Brain-computer interfaces: The innovative key to unlocking neurological conditions

International Journal of Surgery, Год журнала: 2024, Номер 110(9), С. 5745 - 5762

Опубликована: Авг. 14, 2024

Neurological disorders such as Parkinson’s disease, stroke, and spinal cord injury can pose significant threats to human mortality, morbidity, functional independence. Brain–Computer Interface (BCI) technology, which facilitates direct communication between the brain external devices, emerges an innovative key unlocking neurological conditions, demonstrating promise in this context. This comprehensive review uniquely synthesizes latest advancements BCI research across multiple disorders, offering interdisciplinary perspective on both clinical applications emerging technologies. We explore progress its addressing various with a particular focus recent studies prospective developments. Initially, provides up-to-date overview of encompassing classification, operational principles, prevalent paradigms. It then critically examines specific movement consciousness, cognitive mental well sensory highlighting novel approaches their potential impact patient care. reveals trends applications, integration artificial intelligence development closed-loop systems, represent over previous The concludes by discussing prospects directions underscoring need for collaboration ethical considerations. emphasizes importance prioritizing bidirectional high-performance BCIs, areas that have been underexplored reviews. Additionally, we identify crucial gaps current research, particularly long-term efficacy standardized protocols. role neurosurgery spearheading translation is highlighted. Our analysis presents technology transformative approach diagnosing, treating, rehabilitating substantial enhance patients’ quality life advance field neurotechnology.

Язык: Английский

---

A review of Graph Neural Networks for Electroencephalography data analysis

Neurocomputing, Год журнала: 2023, Номер 562, С. 126901 - 126901

Опубликована: Окт. 7, 2023

Electroencephalography (EEG) sensors are flexible and non-invasive sensoring devices for the measurement of electrical brain activity which is extensively used in some areas clinical practice psychological/psychiatric research, such as epilepsy, sleep, emotion, computer interfaces. Although EEG sensor do not provide actual localizations sources, they allow to study functional connectivity. In this paper we review current application a specific family computational methods, Graph Neural Networks (GNN) analysis data. GNNs appear be well suited data modeling deal with signals whose domain defined by graph instead regular lattice Euclidean space. Readings electrodes fall category, hence increasing research on

Язык: Английский

---

BrainGridNet: A two-branch depthwise CNN for decoding EEG-based multi-class motor imagery

Neural Networks, Год журнала: 2023, Номер 170, С. 312 - 324

Опубликована: Ноя. 18, 2023

Язык: Английский

---

TBEEG: A Two-Branch Manifold Domain Enhanced Transformer Algorithm for Learning EEG Decoding

IEEE Transactions on Neural Systems and Rehabilitation Engineering, Год журнала: 2024, Номер 32, С. 1445 - 1455

Опубликована: Янв. 1, 2024

The electroencephalogram-based (EEG) brain-computer interface (BCI) has garnered significant attention in recent research. However, the practicality of EEG remains constrained by lack efficient decoding technology. challenge lies effectively translating intricate into meaningful, generalizable information. signal primarily relies on either time domain or frequency There lacks a method capable simultaneously and extracting both features, as well efficiently fuse these features. Addressing limitations, two-branch Manifold Domain enhanced transformer algorithm is designed to holistically capture EEG's spatio-temporal Our projects time-domain information signals Riemannian spaces fully decode dependence signals. Using wavelet transform, converted information, spatial contained mined through spectrogram. effectiveness proposed TBEEG validated BCIC-IV-2a dataset MAMEM-SSVEP-II datasets.

Язык: Английский

---

A Spiking Neural Network Approach for Classifying Hand Movement and Relaxation from EEG Signal using Time Domain Features

WSEAS TRANSACTIONS ON BIOLOGY AND BIOMEDICINE, Год журнала: 2025, Номер 22, С. 133 - 151

Опубликована: Янв. 21, 2025

High-performance prosthetic and exoskeleton systems based on EEG signals can improve the quality of life hand-impaired people. Effective controlling these assistive devices requires accurate signal classification. Although there have been advancements in Brain-Computer Interface (BCI) systems, still classifying with high accuracy is a great challenge. The objective this research to investigate classification Spiking Neural Network (SNN) classifier for factual exact control individuals hand impairment. dataset has taken from BNCI Horizon 2020 website, which movement-relax events patient spinal cord injury (SCI) operate neuro-prosthetic device attached paralyzed right upper limb. fusion Dispersion Entropy (DE), Fuzzy (FE), Fluctuation (FDE) mean skewness features are extracted Motor Imagery (MI) applied classifier. To compare performance algorithm, same used Convolutional (CNN), Random Forest (RF), Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Logistic Regression (LR) classifiers. It found that SNN given highest 80% precision 80.95%, recall 77.28%, F1-score 79.07%. This indicates five greater potential BCI system-based applications.

Язык: Английский

---

Graph convolution network-based eeg signal analysis: a review

Medical & Biological Engineering & Computing, Год журнала: 2025, Номер unknown

Опубликована: Янв. 30, 2025

Язык: Английский

---

DSTA-Net: Dynamic Spatio-Temporal Feature Augmentation Network for Motor Imagery Classification

Research Square (Research Square), Год журнала: 2025, Номер unknown

Опубликована: Март 18, 2025

Accurate decoding and strong feature interpretability of Motor Imagery (MI) are expected to drive MI applications in stroke rehabilitation. However, the inherent nonstationarity high intra-class variability MI-EEG pose significant challenges extracting reliable spatio-temporal features. We proposed Dynamic Spatio-Temporal Feature Augmentation Network (DSTA-Net), which combines DSTA Convolution (STC) modules. In module, multi-scale temporal convolutional kernels tailored α β frequency bands neurophysiological characteristics, while raw EEG serve as a baseline layer retain original information. Next, Grouped Spatial Convolutions extract multi-level spatial features, combined with weight constraints prevent overfitting. convolution map channel information into new domain, enabling further extraction through dimensional transformation. And STC module extracts features conducts classification. evaluated DSTA-Net on three public datasets applied it self-collected dataset. 10-fold cross-validation, achieved average accuracy improvements 6.29% (p<0.01), 3.05% 5.26%(p<0.01), 2.25% over ShallowConvNet BCI-IV-2a, OpenBMI, CASIA, dataset, respectively. hold-out validation, 3.99% (p<0.01) 4.2% OpenBMI CASIA datasets, Finally, we DeepLIFT, Common Pattern, t-SNE analyze contributions individual channels, patterns, visualize The superiority offers insights for research application MI. code is available https://github.com/CL-Cloud-BCI/DSTANet-code.

Язык: Английский

---

Electroencephalography-based biological and functional characteristics of spinal cord injury patients with neuropathic pain and numbness

Frontiers in Neuroscience, Год журнала: 2024, Номер 18

Опубликована: Май 1, 2024

Objectives To identify potential treatment targets for spinal cord injury (SCI)-related neuropathic pain (NP) by analysing the differences in electroencephalogram (EEG) and brain network connections among SCI patients with NP or numbness. Participants methods The EEG signals during rest, as well left- right-hand feet motor imagination (MI), were recorded. power spectral density (PSD) of θ (4–8 Hz), α (8–12 β (13–30 Hz) bands was calculated applying Continuous Wavelet Transform (CWT) Modified S-transform (MST) to data. We used 21 electrodes nodes performed statistical measurements phase synchronisation between two regions using a phase-locking value, which captures nonlinear synchronisation. Results specificity MST algorithm higher than that CWT. Widespread non-lateralised event-related synchronization observed both groups MI. PWP (patients pain) group had lower PSD values multiple channels including frontal, premotor, motor, temporal compared PWN numbness) (all p &lt; 0.05), but band parietal region 0.05). During left-hand MI, frequency (θ bands), significantly weaker except frontal region. Conversely, (β band), stronger all group. Conclusion connectivity are biological functional characteristics can be distinguish from suggest distinct mechanisms

Язык: Английский

---

Coherence based graph convolution network for motor imagery-induced EEG after spinal cord injury

Frontiers in Neuroscience, Год журнала: 2023, Номер 16

Опубликована: Янв. 13, 2023

Background Spinal cord injury (SCI) may lead to impaired motor function, autonomic nervous system dysfunction, and other dysfunctions. Brain-computer Interface (BCI) based on imagery (MI) can provide more scientific effective treatment solutions for SCI patients. Methods According the interaction between brain regions, a coherence-based graph convolutional network (C-GCN) method is proposed extract temporal-frequency-spatial features functional connectivity information of EEG signals. The algorithm constructs multi-channel coherence networks as graphical signals then classifies MI tasks. Different from traditional neural (GCN), C-GCN uses determine MI-related connections, which are used represent intrinsic connections channels in different rhythms data patients healthy subjects have been analyzed, where served control group. Results experimental results show that achieve best classification performance with certain reliability stability, highest accuracy 96.85%. Conclusion framework an theoretical basis rehabilitation

Язык: Английский

---

vEpiNet: A multimodal interictal epileptiform discharge detection method based on video and electroencephalogram data

Neural Networks, Год журнала: 2024, Номер 175, С. 106319 - 106319

Опубликована: Апрель 14, 2024

To enhance deep learning-based automated interictal epileptiform discharge (IED) detection, this study proposes a multimodal method, vEpiNet, that leverages video and electroencephalogram (EEG) data. Datasets comprise 24 931 IED (from 484 patients) 166 094 non-IED 4-second video-EEG segments. The data is processed by the proposed patient detection with frame difference Simple Keypoints (SKPS) capturing patients' movements. EEG EfficientNetV2. features are fused via multilayer perceptron. We developed comparative model, termed nEpiNet, to test effectiveness of feature in vEpiNet. 10-fold cross-validation was used for testing. showed high areas under receiver operating characteristic curve (AUROC) both models, slightly superior AUROC (0.9902) vEpiNet compared nEpiNet (0.9878). Moreover, model performance real-world scenarios, we set prospective dataset, containing 215 h raw from 50 patients. result shows achieves an area precision-recall (AUPRC) 0.8623, surpassing nEpiNet's 0.8316. Incorporating raises precision 70% (95% CI, 69.8%-70.2%) 76.6% 74.9%-78.2%) at 80% sensitivity reduces false positives nearly third, processing one-hour 5.7 min on average. Our findings indicate can significantly improve especially real clinic It suggests clinically viable effective tool analysis applications.

Язык: Английский

---