Electroencephalogram (EEG) signals provide rich spatiotemporal brain features crucial for epilepsy diagnosis. Though traditional deep neural networks combining attention mechanisms excel in feature extraction, these models do not fully emulate the neural processing efficiency of the brain and they fall short in fully harnessing the spatiotemporal dynamics of EEG data. This study introduces the dual attention-driven and feature fusion-based spiking neural network (DAFF-SNN), which synergizes the spatiotemporal attention mechanism with SNNs' inherent temporal processing prowess to enhance the precision and efficiency for epilepsy detection. The DAFF-SNN utilizes an adaptive spiking fusion module (ASFM) for feature integration, optimizing the feature fusion process by exploiting the spatiotemporal complementarity of EEG data through a spike-driven strategy. Evaluations on the Bonn, Beriut, CHB-MIT, and Siena datasets demonstrate DAFF-SNN's high accuracies (100%, 97.6%, 98.2%, 99.6% ), achieving comparable or superior performance to state-of-the-art ANN methods, highlighting its efficient epilepsy detection capability.

In recent years, brain-computer interfaces (BCIs) leveraging electroencephalography (EEG) signals for the control of external devices have garnered increasing attention. The information transfer rate of BCI has been significantly improved by a lot of cutting-edge methods. The exploration of effective preprocessing in brain-computer interfaces, particularly in terms of identifying suitable preprocessing methods and determining the optimal sequence for their application, remains an area ripe for further investigation. To address this gap, this study explores a range of preprocessing techniques, including but not limited to independent component analysis, surface Laplacian, bandpass filtering, and baseline correction, examining their potential contributions and synergies in the context of BCI applications. In this extensive research, a variety of preprocessing pipelines were rigorously tested across four EEG data sets, all of which were pertinent to motor imagery-based BCIs. These tests incorporated five EEG machine learning models, working in tandem with the preprocessing methods discussed earlier. The study's results highlighted that baseline correction and bandpass filtering consistently provided the most beneficial preprocessing effects. From the perspective of online deployment, after testing and time complexity analysis, this study recommends baseline correction, bandpass filtering and surface Laplace as more suitable for online implementation. An interesting revelation of the study was the enhanced effectiveness of the surface Laplacian algorithm when used alongside algorithms that focus on spatial information. Using appropriate processing algorithms, we can even achieve results (92.91% and 88.11%) that exceed the SOTA feature extraction methods in some cases. Such findings are instrumental in offering critical insights for the selection of effective preprocessing pipelines in EEG signal decoding. This, in turn, contributes to the advancement and refinement of brain-computer interface technologies.

This paper presents the design and analysis of a family of voltage reference generators (VRGs) based on the stacking of a current-source transistor MIand a resistive-load transistor MR, i.e., stacking of MI,R(SMIR) in standard CMOS technology for sub-1V and sub-nW operation. Design guidelines are provided to obtain the reference voltage for various temperature characteristics, namely proportional to absolute temperature (PTAT), complementary to absolute temperature (CTAT), and constant with temperature (CWT), by appropriately sizing the two transistors as current source and load respectively. The proposed 6 such SMIR VRGs in 65nm consume an average power less than 4pW and occupy an area less than 30µm × 100µm when measured from 20 different samples. All VRGs operate at a minimum supply voltage of 0.4V and achieve an average line regulation better than 0.3%/V. For the CWT VRGs, an average temperature coefficient better than 260.9ppm/°C is achieved from -20°C to 80°C.

Electroencephalogram (EEG)-based emotion recognition has become a research hotspot in the field of brain-computer interface. Previous emotion recognition methods have overlooked the fusion of multi-domain emotion-specific information to improve performance, and faced the challenge of insufficient interpretability. In this paper, we proposed a novel EEG emotion recognition model that combined the asymmetry of the brain hemisphere, and the spatial, spectral, and temporal multi-domain properties of EEG signals, aiming to improve emotion recognition performance. Based on the 10-20 standard system, a global spatial projection matrix (GSPM) and a bi-hemisphere discrepancy projection matrix (BDPM) are constructed. A dual-stream spatial-spectral-temporal convolution neural network is designed to extract depth features from the two matrix paradigms. Finally, the transformer-based fusion module is used to learn the dependence of fused features, and to retain the discriminative information. We conducted extensive experiments on the SEED, SEED-IV, and DEAP public datasets, achieving excellent average results of 98.33/2.46 %, 92.15/5.13 %, 97.60/1.68 %(valence), and 97.48/1.42 %(arousal) respectively. Visualization analysis supports the interpretability of the model, and ablation experiments validate the effectiveness of multi-domain and bi-hemisphere discrepancy information fusion.

OBJECTIVE: Deep learning based on convolutional neural networks (CNN) has achieved success in brain-computer interfaces (BCIs) using scalp electroencephalography (EEG). However, the interpretation of the so-called 'black box' method and its application in stereo-electroencephalography (SEEG)-based BCIs remain largely unknown. Therefore, in this paper, an evaluation is performed on the decoding performance of deep learning methods on SEEG signals. METHODS: Thirty epilepsy patients were recruited, and a paradigm including five hand and forearm motion types was designed. Six methods, including filter bank common spatial pattern (FBCSP) and five deep learning methods (EEGNet, shallow and deep CNN, ResNet, and a deep CNN variant named STSCNN), were used to classify the SEEG data. Various experiments were conducted to investigate the effect of windowing, model structure, and the decoding process of ResNet and STSCNN. RESULTS: The average classification accuracy for EEGNet, FBCSP, shallow CNN, deep CNN, STSCNN, and ResNet were 35 ± 6.1%, 38 ± 4.9%, 60 ± 3.9%, 60 ± 3.3%, 61 ± 3.2%, and 63 ± 3.1% respectively. Further analysis of the proposed method demonstrated clear separability between different classes in the spectral domain. CONCLUSION: ResNet and STSCNN achieved the first- and second-highest decoding accuracy, respectively. The STSCNN demonstrated that an extra spatial convolution layer was beneficial, and the decoding process can be partially interpreted from spatial and spectral perspectives. SIGNIFICANCE: This study is the first to investigate the performance of deep learning on SEEG signals. In addition, this paper demonstrated that the so-called 'black-box' method can be partially interpreted.

For a long time, people have carried out various studies on human body communication (HBC) in order to establish a suitable communication link through human body. However, in the galvanic coupled method of HBC, the high current intensity is rarely used to implement the communication link. In the medical field, functional electrical stimulation (FES) is often used to send high intensity electrical pulses to make muscles contract, and this contraction phenomenon will generate surface electromyography (sEMG) signals on the surface of human skins. According to this principle and the galvanic coupling method of HBC, we propose a new digital communication system based on FES and sEMG signal detection with human body as communication channel in this paper. We modulate the transmitted signal into electrical stimulation to stimulate the muscles and detect the sEMG signal caused by it to achieve a complete communication process. The framework of the entire communication system is proposed. Its error performance for different stimulation parameters is tested and evaluated by experiments. Using FES and sEMG signal detection, our work makes a new exploration of HBC at high current intensities and enables a complete communication link. This work is expected to be applied to the HBC design combined with electrical stimulation in medical field.

Molecular communication and the Internet of Nanothings (IoNT) are emerging research hotspots recently, which show great potential in biomedical applications inside the human body. However, how to transmit information from inside body IoNTs to outside devices is seldomly studied. It is well known that the nervous system is responsible for perceiving the external environment and controlling the feedback signals. It exactly works like an interface between the external and internal environment. Inspired by this, this article proposes a novel concept that one can use the modified nervous system to communicate between IoNT devices andin vitroequipments. In our proposed system, nanomachines transmit signals via stimulating the nerve fiber by the electrode. Then, the signals transmit along nerve fibers and muscle fibers. Finally, they cause changes in surface electromyography (sEMG) signals, which can be decoded by the body surface receiver. This article presents the framework of this entire through-body communication system. Each part of the framework is also mathematically modeled. The error probability and mutual information of the system are derived from the communication theory perspective, which are evaluated and analyzed through numerical results. This study can pave the way for the connection of IoNTin vivoto external networks.

It is ambitious to develop a brain-controlled robotic arm for some patients with motor impairments to perform activities of daily living using brain–computer interfaces (BCIs). Despite much progress achieved, this mission is still very challenging mainly due to the poor decoding performance of BCIs. The problem is even exacerbated in the case of noninvasive BCIs. A shared control strategy is developed in this work to realize flexible robotic arm control for reach and grasp of multiple objects. With the intelligent assistance provided by robot vision, the subject was only required to finish gross reaching movement and target selection using a simple motor imagery-based BCI with binary output. Along with the user control, the robotic arm, which identified and localized potential targets within the workspace in the background, was capable of providing both trajectory correction in the reaching phase to reduce trajectory redundancy and autonomous grasping assistance in the phase of grasp. Ten subjects participated in the experiments containing one session of two-block grasping tasks with fixed locations and another one of randomly placed three-block grasping tasks. The results of the experiments demonstrated substantial improvement with the shared control system. Compared with the single BCI control, the success rate of shared control was significantly higher ($p < 0.001$for group performance), and moreover, the task completion time and perceived difficulty were significantly lower ($p < 0.001$for group performance both), indicating the potential of our proposed shared control system in real applications.Note to Practitioners—This article is motivated by the problem of dexterous robotic arm control based on a brain–computer interface (BCI). For people suffering from severe neuromuscular disorders or accident injuries, a brain-controlled robotic arm is expected to provide assistance in their daily lives. A primary bottleneck to achieve the objective is that the information transfer rate of current BCIs is not high enough to produce multiple and reliable commands during the online robotic control. In this work, machine autonomy is incorporated in a BCI-controlled robotic arm system, where the user and machine can work together to reach and grasp multiple objects in a given task. The intelligent robot system autonomously localized the potential targets and provided trajectory correction and grasping assistance accordingly. Meanwhile, the user only needed to complete gross reaching movement and target selection with a basic binary motor imagery-based BCI, which reduced the task difficulty and retained the volitional involvement of the user at the same time. The results of the experiments showed that the accuracy and efficiency of grasping tasks increased significantly in the shared control mode together with a significant decrease in the perceived mental workload, which indicates that our proposed shared control system is effective and user-friendly in practice. In the future, more feedback information will be introduced to enhance the task performance further, and a wheelchair-mounted robotic arm system will be developed for greater flexibility. In addition, more functional task modules (e.g., self-feeding and opening doors) should be integrated for more practical utilities.

This paper presents a context-based approach for the recognition of transition between activities of daily living (ADLs) using wearable sensor data. A Bayesian method is implemented for the recognition of 7 ADLs with data from two wearable sensors attached to the lower limbs of subjects. A second Bayesian method recognises 12 transitions between the ADLs. The second recognition module uses both, data from wearable sensors and the activity recognised from the first Bayesian module. This approach analyses the next most probable transitions based on wearable sensor data and the context or current activity being performed by the subject. This work was validated using the ENABL3S Database composed of data collected from 7 ADLs and 12 transitions performed by participants walking on two circuits composed of flat surfaces, ascending and descending ramps and stairs. The recognition of activities achieved an accuracy of 98.3%. The recognition of transitions between ADLs achieved an accuracy of 98.8%, which improved the 95.3% accuracy obtained when the context or current activity is not considered for the recognition process. Overall, this work proposes an approach capable of recognising transitions between ADLs, which is required for the development of reliable wearable assistive robots.

This paper presents a novel sliding mode controller for trajectory tracking of the piezo-driven stage. The tracking performance of piezoelectric actuator is mainly affected by the hysteresis nonlinearity. Sliding mode control is a possible solution to achieve better tracking performance. However, conventional sliding mode control generates discontinuous control signal which results in chattering. Hence, the hysteresis nonlinearity is first compensated with a hysteresis model, and an uncertainty and disturbance estimator is designed and included to devise a smooth control action. The stability of the proposed method is demonstrated via Lyapunov analysis. Both simulation and experiment are also conducted to verify the effectiveness of the proposed approach. The results are compared with a conventional sliding mode controller and a proportional-integral control with notch filter (PIC-NF).

The lost motor functions of an upper-limb amputee can be restored by means of multi-DOF myoelectric prostheses. However, the somatosensory (tactile and proprioceptive) feedback from a commercial prosthetic hand to the user is still missing, especially the proprioceptive feedback (PF). An object grasping or manipulation actually are organized in phases characterized by muscle synergy and delimited by means of discrete sensory "events". Inspired by the Discrete Event-driven Sensory feedback Control (DESC) policy, we delimited the continuous grasping motion into discrete wrist and finger motions to evaluate human proprioceptive matching ability. In current study, four kinds of typical hand motions (radial flexion, ulnar flexion, wrist flexion and lateral prehension) were passively generated by stimulating respective forearm dominated muscles via non-invasive electrical stimulation (ES) and then actively reappeared with the ipsilateral hands on eight able-bodied subjects. Under two types of matching conditions (PF and PF+ visual feedback (VF)), the human proprioceptive matching ability were evaluated and analyzed. Based on this, a feasible method (interface) used for encoding the grasping movement from fingers of the prosthetic hand to an upper-limb amputee was proposed.

Stereo-electroencephalography (SEEG) creates unique opportunities for clinical applications, brain-machine-interface (BMI) studies and neuroscientific research. However, all of these directions have been under-explored because the lack of an easy-to-use and multi-functional software to accurately localize the SEEG electrodes in 3D space. In this work, we present, for the first time, an SEEG-specific MATLAB-based toolbox (SEEGview) for electrode localization that identifies 3D coordinates and anatomical information for all SEEG contacts within each individual brain, generates cortical activation maps, and maps SEEG contacts from different subjects into one standard brain model. This article first describes how each of these four functions (3D localization, anatomical localization, cortical activation mapping, and SEEG contact standardization) is implemented through SEEGview. We then present localization results based on SEEG data, thereby demonstrating the high localization accuracy and viability for SEEG research.

With the development of technology, Human-Computer Interface (HCI) system is playing a more and more important role in our daily life. HCI is a way to set up connections and to transfer information between human and computer. Pattern recognition based on Surface Electromyography (SEMG) is one of the most important HCI technologies. To make the input device of electronic products more portable to satisfy the people' need of interacting with computer (especially disabled people), this research proposes an SEMG-based handwriting recognition system based on LM-BP (Back Propagation) Neural Network. In the aspect of signal preprocessing, this thesis tries to use some new features of signals to reflect the features of signals better. In the aspect of pattern recognition, LM-BP Neural Networking is applied to design a system that is suitable for SEMG-based model training and recognition. Compared with the existing system based on Dynamic Time Warping (DTW) algorithm and the system based on Hidden Markov Model (HMM), the training times and training time have been reduced a lot, which makes the SEMG-based handwriting recognition system more practical.

Psychophysical tests and standardized questionnaires are often used to analyze tactile sensation based on subjective judgment in conventional studies. In contrast with the subjective evaluation, a novel method based on electroencephalography (EEG) is proposed to explore the possibility of quantifying tactile sensation in an objective way. The proposed experiments adopt cutaneous electrical stimulation to generate two kinds of sensations (vibration and pressure) with three grades (low/medium/strong) on eight subjects. Event-related potentials (ERPs) and event-related synchronization/desynchronization (ERS/ERD) are extracted from EEG, which are used as evaluation indexes to distinguish between vibration and pressure, and also to discriminate sensation grades. Results show that five-phase P1–N1–P2–N2–P3 deflection is induced in EEG. Using amplitudes of latter ERP components (N2 and P3), vibration and pressure sensations can be discriminated on both individual and grand-averaged ERP (p < 0.05). The grand-average ERPs can distinguish the three sensations grades, but there is no significant difference on individuals. In addition, ERS/ERD features of mu rhythm (8–13 Hz) are adopted. Vibration and pressure sensations can be discriminated on grand-average ERS/ERD (p < 0.05), but only some individuals show significant difference. The grand-averaged results show that most sensation grades can be differentiated, and most pairwise comparisons show significant difference on individuals (p < 0.05). The work suggests that ERP- and ERS/ERD-based EEG features may have potential to quantify tactile sensations for medical diagnosis or engineering applications.

Control scheme design based on surface electromyography (sEMG) pattern recognition has been the focus of much research on a myoelectric prosthesis (MP) technology. Due to inherent nonstationarity in sEMG signals, prosthesis systems may need to be recalibrated day after day in daily use applications; thereby, hindering MP usability. In order to reduce the recalibration time in the subsequent days following the initial training, we propose a domain adaptation (DA) framework, which automatically reuses the models trained in earlier days as input for two baseline classifiers: a polynomial classifier (PC) and a linear discriminant analysis (LDA). Two novel algorithms of DA are introduced, one for PC and the other one for LDA. Five intact-limbed subjects and two transradial-amputee subjects participated in an experiment lasting ten days, to simulate the application of a MP over multiple days. The experiment results of four methods were compared: PC-DA (PC with DA), PC-BL (baseline PC), LDA-DA (LDA with DA), and LDA-BL (baseline LDA). In a new day, the DA methods reuse nine pretrained models, which were calibrated by 40 s training data per class in nine previous days. We show that the proposed DA methods significantly outperform nonadaptive baseline methods. The improvement in classification accuracy ranges from 5.49% to 28.48%, when the recording time per class is 2 s. For example, the average classification rates of PC-BL and PC-DA are 83.70% and 92.99%, respectively, for intact-limbed subjects with a nine-motions classification task. These results indicate that DA has the potential to improve the usability of MPs based on pattern recognition, by reducing the calibration time.

This paper proposes the idea that Hindmarsh-Rose (HR) neuronal model can be used to develop a new type of central pattern generator (CPG). Some key properties of HR model are studied and proved to meet the requirements of CPG. Pros and cons of HR model are provided. A CPG network based on HR model is developed and the related properties are investigated. We explore the bipedal primary gaits generated by the CPG network. The preliminary applications of HR model are tested on humanoid locomotion model and functional electrical stimulation (FES) walking system. The positive results of stimulation and experiment show the feasibility of HR model as a valid CPG.

Variations in muscle contraction effort have a substantial impact on performance of pattern recognition based myoelectric control. Though incorporating changes into training phase could decrease the effect, the training time would be increased and the clinical viability would be limited. The modulation of force relies on the coordination of multiple muscles, which provides a possibility to classify motions with different forces without adding extra training samples. This study explores the property of muscle coordination in the frequency domain and found that the orientation of muscle activation pattern vector of the frequency band is similar for the same motion with different force levels. Two novel features based on discrete Fourier transform and muscle coordination were proposed subsequently, and the classification accuracy was increased by around 11% compared to the traditional time domain feature sets when classifying nine classes of motions with three different force levels. Further analysis found that both features decreased the difference among different forces of the same motion ) and maintained the distance among different motions p > 0.1). This study also provided a potential way for simultaneous classification of hand motions and forces without training at all force levels.

One problem in the development of brain-computer interface (BCI) systems is to minimize the amount of subject training on the premise of accurate classification. Hence, the challenge is how to train the BCI system effectively especially in the scenario with small amount of training data. In this paper, we introduce improved semisupervised adaptation based on common spatial pattern (CSP) features. The feature extraction and classification are performed jointly and iteratively. In the iteration step, training data are expanded by part of the testing data with labels which are predicted by a linear discriminant analysis classifier and/or a Bayesian linear discriminant analysis classifier in the previous iteration. Then CSP features are reextracted from the expanded training data, and the classifiers are retrained. Both self-training and cotraining paradigms are proposed for the improved semisupervised adaptation. Throughout the investigation on different number of initial training trials, we find that when a small number of training trials are used, e.g., a training session contains no more than 30 trials, similar classification performance to that of large training data items (40-50 trials) can be achieved. Effectiveness of the algorithms is verified by two competition datasets. Compared with several existing algorithms, the proposed semisupervised algorithms show improvements in classification accuracy for most of the competition datasets especially in the case of small training data.

This paper presents our new progress on research of electromyography (EMG) controlled prosthetic hand. Preliminary clinical study is conducted on amputees. EMG data from the residual muscles of three amputees are used to evaluate the new features, discriminant bispectra (DBS) and discriminant Fourier cepstrum (DFC). The performance is also compared with other traditional features, autoregressive (AR) coefficient, time domain statics (TDS), and power spectral distribution (PSD). Some promising results are presented. It shows that DFC and DBS outperforms the other features. The implementation of EMG algorithm is accomplished. Based on the prosthetic hand (SJT-2 hand), experimental demonstration is conducted by both healthy and amputated subjects.

In this paper, a strategy for pathological tremor compensation based on co-contraction of antagonist muscles induced by Functional Electrical Stimulation (FES) is presented. Although one of the simplest alternatives to apply FES for reducing the effects of tremor, the contribution of different co- contraction levels for joint motion and impedance must be accurately estimated, specially since tremor itself is highly time-varying. In this work, a detailed musculoskeletalmodel of the human wrist actuated by flexor and extensor muscles is used for this purpose. The model takes into account different properties that affect muscle dynamics, such as proprioceptive feed- back and combined natural and artificial activation. The model, analysis of stiffness modulation due to FES-controlled co-contraction and simulation results are presented in the paper.

This paper studies the mechanisms of ionic channels in neurons of lamprey central pattern generator (CPG), such as the N-methyl-D-aspartate (NMDA) receptor channel and the calcium-dependent potassium channel etc. The CPG properties on oscillation attributed to the ionic mechanisms are exploited. The conditions for oscillation, divergence, convergence and the guidelines on selection of the parameters are established. The effects of key parameters on CPG frequency and duty cycle are investigated. Mathematical analysis and simulation study is performed to verify these results. This study will potentially enhance the effective application of biological CPG model into engineering practice such as robotics.

This paper presents theoretical results of the neural control mechanism existed in the spinal cord, central pattern generator (CPG), which has the ability to provide rhythmic movement patterns for the invertebrate and vertebrate. It is known that although CPG is verified by biological methods, it still lacks a complete theoretical investigation. The theoretical analysis about CPG from engineering perspective such as parameter selection and conditions of stable oscillation will strengthen the foundation of CPG theory, and it will significantly enhance the effective application of CPG into motor control systems.

Central pattern generator (CPG) is a neuronal circuit in the nervous system that can generate oscillatory patterns for the rhythmic movements. Its simplified format, neural oscillator, is wildly adopted in engineering application. This paper explores the CPG from an integral view that combines biology and engineering together. Biological CPG and simplified CPG are both studied. Computer simulation reveals the mechanism of CPG. Some properties, such as effect of tonic input and sensory feedback, stable oscillation, robustness, entrainment etc., are further studied. The promising results provide foundation for the potential engineering application in future.

A neural network control system is developed to control the leg rhythmic movements with functional electrical stimulation (FES). The neural network contains two parts. One part is the neural oscillator. It derives from the central pattern generator of biological nervous system and serves as a feedforward controller generating the rhythmic pattern for the leg movements. Another part of the neural network is a radial basis function (RBF) artificial neural network, which maps the signals from the neural oscillator to regulate the FES stimulator, then the stimulator generates appropriate electrical pulse trains to stimulate the muscles. A computer dynamic model of a leg is developed in this paper. The model comprises of skeletal dynamics, revised hill-type muscle model with excitation-contraction dynamics. We have done some simulation studies with this control system on the one-segment leg musculoskeletal model. It shows satisfactory tracking performance and fast tracking response. This idea indicates feasible clinical application for the paralyzed patients in the future work.