstract = {In order to meet the need of autonomous control of patients with severe limb disorders, this paper designs a nursing bed control system based on motor imagery-brain computer interface (MI-BCI). In view of the low decoding performance of cross-subjects and the dynamic fluctuation of cognitive state in the existing MI-BCI technology, the neural network structure optimization and user interaction feedback enhancement are improved. Firstly, the optimized dual-branch graph convolution multi-scale neural network integrates dynamic graph convolution and multi-scale convolution. The average classification accuracy is higher than that of multi-scale attention temporal convolution network, Gram angle field combined with convolution long short term memory hybrid network, Transformer-based graph convolution network and other existing methods. Secondly, a dual visual feedback mechanism is constructed, in which electroencephalogram (EEG) topographic map feedback can improve the discrimination of spatial patterns, and attention state feedback can enhance the temporal stability of signals. Compared with the single EEG topographic map feedback and non-feedback system, the average classification accuracy of the proposed method is also greatly improved. Finally, in the four classification control task of nursing bed, the average control accuracy of the system is 90.84%, and the information transmission rate is 84.78 bits/min. In summary, this paper provides a reliable technical solution for improving the autonomous interaction ability of patients with severe limb disorders, which has important theoretical significance and application value.},
}

Humans
*Brain-Computer Interfaces
*Deep Learning
Electroencephalography
*Feedback, Sensory
Neural Networks, Computer
Beds

@article {pmid41152144,
year = {2025},
author = {Bek, J and Aziz, A and Brady, N},
title = {Transcranial Direct Current Stimulation to Augment Motor Imagery Training: A Systematic Review.},
journal = {The European journal of neuroscience},
volume = {62},
number = {8},
pages = {e70280},
doi = {10.1111/ejn.70280},
pmid = {41152144},
issn = {1460-9568},
support = {101034345//H2020 Marie Skłodowska-Curie Actions/ ; },
mesh = {Humans ; *Transcranial Direct Current Stimulation/methods ; *Motor Cortex/physiology ; *Imagination/physiology ; Brain-Computer Interfaces ; *Imagery, Psychotherapy/methods ; Neuronal Plasticity ; *Stroke Rehabilitation/methods ; },
abstract = {Motor imagery training (MIT) is a widely used technique for motor learning and recovery. To optimize training outcomes, researchers have explored the integration of MIT with complementary approaches. One such approach is transcranial direct current stimulation (tDCS), which also shows promise as a method to enhance motor performance and neuroplasticity. This systematic review aimed to synthesize the current evidence on the synergistic effects of MIT combined with tDCS, with a specific focus on behavioral outcomes. Heterogeneous methods across 16 studies with 432 participants in total, including both healthy and clinical populations, yielded mixed results. Nonetheless, the potential of anodal tDCS applied over the primary motor cortex to augment the beneficial effects of MIT for motor performance in healthy participants is suggested by the current literature. The benefits of combining tDCS with MIT in brain-computer interface (BCI) protocols with stroke patients were less clear, which may relate to population differences, timing of stimulation, or the similarity between outcome measures and trained tasks. Overall, small samples and heterogeneous methods limit interpretation of the findings of combined intervention studies, and further research should aim to measure both behavioral and neurophysiological outcomes in larger samples as well as examining longer-term synergistic effects.},
}

Humans
*Transcranial Direct Current Stimulation/methods
*Motor Cortex/physiology
*Imagination/physiology
Brain-Computer Interfaces
*Imagery, Psychotherapy/methods
Neuronal Plasticity
*Stroke Rehabilitation/methods

@article {pmid40967240,
year = {2025},
author = {Del Sesto, MJ and Negoita, S and Bruzzone Giraldez, M and LaJoie, Z and Akhter Sathi, K and Wong, JK and Widge, AS and Okun, MS and Khalifa, A},
title = {Multitarget neurostimulation of the deep brain: clinical opportunities, challenges, and emerging technologies.},
journal = {Journal of neural engineering},
volume = {22},
number = {5},
pages = {},
pmid = {40967240},
issn = {1741-2552},
support = {DP2 EB037188/EB/NIBIB NIH HHS/United States ; },
mesh = {Humans ; *Deep Brain Stimulation/methods/trends/instrumentation ; Animals ; *Brain/physiology ; *Brain-Computer Interfaces/trends ; Electrodes, Implanted/trends ; },
abstract = {Recent computational, pre-clinical, and clinical studies have demonstrated the potential for using neuromodulation through simultaneous targeting of multiple deep brain regions. This approach has already been used for therapeutic and systems neuroscience applications. However, the broad clinical adoption of invasive distributed deep brain interfaces remains in its early stages. This review explores the barriers to implementation by addressing three key questions: do the benefits of implanting multiple electrodes justify the associated risks for specific applications? What is the risk-benefit ratio, and what technological advancements will be necessary to encourage clinical adoption? We also examine next-generation technologies that could enable multi-target brain interfaces, including system-on-chip micro-stimulators as well as nanoparticles. We highlight the role of novel machine learning algorithms in the optimization of stimulation parameters and for the guidance of device placement. Emerging hardware accelerators equipped with on-chip AI have demonstrated functionality that can be used to decode and to classify distributed neuronal data. This advance in hardware accelerators has also contributed to the potential for enhanced closed-loop stimulation control of devices. Despite these advances, significant technological and translational barriers persist, limiting the widespread clinical application of multi-target brain interfaces. This review provides a critical analysis of recent prototypes and novel hardware for use in multi-target systems. We will discuss both clinical and research applications. We will focus on the utilization of multi-site technologies to meet the needs of neurological diseases. We conclude that there exists a critical need for further innovation and integration of multi-site technologies into clinical practice.},
}

Humans
*Deep Brain Stimulation/methods/trends/instrumentation
Animals
*Brain/physiology
*Brain-Computer Interfaces/trends
Electrodes, Implanted/trends

@article {pmid41151221,
year = {2025},
author = {Liu, S and Su, L and He, Q and Qiu, M and Liang, R},
title = {Comparative evaluation of ChatGPT and Gemini in brain-computer interfaces patient education: A multi-dimensional analysis of reliability, accuracy, comprehensibility, and readability.},
journal = {International journal of medical informatics},
volume = {206},
number = {},
pages = {106164},
doi = {10.1016/j.ijmedinf.2025.106164},
pmid = {41151221},
issn = {1872-8243},
abstract = {BACKGROUND: Brain-Computer Interfaces (BCI) are a type of life-altering neurotechnology, but their inherent complexity poses significant challenges to patient education. Large Language Models (LLMs), such as ChatGPT and Gemini, offer new possibilities to address this challenge. This study aims to conduct a multi-dimensional, rigorous comparative analysis of the performance of these two mainstream AI models in responding to common patient questions related to BCI.

METHODS: Through a structured process combining clinical expert consensus, literature review, and online patient community analysis, we identified 13 key patient questions covering the entire BCI treatment cycle. We then obtained responses to these questions from ChatGPT and Gemini on September 1, 2025. An evaluation panel, composed of clinical experts and non-medical professionals, conducted a blinded assessment of the response quality using standardized Likert scales across three dimensions: reliability, accuracy, and comprehensibility. Concurrently, we performed an objective, quantitative analysis of the response texts using the Flesch-Kincaid readability tests.

RESULTS: On core quality metrics such as reliability, accuracy, and comprehensibility, the performance of the two models was generally comparable, both demonstrating a high level of proficiency with only sporadic statistical differences on a few technical questions. However, a clear significant disparity emerged in the dimension of readability: for 12 of the 13 questions, the text generated by Gemini required a significantly lower reading grade level than that of ChatGPT (p < 0.05) and had significantly higher reading ease scores. This difference stemmed from Gemini's tendency to use shorter sentences and simpler vocabulary.

CONCLUSION: AI chatbots possess immense potential in the field of BCI patient education. Although both ChatGPT and Gemini can provide high-quality information, Gemini demonstrates a clear advantage in the accessibility and approachability of information, making it a potentially more suitable tool for initial application across diverse patient populations. Nevertheless, the limitations of AI in handling highly specialized and dynamically changing knowledge underscore the indispensable role of human expert supervision and validation in any clinical application.},
}

@article {pmid41149457,
year = {2025},
author = {Lee, HH and Siu-Li, N and Pagano, I and Fukui, JA},
title = {Examining a Genomic Test in Predicting Extended Endocrine Benefit and Recurrence Risk in a Diverse Breast Cancer Population.},
journal = {Current oncology (Toronto, Ont.)},
volume = {32},
number = {10},
pages = {},
doi = {10.3390/curroncol32100537},
pmid = {41149457},
issn = {1718-7729},
support = {P30CA071789-17S2/CA/NCI NIH HHS/United States ; },
mesh = {Humans ; *Breast Neoplasms/genetics/drug therapy/pathology ; Female ; Middle Aged ; *Neoplasm Recurrence, Local/genetics ; Aged ; Retrospective Studies ; Adult ; *Genomics/methods ; *Antineoplastic Agents, Hormonal/therapeutic use ; },
abstract = {(1) Background: Extended endocrine therapy (EET) beyond five years can reduce distant recurrence in early-stage hormone receptor-positive (HR+) breast cancer. The Breast Cancer Index (BCI) predicts recurrence risk and EET benefits, yet racial/ethnic differences in its results remain unexplored. This study evaluates such differences in a diverse early-stage HR+ breast cancer population. (2) Methods: We retrospectively analyzed demographics, tumor characteristics and BCI scores of 159 women in Hawaii with early-stage HR+ breast cancer, self-identifying as Caucasian, Filipino, Japanese, Native Hawaiian, Other Asian/Pacific Islander, or Other. Tumor characteristics included size, grade, histology, lymph node/receptor status, Oncotype DX score, and laterality. Logistic regression used demographics and tumor features as predictor variables, with BCI's benefit prediction and recurrence risk as outcome variables. (3) Results: Japanese and other Asian/Pacific Islander patients had significantly lower odds of high recurrence risk compared to Caucasian patients. Higher recurrence risk was associated with greater odds of predicted EET. Racial/ethnic differences in EET benefit prediction were not statistically significant. (4) Conclusions: No racial/ethnic differences in EET benefit prediction suggest BCI's applicability in racially and ethnically diverse populations. Findings among Japanese and other Asian/Pacific Islanders point to potential biological or socioeconomic variation. Limitations include sample size and underrepresentation of certain groups. Future studies should address these gaps and adjust for known risk factors to further clarify BCI's racial and ethnic implications.},
}

Humans
*Breast Neoplasms/genetics/drug therapy/pathology
Female
Middle Aged
*Neoplasm Recurrence, Local/genetics
Aged
Retrospective Studies
Adult
*Genomics/methods
*Antineoplastic Agents, Hormonal/therapeutic use

@article {pmid41149344,
year = {2025},
author = {Kucukselbes, H and Sayilgan, E},
title = {Real-Time EEG Decoding of Motor Imagery via Nonlinear Dimensionality Reduction (Manifold Learning) and Shallow Classifiers.},
journal = {Biosensors},
volume = {15},
number = {10},
pages = {},
doi = {10.3390/bios15100692},
pmid = {41149344},
issn = {2079-6374},
mesh = {Humans ; *Electroencephalography/methods ; Adult ; Male ; Signal Processing, Computer-Assisted ; Female ; Machine Learning ; Brain-Computer Interfaces ; Young Adult ; Dimensionality Reduction ; },
abstract = {This study introduces a real-time processing framework for decoding motor imagery EEG signals by integrating manifold learning techniques with shallow classifiers. EEG recordings were obtained from six healthy participants performing five distinct wrist and hand motor imagery tasks. To address the challenges of high dimensionality and inherent nonlinearity in EEG data, five nonlinear dimensionality reduction methods, t-SNE, ISOMAP, LLE, Spectral Embedding, and MDS, were comparatively evaluated. Each method was combined with three shallow classifiers (k-NN, Naive Bayes, and SVM) to investigate performance across binary, ternary, and five-class classification settings. Among all tested configurations, the t-SNE + k-NN pairing achieved the highest accuracies, reaching 99.7% (two-class), 99.3% (three-class), and 89.0% (five-class). ISOMAP and MDS also delivered competitive results, particularly in multi-class scenarios. The presented approach builds upon our previous work involving EEG datasets from individuals with spinal cord injury (SCI), where the same manifold techniques were examined extensively. Comparative findings between healthy and SCI groups reveal consistent advantages of t-SNE and ISOMAP in preserving class separability, despite higher overall accuracies in healthy subjects due to improved signal quality. The proposed pipeline demonstrates low-latency performance, completing signal processing and classification in approximately 150 ms per trial, thereby meeting real-time requirements for responsive BCI applications. These results highlight the potential of nonlinear dimensionality reduction to enhance real-time EEG decoding, offering a low-complexity yet high-accuracy solution applicable to both healthy users and neurologically impaired individuals in neurorehabilitation and assistive technology contexts.},
}

Humans
*Electroencephalography/methods
Adult
Male
Signal Processing, Computer-Assisted
Female
Machine Learning
Brain-Computer Interfaces
Young Adult
Dimensionality Reduction

@article {pmid41146476,
year = {2025},
author = {Yue, X and Lu, L and Liu, H and Zang, Y},
title = {LRR-UNet: A Deep Unfolding Network With Low-Rank Recovery for EEG Signal Denoising.},
journal = {CNS neuroscience & therapeutics},
volume = {31},
number = {10},
pages = {e70632},
doi = {10.1111/cns.70632},
pmid = {41146476},
issn = {1755-5949},
support = {2023YFF1204200//National Key Research and Development Program of China/ ; 62476197//National Natural Science Foundation of China/ ; },
mesh = {*Electroencephalography/methods ; Humans ; *Deep Learning ; *Neural Networks, Computer ; *Signal Processing, Computer-Assisted ; Signal-To-Noise Ratio ; Artifacts ; Algorithms ; *Brain/physiology ; },
abstract = {BACKGROUND: Electroencephalogram (EEG) signals are crucial for brain-computer interface research but are highly susceptible to noise contamination, necessitating effective denoising. While deep learning has been widely applied, its "black-box" nature limits interpretability. In contrast, traditional model-based methods like Low-Rank Recovery (LRR) offer strong interpretability by decomposing signals into low-rank and sparse components.

OBJECTIVE: This paper aims to develop an interpretable deep-learning model for EEG denoising that combines the performance of deep learning with the interpretability of traditional LRR methods.

METHODS: We propose LRR-Unet, a deep unfolding network that transforms the traditional iterative LRR algorithm into a neural network architecture. Specifically, the time-consuming Singular Value Decomposition (SVD) and sparse optimization processes in LRR are replaced with learnable neural network modules.

RESULTS: Extensive experiments demonstrate that LRR-Unet outperforms other state-of-the-art models in removing ocular and electromyographic artifacts, achieving superior performance on both quantitative and qualitative metrics. Furthermore, in downstream classification tasks, EEG signals preprocessed with LRR-Unet yield better results across various evaluation indicators.

CONCLUSION: The proposed LRR-Unet provides an effective and interpretable solution for EEG denoising. Its superiority in denoising performance and practical utility in enhancing downstream application performance is validated through comprehensive experiments.},
}

*Electroencephalography/methods
Humans
*Deep Learning
*Neural Networks, Computer
*Signal Processing, Computer-Assisted
Signal-To-Noise Ratio
Artifacts
Algorithms
*Brain/physiology

@article {pmid41146424,
year = {2025},
author = {Yang, C and Wang, X and Ye, X and Shen, Y and Tong, J and Zhang, X and Zhou, Y},
title = {Spatiotemporal Immune Dynamics in Experimental Retinal Ganglion Cell Injury Models.},
journal = {Immunity, inflammation and disease},
volume = {13},
number = {10},
pages = {e70284},
doi = {10.1002/iid3.70284},
pmid = {41146424},
issn = {2050-4527},
support = {//This study was supported by a National Natural Science Foundation of China (NSFC) Key Program grant 82430038, Key R&D Program of Zhejiang Province 2025C02109, NSFC grants 82371455 and 82371084, a National Key Research and Development Program of China grant 2023YFC2506200, a China Postdoctoral Science Foundation grant 2024M752831, and the Open Project Program of Shaanxi Provincial Key Laboratory of Biological Psychiatry XJJSHTS-2504./ ; },
mesh = {*Retinal Ganglion Cells/immunology/pathology ; Animals ; *Optic Nerve Injuries/immunology/pathology ; Disease Models, Animal ; *Glaucoma/immunology/pathology ; Humans ; },
abstract = {BACKGROUND: The damage and regeneration of retinal ganglion cells (RGCs) have been extensively studied. Among them, immune cells in different parts of the visual pathway play an important role in injury, regeneration and repair, but a comprehensive analysis of their spatial and temporal distribution is lacking.

PURPOSE: This review emphasizes the unique characteristics of immune cells within the visual input pathway, focusing on their spatiotemporal dynamics in the retina, optic nerve head (ONH), and optic nerve during glaucoma and traumatic optic nerve injury.

METHODS: A comprehensive search was conducted across PubMed and Web of Science up to April 2025. Studies were included if they reported immune cells under glaucoma or optic nerve crush (ONC) animal models.

FINDINGS: Each region of the visual input pathway displays a distinct immune cell composition, including Müller cells, microglia, astrocytes, T cells, and oligodendrocytes, all of which work together to maintain homeostasis and respond to injury. Some immune cells are specific to certain regions, while others are shared across areas. Furthermore, even within a single glial cell type, there are different subtypes with unique developmental origins or marker profiles, reflecting a range of functions. In both glaucoma and traumatic optic nerve injury, retinal immune cells are rapidly activated, regardless of whether the initial impairment occurs in the soma or axon of RGCs, in the subacute or chronic course. The early stages of injury also see the presence of adaptive immune cells, such as T cells and neutrophils. Macrophages and microglia typically play complementary roles, while astrocytes show prolonged activation compared to microglia in the optic nerve, though this pattern does not hold in the retina following ONC.

CONCLUSIONS: Understanding the spatiotemporal dynamics of these immune responses in glaucoma and traumatic optic nerve injury is crucial for developing targeted therapies that can reduce RGC loss, mitigate neurotoxicity, and promote functional recovery, ultimately preventing vision impairment. Targeting specific immune cell subsets may provide new strategies for delaying RGC damage.},
}

*Retinal Ganglion Cells/immunology/pathology
Animals
*Optic Nerve Injuries/immunology/pathology
Disease Models, Animal
*Glaucoma/immunology/pathology
Humans

@article {pmid41146245,
year = {2025},
author = {López-Larraz, E and Sarasola-Sanz, A and Birbaumer, N and Ramos-Murguialday, A},
title = {Uncovering attempted movements of the paralyzed upper limb after stroke through EEG and EMG.},
journal = {Journal of neuroengineering and rehabilitation},
volume = {22},
number = {1},
pages = {221},
pmid = {41146245},
issn = {1743-0003},
support = {2422-0-1//Fortüne program - University of Tübingen/ ; E! 113928 SubliminalHomeRehab//Eureka-Eurostars/ ; MAIA 951910//European Union H2020-FETPROACT-EIC-2018-2020/ ; NanoNeuro//Basque IKUR initiative/ ; 01QE2023//Bundesministerium für Bildung und Forschung/ ; },
mesh = {Humans ; *Electroencephalography/methods ; Male ; *Stroke/physiopathology/complications ; Female ; *Electromyography/methods ; Middle Aged ; *Upper Extremity/physiopathology ; Aged ; *Paralysis/physiopathology/etiology/rehabilitation ; Adult ; Movement/physiology ; *Stroke Rehabilitation ; },
abstract = {Detecting attempted movements of a paralyzed limb is a key step for neural interfaces for motor rehabilitation and restoration after a stroke. In this paper, we present a systematic evaluation of electroencephalographic (EEG) and electromyographic (EMG) activity to decode when stroke patients with severe upper-limb paralysis attempt to move their affected arm. EEG and EMG recordings of 35 chronic stroke patients were analyzed. We trained classifiers to discriminate between rest and movement attempt states relying on brain, muscle, or both types of signals combined. Our results reveal that: (i) EEG and residual EMG activity provide complementary information to detect attempted movements, obtaining significantly higher decoding accuracy when both sources of activity are combined; (ii) EMG-based, but not EEG-based, decoding accuracy correlates with the degree of impairment of the patient; and (iii) the percentage of patients that achieve decoding accuracy above the chance level strongly depends on the type of features considered, and can be as low as 50% of them if only ipsilesional EEG is used. These results offer new perspectives to develop improved neurotechnologies that establish a more accurate contingent link between the central and peripheral nervous system after a stroke, leveraging Hebbian learning and facilitating functional plasticity and recovery.},
}

Humans
*Electroencephalography/methods
Male
*Stroke/physiopathology/complications
Female
*Electromyography/methods
Middle Aged
*Upper Extremity/physiopathology
Aged
*Paralysis/physiopathology/etiology/rehabilitation
Adult
Movement/physiology
*Stroke Rehabilitation

@article {pmid41145802,
year = {2025},
author = {Hazrati, H and Daliri, MR},
title = {Decoding covert visual attention of electroencephalography signals using continuous wavelet transform and deep learning approach.},
journal = {Scientific reports},
volume = {15},
number = {1},
pages = {37503},
pmid = {41145802},
issn = {2045-2322},
mesh = {Humans ; *Deep Learning ; *Electroencephalography/methods ; *Attention/physiology ; *Wavelet Analysis ; Male ; Adult ; Female ; Young Adult ; Brain-Computer Interfaces ; *Visual Perception/physiology ; Signal Processing, Computer-Assisted ; },
abstract = {Covert visual attention decoding from EEG signals is a key challenge in cognitive neuroscience and brain-computer interface applications. Traditional approaches often rely on manual feature extraction and handcrafted pipelines, which limit scalability and generalization. In this study, we propose a deep learning-based framework that leverages time-frequency representations, specifically Continuous Wavelet Transform (CWT), to enable end-to-end classification of covert attention states without manual feature engineering. EEG data were recorded from ten healthy participants performing spatial and feature-based attention tasks. Among the tested models, ShallowConvNet achieved 100% accuracy in binary classification and over 90% in four-class conditions. EEGNet also performed competitively, exceeding 97% and 88% accuracy in two- and four-class scenarios, respectively. These findings demonstrate that integrating CWT with deep neural architectures significantly enhances decoding performance compared to conventional raw-signal approaches, offering a scalable and efficient solution for real-time attention monitoring.},
}

Humans
*Deep Learning
*Electroencephalography/methods
*Attention/physiology
*Wavelet Analysis
Male
Adult
Female
Young Adult
Brain-Computer Interfaces
*Visual Perception/physiology
Signal Processing, Computer-Assisted

@article {pmid41145516,
year = {2025},
author = {Wei, Z and Lin, X and Zhang, L and Guo, L and Liu, J and Hu, L and Liu, Y and Kong, Y},
title = {CoSpine open access simultaneous cortico-spinal fMRI database of thermal pain and motor tasks.},
journal = {Scientific data},
volume = {12},
number = {1},
pages = {1696},
pmid = {41145516},
issn = {2052-4463},
support = {82072010//National Natural Science Foundation of China (National Science Foundation of China)/ ; IS23108//Natural Science Foundation of Beijing Municipality (Beijing Natural Science Foundation)/ ; },
mesh = {*Magnetic Resonance Imaging ; Humans ; *Pain/physiopathology/diagnostic imaging ; Brain/diagnostic imaging ; *Spinal Cord/diagnostic imaging ; },
abstract = {Simultaneous cortico-spinal functional magnetic resonance imaging (fMRI) enables non-invasive investigation of integrated central nervous system function, but acquisition challenges have restricted the availability of public datasets and slowed the development of advanced analytic methods. Here, we introduce the CoSpine database, the first open-access, BIDS-compliant cortico-spinal task-based fMRI resource (N = 61), acquired using a novel single-field-of-view (FOV) imaging protocol covering the whole brain (including cortical, subcortical, brainstem, and cerebellar regions) and cervical spinal cord. The dataset contains raw images, field maps, physiological recordings, and BIDS event files from thermal pain and voluntary motor tasks. An optimized acquisition and preprocessing framework is provided, validated by quality-control metrics such as temporal signal-to-noise ratio and alignment precision. Spanning a broad age range and standardized paradigms, CoSpine serves as a reference for neuroimaging methods development (e.g., hyperalignment) and for artificial intelligence (AI) model benchmarking. Potential applications include sensorimotor phenotyping, studies of age-related neurodegeneration, and exploratory work in neurorehabilitation, while also supporting early-stage development of brain-computer interface (BCI) systems involving spinal activity and personalized neuromodulation strategies.},
}

*Magnetic Resonance Imaging
Humans
*Pain/physiopathology/diagnostic imaging
Brain/diagnostic imaging
*Spinal Cord/diagnostic imaging

@article {pmid41144819,
year = {2025},
author = {Liu, H and Cao, X and Li, J and Zheng, L and Li, J and Li, Q and Xie, M and Li, H and Wang, X and Wu, Y and Zhang, X and Wang, Y and Gao, X and Sheng, T and Du, N and Xu, C and Zhou, K and Xu, J and Yan, C and Liu, L and Gao, L and Li, X and Zhang, M},
title = {Deciphering Neural Mechanisms Underlying Marmoset Dynamic Natural Behaviors Using a Miniaturized Wireless Large-Scale Coverage Neural Recorder.},
journal = {Advanced science (Weinheim, Baden-Wurttemberg, Germany)},
volume = {},
number = {},
pages = {e07110},
doi = {10.1002/advs.202507110},
pmid = {41144819},
issn = {2198-3844},
support = {2023YFB4705500//National Key Research and Development Program of China/ ; 62350710211//National Natural Science Foundation of China/ ; },
abstract = {Deciphering neural mechanisms underlying dynamic natural behaviors of freely moving species requires long-term recordings of large-scale brain activities. However, most conventional neural recorders are limited by their weights and measures, electrode coverage, and signal throughput, hindering the dissection of underlying neural mechanisms. This study reports real-time large-scale recordings and deciphering of brain activities from frontal and temporal cortices of freely moving marmoset across various natural behavioral repertoire using a miniaturized wireless neural recorder comprising a custom-designed 120-channel flexible µECoG array. Behavior-specific highly resolved spatiotemporal neural dynamics are observed, including alpha-band activations during drinking, anticipatory responses before vocalization, and transient high-gamma increase during vigilance to human intruders. Three phases of drinking behavior are identified using multi-area neural features captured by the recorder with an accuracy exceeding 87%. After over 16 months (March 13, 2024-August 1, 2025, remaining actively recording) of recordings, the neural signals acquired using the recorder maintain high fidelity and low attenuation during both the resting and drinking states, enabling potential long-term dissection of the neural mechanisms of natural behaviors in freely moving marmosets.},
}

@article {pmid41144414,
year = {2025},
author = {Lu, B and Chen, J and Wang, F and Wen, G and Fu, R and Hua, C},
title = {Causality-Driven Convolutional Manifold Attention Network for Electroencephalogram Signal Decoding.},
journal = {IEEE transactions on pattern analysis and machine intelligence},
volume = {PP},
number = {},
pages = {},
doi = {10.1109/TPAMI.2025.3625631},
pmid = {41144414},
issn = {1939-3539},
abstract = {Deep learning-based methods have achieved remarkable success in brain-computer interfaces (BCIs). However, its inherent assumption of independent and identically distributed (i.i.d.) data renders it vulnerable to out-of distribution (OOD) scenarios. To address this limitation, the present study proposed a causality-driven convolutional manifold attention network (CD-CMAN) that learned invariant representations from electroencephalogram (EEG) signals to enhance OOD generalization. The framework began with a spatiotemporal convolution module to extract rich temporal and spatial features. Guided by the defined structural causal model and leveraging the strengths of Riemannian geometry and deep learning, dual latent encoders with manifold attention units were crafted to explicitly separate spatiotemporal feature maps into semantic and variation latent factors. A reconstruction module with a dedicated loss was implemented to ensure these factors retaining informative, while the Hilbert-Schmidt independence criterion (HSIC) was introduced to enforce their statistical independence. Further, a variational information bottleneck and gradient reversal layer were incorporated to compress and disentangle the semantic and variation factors. Evaluations on two public datasets under both subject-dependent and subject independent settings demonstrated that CD-CMAN consistently outperforms comparative baselines. These findings suggest that the proposed model could provide a new solution for the practical application of BCI technology.},
}

@article {pmid41143908,
year = {2025},
author = {Xie, X and Hu, F and Yuan, S and Wen, D and Duan, D},
title = {MS-CANet: lightweight multi-scale channel attention network with depthwise residual blocks for EEG-based spatial cognition evaluation.},
journal = {Medical & biological engineering & computing},
volume = {},
number = {},
pages = {},
pmid = {41143908},
issn = {1741-0444},
support = {62276022//National Natural Science Foundation of China/ ; 2023YFF1203702//National Key Research and Development Program of China/ ; 62206014//National Key Research and Development Program of China/ ; FRFBD-25-052//Fundamental Research Funds for the Central Universities/ ; },
abstract = {Objective assessment of spatial cognitive ability is crucial for screening cognitive impairment and in neurorehabilitation. While deep learning methods for electroencephalogram (EEG) analysis show great promise, they increasingly rely on complex, parameter-heavy architectures. This complexity often leads to poor generalization due to overfitting on small datasets and hinders deployment on mobile healthcare devices. To overcome these limitations, we propose a novel lightweight multi-scale channel attention network with depthwise residual blocks. The model incorporates multi-scale convolutional layers to capture diverse temporal and spatial patterns in EEG signals. It then leverages channel attention mechanisms to dynamically prioritize informative channels, focusing on task-critical features. Furthermore, a novel depthwise separable residual block is introduced to significantly reduce computational complexity and maintain stable model performance. Evaluations on a spatial cognition EEG dataset demonstrate that our network achieves higher accuracy than baselines with only 8.453M parameters, making it an efficient and practical solution for mobile deployment. It also holds strong potential for extension to early screening and intervention in a wider range of cognitive disorders.},
}

@article {pmid41142108,
year = {2025},
author = {Peng, Q and Huang, J and Li, C and Jiang, M and Huang, C and Luo, J and Li, H and Yin, T and Cai, M and Fu, S and Ma, G and Liu, Z and Xu, T},
title = {Magnetically Actuated Soft Electrodes for Multisite Bioelectrical Monitoring of Ex Vivo Tissues.},
journal = {Cyborg and bionic systems (Washington, D.C.)},
volume = {6},
number = {},
pages = {0434},
pmid = {41142108},
issn = {2692-7632},
abstract = {Multisite electrophysiological monitoring of ex vivo tissues and organ models is essential for basic research and drug toxicity evaluation. However, conventional microelectrode arrays with fixed positions and rigid structures are insufficient for dynamic, curved tissue surfaces. Here, we present a magnetically actuated soft electrode (MSE) with precise navigation, adaptive attachment, and high-fidelity signal acquisition. Operating in a "locate-adhere-record-detach" cycle, the MSE enabled continuous multisite detection on beating ex vivo tissues. In isolated rat heart experiments, the MSE demonstrated millimeter-level navigation accuracy, stable contact, and high signal-to-noise ratio (average 28 dB). By integrating magnetic locomotion with electrophysiological sensing, this work establishes a programmable, actively addressable platform for multisite electrophysiological monitoring of organ models, tissue slices, and engineered constructs, offering broad potential for cardiotoxicity screening and cardiovascular research.},
}

@article {pmid41141194,
year = {2025},
author = {Jayalaksshme Srinivasan, K and Periasamy, P and Gunasekaran, S},
title = {Motor Imagery and Motor Execution: A Narrative Review of Electroencephalographic (EEG) Signatures, Methodological Consistency, and Translational Applications.},
journal = {Cureus},
volume = {17},
number = {9},
pages = {e93011},
pmid = {41141194},
issn = {2168-8184},
abstract = {This narrative review evaluates when electroencephalography (EEG) signatures elicited by kinesthetic motor imagery (MI) genuinely approximate those of motor execution (ME), appraises methodological consistency across studies, and outlines pragmatic routes to translation in brain-computer interfaces (BCIs) and neurorehabilitation. A keyword-driven search of Web of Science, Scopus, PubMed, and conference repositories was used to extract empirical, English-language EEG studies reporting sensorimotor rhythm (mu 8-13 Hz; beta 13-30 Hz) event-related desynchronization/synchronization (ERD/ERS) metrics and/or decoding performance for MI and/or ME, with structured extraction of task/sample features, imagery protocol, EEG methods/signatures, MI-ME overlap, translational readouts, and limitations. Across convergent datasets, MI reliably evokes contralateral mu/beta ERD with timing and topography akin to ME, typically with smaller amplitudes and broader fields; realistic decoding benchmarks cluster around the mid-70% for MI versus low-80% for ME, with ≈70% a usability threshold and 15%-30% of naïve users below it. Convergence and performance improve with first-person kinesthetic instructions, higher imagery vividness, synchronised action observation, object-oriented tasks, EMG monitoring, and contingent neurofeedback; source-space modelling and synergy-aware features can lift MI accuracy into the ~82%-95% range in constrained settings, though offline gains often overestimate online control. In stroke cohorts, most patients exhibit clear ERD/ERS, and a meaningful subset exceeds operational thresholds; however, calibration-to-online drops (e.g., ~80% to ~70%) are common and partially recover with adaptive retraining. The principal barriers to translation are heterogeneous protocols (band definitions, referencing, validation), small and selective samples, sparse EMG to exclude covert movement, non-stationarity across sessions, and persistent non-responders. To move from plausibility to practice, future studies should standardise mu/beta windows and baselines, report closed-loop outcomes, personalise training with vividness assessment and synchronised action observation, anticipate drift with adaptive algorithms and periodic recalibration, and integrate MI with robotics, functional electrical stimulation, or virtual reality in multisite trials that track durable functional gains.},
}

@article {pmid41140580,
year = {2025},
author = {Boonstra, JT},
title = {Ethical imperatives in the commercialization of brain-computer interfaces.},
journal = {IBRO neuroscience reports},
volume = {19},
number = {},
pages = {718-724},
pmid = {41140580},
issn = {2667-2421},
abstract = {The rapid commercialization of brain-computer interfaces (BCIs) raises urgent ethical and scientific challenges for human research oversight. While BCIs hold transformative potential for treating neurological disorders, their premature translation into consumer markets risks outpacing neuroscientific understanding and ethical frameworks. This essay critically examines the mismatch between commercial claims and the technical limitations of current BCI systems, decoding accuracy and biocompatibility, unresolved ethical dilemmas posed by neural data commodification and procedural risks, and the inadequacy of existing governance to address vulnerabilities in consent, privacy, and long-term safety. Responsible innovation demands proactive measures and robust public engagement to align development with societal values. Without such safeguards, the rush to commercialize BCIs risks prioritizing market interests over patient welfare and eroding public trust in neurotechnology.},
}

@article {pmid41139722,
year = {2025},
author = {Li, J and Chen, T and Yan, X and Luo, L},
title = {The effect of device-based neuromodulation on the motor recovery of patients with spinal cord injury.},
journal = {Spinal cord},
volume = {},
number = {},
pages = {},
pmid = {41139722},
issn = {1476-5624},
abstract = {STUDY DESIGN: This paper systematically analyzes literature from PubMed, MEDLINE, Embase, and Cochrane databases over the past 10 years (up to May 25, 2025). It employs defined search terms, inclusion/exclusion criteria, and a documented search flow to evaluate mechanisms, efficacy, challenges, and future directions of neuromodulation technologies for spinal cord injury rehabilitation. The results synthesize findings from clinical trials, and representative papers.

OBJECTIVE: This review aims to evaluate the mechanisms and clinical applications of device-based neuromodulation technologies in spinal cord injury (SCI) rehabilitation, focusing on their efficacy, challenges, and future directions.

SETTING: The countries and regions worldwide participating in neuromodulation.

METHODS: We systematically analyzed advancements in neuromodulation over the past decade, including brain-spinal interfaces (BSI), brain-computer interfaces (BCI), cranial stimulation techniques (DBS, TMS, tDCS), spinal cord stimulation (SCS), robotic exoskeletons. The review integrates findings from clinical trials.

RESULTS: Neuromodulation technologies demonstrate significant potential in restoring motor and sensory function post-SCI. BSI and BCI improve mobility but face infection and cybersecurity risks. Cranial stimulation techniques enhance neuroplasticity, with DBS and TMS showing efficacy, while tDCS requires further validation. Epidural SCS enables motor recovery in complete paralysis but has high infection rates. Robotic exoskeletons benefit younger patients.

CONCLUSION: Neuromodulation technologies represent promising interventions for SCI, yet challenges remain in precision, safety, and efficacy. Future research should prioritize AI-driven parameter optimization, wearable device development, and multicenter randomized trials to establish these methods as core treatments, ultimately improving patient outcomes and quality of life.},
}

@article {pmid41138930,
year = {2025},
author = {Wang, J and Wang, X and Qiao, S and La, H and Yu, Y and An, X},
title = {Investigation of Fatigue Mechanisms and Detection Methods for Anesthesiologists Based on Multimodal Physiological Signals.},
journal = {Brain research bulletin},
volume = {},
number = {},
pages = {111597},
doi = {10.1016/j.brainresbull.2025.111597},
pmid = {41138930},
issn = {1873-2747},
abstract = {Anesthesiologists are highly susceptible to fatigue due to the demanding intensity and critical responsibility of their work, which poses substantial risks to both clinician health and patient safety. To elucidate fatigue mechanisms, this study systematically assessed cognitive and physiological alterations before and after prolonged high-intensity work. Cognitive performance was evaluated with paradigms targeting attention (0-back), working memory (2-back), and visuospatial processing, complemented by multimodal physiological monitoring with electroencephalogram (EEG) and electrocardiogram (ECG) recordings. Prolonged work was associated with significant declines in n-back accuracy, reflecting impaired attention and working memory, while visuospatial performance showed marked increases in both error rate and reaction time, indicating deterioration of spatial cognition and executive control. Concurrently, physiological analyses revealed enhanced EEG alpha-band connectivity, shortened RR intervals, a reduced LF/HF ratio, and elevated multiscale entropy, collectively indicating autonomic imbalance and central-autonomic dysregulation under fatigue. Building on these mechanistic findings, we applied transfer learning algorithms to statistically significant multimodal physiological features, achieving 99.4% cross-subject classification accuracy. This integration of mechanistic insights with computational modeling underscores the reliability of the proposed strategy and its translational potential for real-world clinical fatigue monitoring.},
}

@article {pmid41137585,
year = {2025},
author = {Zadeh Makouei, ST and Uyulan, C and Erguzel, TT and Tarhan, N},
title = {Advanced Facial Expression Recognition Using Model Averaging Ensembles of Convolutional Neural Networks and CAM Analysis.},
journal = {Clinical EEG and neuroscience},
volume = {},
number = {},
pages = {15500594251366792},
doi = {10.1177/15500594251366792},
pmid = {41137585},
issn = {2169-5202},
abstract = {Facial expressions play a vital role in non-verbal communication, conveying a wide range of emotions and messages. Although prior research achieved notable advances through architecture design or dataset-specific optimization, few studies have integrated multiple advanced techniques into a unified facial expression recognition (FER) pipeline. Addressing this gap, we propose a comprehensive approach that combines (i) multiple pre-trained CNNs, (ii) MTCNN-based face detection for improved facial region localization, and (iii) Grad-CAM-based interpretability. While MTCNN enhances the quality of face localization, it may slightly affect classification accuracy by focusing on cleaner yet more challenging samples. We evaluate four pre-trained models - DenseNet121, ResNet-50, ResNet18, and MobileNetV2 - on two datasets: Raf-DB and Cleaned-FER2013. The proposed pipeline demonstrates consistent improvements in interpretability and overall system robustness. The results emphasize the strength of integrating face detection, transfer learning, and interpretability techniques within a single framework can significantly enhance the transparency and reliability of FER systems. Combining FER with EEG-based systems significantly enhances the emotional intelligence of brain-computer interfaces, enabling more adaptive and personalized user experiences. With this approach the paper bridges the gap between affective computing and cognitive neuroscience, aligning closely EEG-centered interaction methodologies. Besides understanding the relationship between facial expressions of emotions and EEG signals will be an important study for literature.},
}

@article {pmid41136747,
year = {2025},
author = {Zuo, H and Zhang, W and Wang, L and Wu, Y and Zheng, Y and Hao, S and Chen, QY and Cao, P and Ouyang, M and Huang, S and Zhou, W and Xue, YX and Pan, Y and Wei, W and Zhuo, M and Yuan, T and Zha, R and Zhang, Z and Zhang, X},
title = {Transcranial direct current stimulation restores addictive behavior via prefrontal-striatal circuit.},
journal = {Molecular psychiatry},
volume = {},
number = {},
pages = {},
pmid = {41136747},
issn = {1476-5578},
support = {32171080//National Natural Science Foundation of China (National Science Foundation of China)/ ; 32400919//National Natural Science Foundation of China (National Science Foundation of China)/ ; 32200914//National Natural Science Foundation of China (National Science Foundation of China)/ ; },
abstract = {Dependence on methamphetamine (METH) is a severe brain disorder characterized by high relapse rates and cognitive decline following detoxification. Recent research suggests that transcranial direct current stimulation (tDCS) may treat addiction, but the underlying neural mechanisms remain unknown. Here, we employed METH-conditioned place preference (CPP) paradigm integrated with fMRI, electrophysiology, chemogenetics, in vivo fiber photometry recordings and a novel rodent tDCS model to examine the neural circuit underlying tDCS modulation on METH-induced addictive behavior. We demonstrated that tDCS targeted at the medial prefrontal cortex (mPFC) prevents relapse. Specifically, tDCS enhanced the activity of neurons in both the infralimbic cortex (IL) and the nucleus accumbens shell (NAcSh) simultaneously. Furthermore, chemogenetic inhibition of the IL-NAcSh circuit eliminated the modulatory effects of tDCS, while activation of the IL-NAcSh circuit was sufficient to suppress the relapse. These findings reveal that the IL-NAcSh pathway functions as a descending regulatory circuit mediating the therapeutic outcomes of tDCS in the treatment of substance use disorder, offering new insights into circuit-based neuro-modulatory treatments for addiction.},
}

@article {pmid41135743,
year = {2025},
author = {Wang, Y and Chen, HJ and Cheng, Y and Xie, Y and Cheng, Y and Zhao, S and Jiang, Y and Bai, T and Huo, Y and Wang, K and Zhang, M and Huang, W and Feng, G and Han, Y and Shu, N},
title = {Multimodal Integration of Plasma Biomarkers, MRI, and Genetic Risk to Predict Cerebral Amyloid Burden in Alzheimer's Disease.},
journal = {NeuroImage},
volume = {},
number = {},
pages = {121550},
doi = {10.1016/j.neuroimage.2025.121550},
pmid = {41135743},
issn = {1095-9572},
abstract = {Alzheimer's disease (AD), the most prevalent neurodegenerative disorder, is marked by the accumulation of amyloid-β (Aβ) plaques. Although cerebral Aβ positron emission tomography (Aβ-PET) remains the gold standard for assessing cerebral Aβ burden, its clinical utility is hindered by cost, radiation exposure, and limited availability. Plasma biomarkers have emerged as promising, non‑invasive indicators of Aβ pathology, yet they do not incorporate individual genetic risk or neuroanatomical context. To address this gap, we developed a multimodal machine‑learning framework that integrates plasma biomarkers, MRI‑derived brain structural features (regional volumes, cortical thickness, cortical area and structural connectivity), and genetic risk profiles to predict cerebral Aβ burden. This approach was evaluated in 150 participants from the Alzheimer's Disease Neuroimaging Initiative (ADNI) and 101 participants from a domestic Chinese Sino Longitudinal Study of Cognitive Decline (SILCODE). Incorporating multimodal features substantially improved predictive performance: the baseline model using plasma and clinical variables alone achieved an R[2] of 0.56, whereas integrating neuroimaging and genetic information increased accuracy (R[2] = 0.63 with apolipoprotein E genotypes and R[2] = 0.64 with polygenic risk scores). Furthermore, a multiclass classifier trained on the same multimodal features achieved robust discrimination of cognitive status, with area‑under‑the‑curve values of 0.87 for normal controls, 0.76 for mild cognitive impairment, and 0.95 for AD dementia. These findings highlight the value of combining plasma, imaging, and genetic data to non-invasively estimate cerebral Aβ burden, offering a potential alternative to PET imaging for early AD risk assessment.},
}

@article {pmid41135661,
year = {2025},
author = {Pan, Y and Yang, X and Wu, M and Hu, S},
title = {Latent profile analysis of childhood trauma in Chinese individuals with bipolar disorder: Differential associations with suicidality and clinical symptomatology.},
journal = {Journal of affective disorders},
volume = {394},
number = {Pt A},
pages = {120490},
doi = {10.1016/j.jad.2025.120490},
pmid = {41135661},
issn = {1573-2517},
abstract = {BACKGROUND: Childhood trauma is a well-established risk factor for poor clinical outcomes in bipolar disorder (BD), yet most studies have relied on cumulative trauma scores, potentially overlooking heterogeneity in trauma exposure and its differential impact on psychopathology.

METHODS: This study employed latent profile analysis (LPA) to identify distinct subtypes of childhood trauma based on the Childhood Trauma Questionnaire (CTQ) among 725 individuals with BD in a Chinese clinical sample. Differences across trauma profiles were examined in relation to demographic features, psychiatric symptoms (anxiety, depression, mania), and suicidal ideation (Beck Scale for Suicide Ideation, BSSI).

RESULTS: A four-class solution was identified, and the relationship with mental health outcomes was analyzed. Class 4 group, characterized by the most severe emotional abuse and physical neglect, along with the lowest emotional neglect, reported the highest levels of anxiety (HAMA), depression (HAMD), and suicidal ideation (BSSI). In contrast, manic symptoms (YMRS) were present across all groups but did not differ significantly between trauma profiles. Logistic regression indicated that emotional abuse was the strongest predictor of trauma class membership.

CONCLUSIONS: Distinct trauma profiles in BD are differentially associated with symptom severity and suicide risk. These findings highlight the clinical value of moving beyond cumulative trauma scores to identify trauma-specific subtypes. Early identification of high-risk trauma configurations may inform personalized assessment and intervention strategies for individuals with BD.},
}

@article {pmid41135523,
year = {2025},
author = {Wang, Z and Tang, Q and Li, K and Mou, J and Chen, Y and Kuang, W and Sun, L and Ma, Z and Wei, Y and Bao, R and Sun, X and Wang, S and Lu, W and Xu, GY and Tang, YQ and Duan, S and Ni, JD},
title = {An enteric-DRG pathway for interoception and visceral pain in mice.},
journal = {Neuron},
volume = {},
number = {},
pages = {},
doi = {10.1016/j.neuron.2025.09.035},
pmid = {41135523},
issn = {1097-4199},
abstract = {Sensory afferents are major interoceptive pathways for organ-brain communication. Within the distal colon, dorsal root ganglia (DRGs) afferents regulate key gut physiology. Inflammation causes hypersensitivity of DRG pathways, leading to visceral pain. However, whether enteric neurons contribute to interoception and visceral pain remains unclear. Here, we surveyed the DRG innervation along the gastrointestinal tract in mice and found extensive associations between DRG terminals and enteric neurons. Optogenetic activation of different DRG terminals in the distal colon elicited variable degrees of behavioral responses, but only designated subpopulations induced aversion. Notably, optogenetic activation of colon cholinergic, but not nitrergic, enteric neurons signaled through the DRG-spinal pathway to evoke a non-aversive nociceptive-like reflex. Acetylcholine is part of the enteric-DRG signaling. Remarkably, inflammation shifted the nature of the enteric-DRG pathway from non-aversive to aversive. These findings expand the previous understanding of DRG-mediated visceral sensation, highlighting the contribution of enteric neuron-DRG communication to inflammation-induced visceral pain.},
}

@article {pmid41134961,
year = {2025},
author = {Jin, J and Qin, K and Allison, BZ and Li, S and Zhang, Y and Wang, X and Cichocki, A},
title = {A Transfer Learning SSVEP Decoding Algorithm Calibrated With Single-Trial Data.},
journal = {IEEE transactions on neural networks and learning systems},
volume = {PP},
number = {},
pages = {},
doi = {10.1109/TNNLS.2025.3617508},
pmid = {41134961},
issn = {2162-2388},
abstract = {Training-based algorithms significantly outperform training-free methods in terms of recognition performance for steady-state visual-evoked potential (SSVEP)-based brain-computer Interfaces (BCIs). However, collecting training data requires calibration experiments that are effort-intensive and often costly. These calibration demands limit the practicality of BCI, as users (and even system operators) may experience fatigue or lose interest in continued use. Transfer learning (TL) offers an effective solution, but it typically relies on either a certain amount of target domain data or extensive source domain data. To address this limitation, we introduce the concept of cross-dataset TL in SSVEP for the first time to extract transfer knowledge from other datasets. During this process, we identified a data mismatch problem that severely compromises the generalizability of transfer knowledge. To overcome this challenge, we propose a TL-SSVEP decoding algorithm calibrated with single-trial data (TL-CSTD). Specifically, we use 2 s of 8 Hz single-trial calibration data from the target domain to obtain matched transfer templates from the source domain. These templates are then corrected to extract holistic and single-period transfer knowledge, which are subsequently employed to construct an efficient TL-SSVEP decoding model for the target subject. Experimental results on three large SSVEP datasets demonstrate that TL-CSTD effectively addresses the data mismatch problem and achieves excellent SSVEP recognition performance using only 2 s of single-trial calibration data, showing its significant application potential and practicality.},
}

@article {pmid41134943,
year = {2025},
author = {Sun, J and Lin, PJ and Zhai, X and Wang, W and Jia, T and Li, Z and Pan, Y and Ji, L and Zhou, B and Li, C},
title = {Multimodal behavioral data predict stroke patient's response to BCI treatment through explainable AI.},
journal = {IEEE transactions on neural systems and rehabilitation engineering : a publication of the IEEE Engineering in Medicine and Biology Society},
volume = {PP},
number = {},
pages = {},
doi = {10.1109/TNSRE.2025.3625222},
pmid = {41134943},
issn = {1558-0210},
abstract = {Brain-computer interface (BCI)-based neurorehabilitation holds promise in enhancing motor recovery after stroke. However, recent research has reported heterogeneous results, indicating both responders and non-responders to BCI therapy. Using explainable artificial intelligence (XAI) methods, this study aims to investigate the independent and combined importance of multimodal behavioral data to predict patients' response to BCI therapy. Forty-two subacute stroke patients with lower-limb motor impairment underwent behavioral assessments, and received two-week BCI rehabilitation training. Linear regression, elastic net and artificial neural network models were developed to predict response to BCI therapy. Two XAI techniques, the stepwise method and Shapley additive explanation, were used to interpret model outcomes. The multivariate model (R[2]=0.852, P<0.001) that combines an optimal subset of multimodal behavioral data outperformed the univariate model (R[2]=0.758, P<0.001) trained on a single variable. Elastic net and artificial neural network models both demonstrated high prediction performance, as indicated by classification accuracies of 0.810 and 0.762, and areas under the receiver operating characteristic curve of 0.782 and 0.771. Our results revealed that multimodal behavioral data, including demographic, clinical, and biomechanical characteristics, provided unique and complementary information for interpreting the response of subacute patients to BCI therapy. Particularly, baseline motor impairment, muscle spasticity and balance function were primary predictors. Our findings highlight the core role of XAI methods towards precision medicine, which can help clinicians to identify individual recovery potentials and plan optimal treatment strategies.},
}

@article {pmid41132843,
year = {2025},
author = {Jin, S and Lin, C and Li, P and Wang, X and Wang, Y and Zhang, C and Wang, X and Peng, Y and Li, H and Lu, Y and Wang, X},
title = {Cannabidiol alleviates methamphetamine addiction via targeting ATP5A1 and modulating the ATP-ADO-A1R signaling pathway.},
journal = {Acta pharmaceutica Sinica. B},
volume = {15},
number = {10},
pages = {5261-5276},
pmid = {41132843},
issn = {2211-3835},
abstract = {Cannabidiol (CBD), a non-psychoactive cannabinoid, shows great promise in treating methamphetamine (METH) addiction. Nonetheless, the molecular target and the mechanism through which CBD treats METH addiction remain unexplored. Herein, CBD was shown to counteract METH-induced locomotor sensitization and conditioned place preference. Additionally, CBD mitigated the adverse effects of METH, such as cristae loss, a decline in ATP content, and a reduction in membrane potential. Employing an activity-based protein profiling approach, a target fishing strategy was used to uncover CBD's direct target. ATP5A1, a subunit of ATP synthase, was identified and validated as a CBD target. Moreover, CBD demonstrated the ability to ameliorate METH-induced ubiquitination of ATP5A1 via the D376 residue, thereby reversing the METH-induced reduction of ATP5A1 and promoting the assembly of ATP synthase. Pharmacological inhibition of the ATP efflux channel pannexin 1, blockade of ATP hydrolysis by a CD39 inhibitor, and blocking the adenosine A1 receptor (A1R) all attenuated the therapeutic benefits of CBD in mitigating METH-induced behavioral sensitization and CPP. Moreover, the RNA interference of ATP5A1 in the ventral tegmental area resulted in the reversal of CBD's therapeutic efficacy against METH addiction. Collectively, these data show that ATP5A1 is a target for CBD to inhibit METH-induced addiction behaviors through the ADO-A1R signaling pathway.},
}

@article {pmid41132728,
year = {2025},
author = {Berlet, R and Azapagic, A and Jha, NK and Aksenov, D and Bookwalter, J and Ullah, A and Bobustuc, G and Lee, J and Sant, H and McDaid, J and Walker, M and Shea, J and Graff, D and Barlow, AK and Frigerio, R and Aliee, D and Bailes, C and Gale, BK and Bailes, JE},
title = {An implantable, intracerebral osmotic pump for convection-enhanced drug delivery in glioblastoma multiforme.},
journal = {Frontiers in oncology},
volume = {15},
number = {},
pages = {1676691},
pmid = {41132728},
issn = {2234-943X},
abstract = {BACKGROUND: Glioblastoma multiforme (GBM; WHO Grade 4) is an aggressive brain tumor that invariably recurs after surgical resection, chemoradiation, and adjuvant chemotherapy. Treatment is limited, in part, because the blood-brain barrier (BBB) restricts entry of chemotherapeutic agents to the brain. Introducing drugs directly into the brain circumvents the BBB, but diffusion of these typically large drug molecules within brain parenchyma is limited. Convection-enhanced delivery (CED), based on the principles of bulk flow, can achieve drug distribution over a wider area to target residual cancer cells and thus remains a promising technique for treating GBM and other neuro-oncologic pathologies. Here, we propose a new method that combines direct brain delivery and CED using a fully implantable, microfluidic pump placed at the time of initial resection surgery.

METHODS: In this initial proof-of-concept study, we evaluated the function of a 3D-printed pump in an in vitro system and in vivo in a rat C6 glioma model.

RESULTS: In vitro osmosis-driven distribution of a high molecular-weight marker dye extended up to 18 mm from the pump with minimal reflux, including under simulations of increased intracranial pressure. In vivo, MRI imaging demonstrated wide distribution of superparamagnetic iron oxide particles from a pump implanted after the resection of a C6 glioma. Histological staining indicated that pump implantation did not cause additional inflammatory changes compared to controls.

CONCLUSION: This preliminary study demonstrated the feasibility of using an implantable, osmosis-driven pump to bypass the BBB and provide targeted delivery for treatment of GBM.},
}

@article {pmid41129590,
year = {2025},
author = {Cao, D and Yu, Z and Wang, J and Wu, Y},
title = {SMMTM: Motor imagery EEG decoding algorithm using a hybrid multi-branch separable convolutional self-attention temporal convolutional network.},
journal = {PloS one},
volume = {20},
number = {10},
pages = {e0333805},
pmid = {41129590},
issn = {1932-6203},
mesh = {Humans ; *Brain-Computer Interfaces ; *Electroencephalography/methods ; *Algorithms ; Neural Networks, Computer ; *Imagination/physiology ; },
abstract = {Motor imagery (MI) is a brain-computer interface (BCI) technology with the potential to change human life in the future. MI signals have been widely applied in various BCI applications, including neurorehabilitation, smart home control, and prosthetic control. However, the limited accuracy of MI signals decoding remains a significant barrier to the broader growth of the BCI applications. In this study, we propose the SMMTM model, which combines spatiotemporal convolution (SC), multi-branch separable convolution (MSC), multi-head self-attention (MSA), temporal convolution network (TCN), and multimodal feature fusion (MFF). Specifically, we use the SC module to capture both temporal and spatial features. We design a MSC to capture temporal features at multiple scales. In addition, MSA is designed to extract valuable global features with long-term dependence. The TCN is employed to capture higher-level temporal features. The MFF consists of feature fusion and decision fusion, using the features output from the SMMTM to improve robustness. The SMMTM was evaluated on the public benchmark BCI Comparison IV 2a and 2b datasets, the results showed that the within-subject classification accuracies for the datasets were 84.96% and 89.26% respectively, with kappa values of 0.797 and 0.756. The cross-subject classification accuracy for the 2a dataset was 69.21%, with a kappa value of 0.584. These results indicate that the SMMTM significantly enhances decoding performance, providing a strong foundation for advancing practical BCI implementations.},
}

Humans
*Brain-Computer Interfaces
*Electroencephalography/methods
*Algorithms
Neural Networks, Computer
*Imagination/physiology

@article {pmid41129446,
year = {2025},
author = {Dang, W and Ren, Z and Sun, J and Lv, D and Xiong, Z and Guo, W and Gao, Z and Yu, H},
title = {ML-TGNet: A Multi-Level Topology Guidance Network for Motor Imagery Decoding.},
journal = {IEEE journal of biomedical and health informatics},
volume = {PP},
number = {},
pages = {},
doi = {10.1109/JBHI.2025.3624298},
pmid = {41129446},
issn = {2168-2208},
abstract = {Brain-computer interfaces (BCIs) based on motor imagery electroencephalogram (MI-EEG) signals have been extensively applied in various neural rehabilitation scenarios. However, existing methods primarily focus on designing complex architectures to extract spatio-temporal features from MI-EEG signals, often neglecting the brain dynamics information embedded within them. This oversight leads to the extraction of redundant information, ultimately reducing decoding performance. To address these challenges, we design a multi-level topology-guidance network (ML-TGNet) that leverages topological brain synchronization information to more effectively extract features related to MI tasks. ML-TGNet specifically comprises a multi-level topology guidance module, a feature pool module, and a multi-branch decoding module. To evaluate its performance, extensive experiments are conducted on three publicly available MI datasets: the BCI Competition IV-2a dataset, the High Gamma dataset, and the OpenBMI dataset. ML-TGNet achieves classification accuracies of 82.33%, 96.42%, and 85.26% on these three datasets, respectively, outperforming current state-of-the-art models. These findings confirm the efficacy of using brain synchronization information to guide MI decoding, thereby opening a novel approach for EEG-based brain state decoding by integrating brain dynamics into deep learning.},
}

@article {pmid41129442,
year = {2025},
author = {Wang, Z and Wang, H and Jia, T and He, X and Li, S and Wu, D},
title = {DBConformer: Dual-Branch Convolutional Transformer for EEG Decoding.},
journal = {IEEE journal of biomedical and health informatics},
volume = {PP},
number = {},
pages = {},
doi = {10.1109/JBHI.2025.3622725},
pmid = {41129442},
issn = {2168-2208},
abstract = {Electroencephalography (EEG)-based brain computer interfaces (BCIs) transform spontaneous/evoked neural activity into control commands for external communication. While convolutional neural networks (CNNs) remain the mainstream backbone for EEG decoding, their inherently short receptive field makes it difficult to capture long-range temporal dependencies and global inter-channel relationships. Recent CNN-Transformer (Con former) hybrids partially address this issue, but most adopt a serial design, resulting in suboptimal integration of local and global features, and often overlook explicit channel-wise modeling. To address these limitations, we propose DBConformer, a dual-branch convolutionalTrans former network tailored for EEG decoding. It integrates a temporal Conformer to model long-range temporal dependencies and a spatial Conformer to extract inter-channel interactions, capturing both temporal dynamics and spatial patterns in EEG signals. A lightweight channel attention module further refines spatial representations by assigning data-driven importance to EEG channels. Extensive experiments under four evaluation settings on three paradigms, including motor imagery, seizure detection, and steady state visual evoked potential, demonstrated that DBCon former consistently outperformed 13 competitive baseline models, with over an eight-fold reduction in parameters than current high-capacity EEG Conformer architecture. Furthermore, the visualization results confirmed that the features extracted by DBConformer are physiologically in terpretable and aligned with prior knowledge. The superior performance and interpretability of DBConformer make it reliable for accurate, robust, and explainable EEG decoding. Code is publicized at https://github.com/wzwvv/ DBConformer.},
}