Research on Brain Networks of Human Balance Based on Phase Estimation Synchronization.

Phase synchronization serves as an effective method for analyzing the synchronization of electroencephalogram (EEG) signals among brain regions and the dynamic changes of the brain. The purpose of this paper is to study the construction of the functional brain network (FBN) based on phase synchronization, with a special focus on neural processes related to human balance regulation. This paper designed four balance paradigms of different difficulty by blocking vision or proprioception and collected 19-channel EEG signals. Firstly, the EEG sequences are segmented by sliding windows. The phase-locking value (PLV) of core node pairs serves as the phase-screening index to extract the valid data segments, which are recombined into new EEG sequences. Subsequently, the multichannel weighted phase lag index (wPLI) is calculated based on the new EEG sequences to construct the FBN. The experimental results show that due to the randomness of the time points of body balance adjustment, the degree of phase synchronization of the datasets screened by PLV is more obvious, improving the effective information expression of the subsequent EEG data segments. The FBN topological structures of the wPLI show that the connectivity of various brain regions changes structurally as the difficulty of human balance tasks increases. The frontal lobe area is the core brain region for information integration. When vision or proprioception is obstructed, the EEG synchronization level of the corresponding occipital lobe area or central area decreases. The synchronization level of the frontal lobe area increases, which strengthens the synergistic effect among the brain regions and compensates for the imbalanced response caused by the lack of sensory information. These results show the brain regional characteristics of the process of human balance regulation under different balance paradigms, providing new insights into endogenous neural mechanisms of standing balance and methods of constructing brain networks.

Skin marker combined with surface-guided auto-positioning for breast DIBH radiotherapy daily initial patient setup: An optimal schedule for both accuracy and efficiency.

BACKGROUND AND PURPOSE: By employing three surface-guided radiotherapy (SGRT)-assisted positioning methods, we conducted a prospective study of patients undergoing SGRT-based deep inspiration breath-hold (DIBH) radiotherapy using a Sentine/Catalys system. The aim of this study was to optimize the initial positioning workflow of SGRT-DIBH radiotherapy for breast cancer. MATERIALS AND METHODS: A total of 124 patients were divided into three groups to conduct a prospective comparative study of the setup accuracy and efficiency for the daily initial setup of SGRT-DIBH breast radiotherapy. Group A was subjected to skin marker plus SGRT verification, Group B underwent SGRT optical feedback plus auto-positioning, and Group C was subjected to skin marker plus SGRT auto-positioning. We evaluated setup accuracy and efficiency using cone-beam computed tomography (CBCT) verification data and the total setup time. RESULTS: In groups A, B, and C, the mean and standard deviation of the translational setup-error vectors were small, with the highest values of the three directions observed in group A (2.4 ± 1.6, 2.9 ± 1.8, and 2.8 ± 2.1 mm). The rotational vectors in group B (1.8 ± 0.7°, 2.1 ± 0.8°, and 1.8 ± 0.7°) were significantly larger than those in groups A and C, and the Group C setup required the shortest amount of time, at 1.5 ± 0.3 min, while that of Group B took the longest time, at 2.6 ± 0.9 min. CONCLUSION: SGRT one-key calibration was found to be more suitable when followed by skin marker/tattoo and in-room laser positioning, establishing it as an optimal daily initial set-up protocol for breast DIBH radiotherapy. This modality also proved to be suitable for free-breathing breast cancer radiotherapy, and its widespread clinical use is recommended.

Advances in nanotechnology-based targeted-contrast agents for computed tomography and magnetic resonance.

X-ray computed tomography (CT) and magnetic resonance (MR) imaging are essential tools in modern medical diagnosis and treatment. However, traditional contrast agents are inadequate in the diagnosis of various health conditions. Consequently, the development of targeted nano-contrast agents has become a crucial area of focus in the development of medical image-enhancing contrast agents. To fully understand the current development of nano-contrast agents, this review provides an overview of the preparation methods and research advancements in CT nano-contrast agents, MR nano-contrast agents, and CT/MR multimodal nano-contrast agents described in previous publications. Due to the physicochemical properties of nanomaterials, such as self-assembly and surface modifiability, these specific nano-contrast agents can greatly improve the targeting of lesions through various preparation methods and clearly highlight the distinction between lesions and normal tissues in both CT and MR. As a result, they have the potential to be used in the early stages of disease to improve diagnostic capacity and level in medical imaging.

Influence of Anodal tDCS on the Brain Functional Networks and Muscle Synergy of Hand Movements.

BACKGROUND: Transcranial direct current stimulation (tDCS) is a non-invasive technique that has demonstrated potential in modulating cortical neuron excitability. The objective of this paper is to investigate the effects of tDCS on characteristic parameters of brain functional networks and muscle synergy, as well as to explore its potential for enhancing motor performance. METHODS: By applying different durations of tDCS on the motor cortex of the brain, the 32-lead electroencephalogram (EEG) of the cerebral cortex and 4-lead electromyography (EMG) signals of the right forearm were collected for 4 typical hand movements which are commonly used in rehabilitation training, including right-hand finger flexion, finger extension, wrist flexion, and wrist extension. RESULTS: The study showed that tDCS can enhance the brain's electrical activity in the beta band of the C3 node of the cerebral cortex during hand movements. Furthermore, the structure of muscle synergy remains unaltered; however, the associated muscle activity is amplified (p < 0.05). CONCLUSIONS: Based on the study results, it can be inferred that tDCS enhances the control strength between the motor area of the cerebral cortex and the muscles during hand movements.

Analysis of Cortico-Muscular Coupling and Functional Brain Network under Different Standing Balance Paradigms.

Maintaining standing balance is essential for people to engage in productive activities in daily life. However, the process of interaction between the cortex and the muscles during balance regulation is understudied. Four balance paradigms of different difficulty were designed by closing eyes and laying sponge pad under feet. Ten healthy subjects were recruited to stand for ten 15 s trials in each paradigm. This study used simultaneously acquired electroencephalography (EEG) and electromyography (EMG) to investigate changes in the human cortico-muscular coupling relationship and functional brain network characteristics during balance control. The coherence and causality of EEG and EMG signals were calculated by magnitude-squared coherence (MSC) and transfer entropy (TE). It was found that changes in balance strategies may lead to a shift in cortico-muscular coherence (CMC) from the beta band to the gamma band when the difficulty of balance increased. As subjects performed the four standing balance paradigms, the causality of the beta band and the gamma band was stronger in the descending neural pathway than that in the ascending neural pathway. A multi-rhythmic functional brain network with 19 EEG channels was constructed and analyzed based on graph theory, showing that its topology also changed with changes in balance difficulty. These results show an active adjustment of the sensorimotor system under different balance paradigms and provide new insights into the endogenous physiological mechanisms underlying the control of standing balance.

Chinese sign language recognition based on surface electromyography and motion information.

Sign language (SL) has strong structural features. Various gestures and the complex trajectories of hand movements bring challenges to sign language recognition (SLR). Based on the inherent correlation between gesture and trajectory of SL action, SLR is organically divided into gesture-based recognition and gesture-related movement trajectory recognition. One hundred and twenty commonly used Chinese SL words involving 9 gestures and 8 movement trajectories, are selected as research and test objects. The method based on the amplitude state of surface electromyography (sEMG) signal and acceleration signal is used for vocabulary segmentation. The multi-sensor decision fusion method of coupled hidden Markov model is used to complete the recognition of SL vocabulary, and the average recognition rate is 90.41%. Experiments show that the method of sEMG signal and motion information fusion has good practicability in SLR.

ULFAC-Net: Ultra-Lightweight Fully Asymmetric Convolutional Network for Skin Lesion Segmentation.

Segmentation of skin lesions is a critical step in the process of skin lesion diagnosis. Such segmentation is challenging due to the irregular shape, fuzzy contours and severe noise interference in the skin lesion region. Existing deep learning-based skin lesion segmentation methods are usually computationally expensive, hindering their deployment in dermoscopic devices with poor computational power. To address these challenges, we propose an ultralightweight fully asymmetric convolutional network for skin lesion segmentation, called ULFAC-Net. we use a parallel asymmetric convolutional (PAC) module to extract features instead of the traditional square convolution, and innovatively propose a PAC module with dual attention (Att-PAC) to enhance the feature representation. Based on the PAC and Att-PAC modules, we further propose a lightweight textual information submodule. To balance the number of parameters and performance of the model, we also hand-design an asymmetric encoder-decoder architecture. In this paper, we validate the effectiveness and robustness of the proposed ULFAC-Net on four publicly available skin lesion segmentation datasets (ISIC2018, ISBI2017, ISIC2016 and PH2 datasets). The experimental results show that ULFAC-Net achieves competitive segmentation performance with only 0.842 million(0.842M) parameters and 3.71 gigabytes of floating point operations (GFLOPs) compared to other state-of-the-art methods.

SGRT-based DIBH radiotherapy practice for right-sided breast cancer combined with RNI: A retrospective study on dosimetry and setup accuracy.

BACKGROUND: We retrospectively studied the dosimetry and setup accuracy of deep inspiration breath-hold (DIBH) radiotherapy in right-sided breast cancer patients with regional nodal irradiation (RNI) who had completed treatment based on surface-guided radiotherapy (SGRT) technology by Sentinel/Catalyst system, aiming to clarify the clinical application value and related issues. METHODS: Dosimetric indicators of four organs at risk (OARs), namely the heart, right coronary artery (RCA), right lung, and liver, were compared on the premise that the planning target volume met dose-volume prescription requirements. Meanwhile, the patients were divided into the edge of the xiphoid process (EXP), sternum middle (SM), and left breast wall (LBW) groups according to different positions of respiratory gating primary points. The CBCT setup error data of the three groups were contrasted for the treatment accuracy study, and the effects of different gating window heights on the right lung volume increases were compared among the three groups. RESULTS: Compared with free breath (FB), DIBH reduced the maximum dose of heart and RCA by 739.3 ± 571.2 cGy and 509.8 ± 403.8 cGy, respectively (p < 0.05). The liver changed the most in terms of the mean dose (916.9 ± 318.9 cGy to 281.2 ± 150.3 cGy, p < 0.05). The setup error of the EXP group in the anterior-posterior (AP) direction was 3.6 ± 4.5 mm, which is the highest among the three groups. The right lung volume increases in the EXP, SM, and LBW groups were 72.3%, 69.9%, and 67.2%, respectively (p = 0.08), and the corresponding breath-holding heights were 13.5 ± 3.7 mm, 10.3 ± 2.4 mm, and 9.6 ± 2.8 mm, respectively (p < 0.05). CONCLUSIONS: SGRT-based DIBH radiotherapy can better protect the four OARs of right-sided breast cancer patients with RNI. Different respiratory gating primary points have different setup accuracy and breath-hold height.

Research on the Effect of MT+FES Training on Sensorimotor Cortex.

Purpose: Aiming at the motor recovery of patients with unilateral upper limb motor dysfunction after stroke, we propose a mirror therapy (MT) training method, which uses surface electromyography (sEMG) to identify movements on one side and control the other side to perform functional electrical stimulation (FES) while mirror therapy is used. And we verify the effect of this training method by analyzing the activity changes of the sensorimotor cortex. Method: Ten subjects (6 men and 4 women) were randomly divided into two groups according to 3 men and 2 women in each group: the experimental group (n = 5) received FES+MT training, and the control group (n = 5) received MT training. Both groups were trained at a fixed time at 9 : 00 am every day, each time lasting 20 minutes, once a day, 5 days a week, continuous training for 4 weeks, and the training action was elbow flexion training. During the training of the elbow flexion exercise, the experimental group applied FES with a frequency of 30 Hz, a pulse width of 100 µs, and a current of 10 mA to the muscles corresponding to the elbow flexion exercise, and rested for 10 s after 10-s stimulation. We collect the EEG of the elbow flexion motor imagery of all subjects before and after training, and calculate the eigenvalue E, and analyze the effect of FES+MT training on the activity of the cerebral sensorimotor cortex. Results: After repeated measure (RM) two-way ANOVA of the two groups, comparing the subjects' µ rhythm elbow flexion motor imagery eigenvalue E, the experimental group (after training) > the control group (after training) > before training. Conclusion: The FES+MT training method has obvious activation effect on the cerebral sensorimotor cortex.

Functional integration of mirror neuron system and sensorimotor cortex under virtual self-actions visual perception.

Virtual reality (VR) technology, with the advantage of immersive visual experience, has been increasingly applied in the rehabilitation therapy of motor deficits. The functional integration of the mirror neuron system and the sensorimotor cortex under the visual perception of actions is one of the theoretical bases for the application of action observation in the neurorehabilitation of motor deficits. Whether the visual experience changes brought by VR technology can further promote this functional integration to be further confirmed. Using the exact low-resolution brain electromagnetic tomography (eLORETA) source localization method, we calculated and statistically tested the whole brain cortical voxel current density estimation under the Electroencephalogram (EEG) signals collected during action observation under the first-person and third-person perspectives in the VR scene for twenty healthy adults. Furthermore, the functional connectivity between the mirror neuron system and the sensorimotor cortex was analyzed using the lagged phase synchronization method. Under the first-person perspective in the VR scene, the current density changes of the core cortices of the mirror neuron system lead to a larger average event-related potential, more significant suppression in the α1 and α2 frequency bands of EEG signals, and a significant enhancement of functional connectivity between the core cortices of the mirror neuron system and the sensorimotor cortex. These findings indicate that compared with the traditional action observation, the visual reappearance of self-actions in the VR scene further stimulates the activity of the core cortices of the mirror neuron system, and promotes the functional integration of the core cortices of the mirror neuron system and the sensorimotor cortex.

Research on AR-AKF Model Denoising of the EMG Signal.

Electromyography (EMG) signals can be used for clinical diagnosis and biomedical applications. It is very important to reduce noise and to acquire accurate signals for the usage of the EMG signals in biomedical engineering. Since EMG signal noise has the time-varying and random characteristics, the present study proposes an adaptive Kalman filter (AKF) denoising method based on an autoregressive (AR) model. The AR model is built by applying the EMG signal, and the relevant parameters are integrated to find the state space model required to optimally estimate AKF, eliminate the noise in the EMG signal, and restore the damaged EMG signal. To be specific, AR autoregressive dynamic modeling and repair for distorted signals are affected by noise, and AKF adaptively can filter time-varying noise. The denoising method based on the self-learning mechanism of AKF exhibits certain capabilities to achieve signal tracking and adaptive filtering. It is capable of adaptively regulating the model parameters in the absence of any prior statistical knowledge regarding the signal and noise, which is aimed at achieving a stable denoising effect. By comparatively analyzing the denoising effects exerted by different methods, the EMG signal denoising method based on the AR-AKF model is demonstrated to exhibit obvious advantages.

Sparse representation-based classification with two-dimensional dictionary optimization for motor imagery EEG pattern recognition.

BACKGROUND: Sparse representation-based classification (SRC) has more advantages in motor imagery EEG pattern recognition, and the quality of dictionary construction directly determines the performance of SRC. In this paper, we proposed a two-dimensional dictionary optimization (TDDO) method to directly improve the performance of SRC. NEW METHOD: Firstly, an initial dictionary was constructed with multi-band features extracted by filter band common spatial pattern (FBCSP). Then Lasso regression is used to select significant features in each atom synchronously in the horizontal direction, and the KNN-based method is used to clean up noise atoms in the vertical direction. Finally, an SRC method by training samples linearly representing test samples was implemented in classification. RESULTS: The results show the necessity and rationality of TDDO-SRC method. The highest average classification accuracy of 86.5% and 92.4% is obtained on two public datasets. COMPARISON WITH EXISTING METHOD(S): The proposed method has more superior classification accuracy compared to traditional methods and existing winners' methods. CONCLUSIONS: The quality of dictionary construction has a great impact on the robustness of SRC. And compared with the original SRC, the classification accuracy of the optimized TDDO-SRC is greatly improved.

Optimal channel-based sparse time-frequency blocks common spatial pattern feature extraction method for motor imagery classification.

Common spatial pattern (CSP) as a spatial filtering method has been most widely applied to electroencephalogram (EEG) feature extraction to classify motor imagery (MI) in brain-computer interface (BCI) applications. The effectiveness of CSP is determined by the quality of interception in a specific time window and frequency band. Although numerous algorithms have been designed to optimize CSP by splitting the EEG data with a sliding time window and dividing the frequency bands with a set of band-pass filters, simultaneously. However, they did not consider the drawbacks of the rapid increase in data volume and feature dimensions brought about by this method, which would reduce the classification accuracy and calculation efficiency of the model. Therefore, we propose an optimal channel-based sparse time-frequency blocks common spatial pattern (OCSB-CSP) feature extraction method to improve the model classification accuracy and computational efficiency. Comparative experiments on two public EEG datasets show that the proposed method can quickly select significant time-frequency blocks and improve classification performance. The average classification accuracies are higher than those of other winners' methods, providing a new idea for the improvement of BCI applications.

Cortico-muscular functional network: an exploration of cortico-muscular coupling in hand movements.

Objective. The main objective of this research was to study cortico-muscular, intra-cortical, and inter-muscular coupling. Herein, we established a cortico-muscular functional network (CMFN) to assess the network differences associated with making a fist, opening the hand, and wrist flexion.Approach. We used transfer entropy (TE) to calculate the causality between electroencephalographic and electromyographic data and established the TE connection matrix. We then applied graph theory to analyze the clustering coefficient, global efficiency, and small-world attributes of the CMFN. We also used Relief-F to extract the features of the TE connection matrix of the beta2 band for the different hand movements and observed high accuracy when this feature was used for action recognition.Main results. We found that the CMFN of the three actions in the beta band had small-world attributes, among which the beta2 band's small-world was stronger. Moreover, we found that the extracted features were mainly concentrated in the left frontal area, left motor area, occipital lobe, and related muscles, suggesting that the CMFN could be used to assess the coupling differences between the cortex and the muscles that are associated with different hand movements. Overall, our results showed that the beta2 (21-35 Hz) wave is the main information carrier between the cortex and the muscles, and the CMFN can be used in the beta2 band to assess cortico-muscular coupling.Significance. Our study preliminarily explored the CMFN associated with hand movements, providing additional insights regarding the transmission of information between the cortex and the muscles, thereby laying a foundation for future rehabilitation therapy targeting pathological cortical areas in stroke patients.

Research on Recognition of Motor Imagination Based on Connectivity Features of Brain Functional Network.

Feature extraction is essential for classifying different motor imagery (MI) tasks in a brain-computer interface. To improve classification accuracy, we propose a novel feature extraction method in which the connectivity increment rate (CIR) of the brain function network (BFN) is extracted. First, the BFN is constructed on the basis of the threshold matrix of the Pearson correlation coefficient of the mu rhythm among the channels. In addition, a weighted BFN is constructed and expressed by the sum of the existing edge weights to characterize the cerebral cortex activation degree in different movement patterns. Then, on the basis of the topological structures of seven mental tasks, three regional networks centered on the C3, C4, and Cz channels are constructed, which are consistent with correspondence between limb movement patterns and cerebral cortex in neurophysiology. Furthermore, the CIR of each regional functional network is calculated to form three-dimensional vectors. Finally, we use the support vector machine to learn a classifier for multiclass MI tasks. Experimental results show a significant improvement and demonstrate the success of the extracted feature CIR in dealing with MI classification. Specifically, the average classification performance reaches 88.67% which is higher than other competing methods, indicating that the extracted CIR is effective for MI classification.

Effect of muscle fatigue on the cortical-muscle network: A combined electroencephalogram and electromyogram study.

Electroencephalogram (EEG) and electromyogram (EMG) signals during motion control reflect the interaction between the cortex and muscle. Therefore, dynamic information regarding the cortical-muscle system is of significance for the evaluation of muscle fatigue. We treated the cortex and muscle as a whole system and then applied graph theory and symbolic transfer entropy to establish an effective cortical-muscle network in the beta band (12-30 Hz) and the gamma band (30-45 Hz). Ten healthy volunteers were recruited to participate in the isometric contraction at the level of 30% maximal voluntary contraction. Pre- and post-fatigue EEG and EMG data were recorded. According to the Borg scale, only data with an index greater than 14<19 were selected as fatigue data. The results show that after muscle fatigue: (1) the decrease in the force-generating capacity leads to an increase in STE of the cortical-muscle system; (2) increases of dynamic forces in fatigue leads to a shift from the beta band to gamma band in the activity of the cortical-muscle network; (3) the areas of the frontal and parietal lobes involved in muscle activation within the ipsilateral hemibrain have a compensatory role. Classification based on support vector machine algorithm showed that the accuracy is improved compared to the brain network. These results illustrate the regulation mechanism of the cortical-muscle system during the development of muscle fatigue, and reveal the great potential of the cortical-muscle network in analyzing motor tasks.

Gesture Recognition Based on Multiscale Singular Value Entropy and Deep Belief Network.

As an important research direction of human-computer interaction technology, gesture recognition is the key to realizing sign language translation. To improve the accuracy of gesture recognition, a new gesture recognition method based on four channel surface electromyography (sEMG) signals is proposed. First, the S-transform is applied to four channel sEMG signals to enhance the time-frequency detail characteristics of the signals. Then, multiscale singular value decomposition is applied to the multiple time-frequency matrix output of S-transform to obtain the time-frequency joint features with better robustness. The corresponding singular value permutation entropy is calculated as the eigenvalue to effectively reduce the dimension of multiple eigenvectors. The gesture features are used as input into the deep belief network for classification, and nine kinds of gestures are recognized with an average accuracy of 93.33%. Experimental results show that the multiscale singular value permutation entropy feature is especially suitable for the pattern classification of the deep belief network.

Multi-class motor imagery EEG classification using collaborative representation-based semi-supervised extreme learning machine.

Both labeled and unlabeled data have been widely used in electroencephalographic (EEG)-based brain-computer interface (BCI). However, labeled EEG samples are generally scarce and expensive to collect, while unlabeled samples are considered to be abundant in real applications. Although the semi-supervised learning (SSL) allows us to utilize both labeled and unlabeled data to improve the classification performance as against supervised algorithms, it has been reported that unlabeled data occasionally undermine the performance of SSL in some cases. To overcome this challenge, we propose a collaborative representation-based semi-supervised extreme learning machine (CR-SSELM) algorithm to evaluate the risk of unlabeled samples by a new safety-control mechanism. Specifically, the ELM model is firstly used to predict unlabeled samples and then the collaborative representation (CR) approach is employed to reconstruct the unlabeled samples according to the obtained prediction results, from which the risk degree of unlabeled sample is defined. A risk-based regularization term is then constructed accordingly and embedded into the objective function of the SS-ELM. Experiments conducted on benchmark and EEG datasets demonstrate that the proposed method outperforms the ELM and SS-ELM algorithm. Moreover, the proposed CR-SSELM even offers the best performance while SS-ELM yields worse performance compared with its supervised counterpart (ELM). Graphical abstract This paper proposes a collaborative representation-based semi-supervised extreme learning machine (CR-SSELM) algorithm to evaluate the risk of unlabeled samples by a new safety-control mechanism. It is aim to solve the safety problem of SS-ELM method that SS-ELM yields worse performance than ELM. With the help of safety mechanism, the performance of our method is still better than supervised ELM method.

Double-Criteria Active Learning for Multiclass Brain-Computer Interfaces.

Recent technological advances have enabled researchers to collect large amounts of electroencephalography (EEG) signals in labeled and unlabeled datasets. It is expensive and time consuming to collect labeled EEG data for use in brain-computer interface (BCI) systems, however. In this paper, a novel active learning method is proposed to minimize the amount of labeled, subject-specific EEG data required for effective classifier training, by combining measures of uncertainty and representativeness within an extreme learning machine (ELM). Following this approach, an ELM classifier was first used to select a relatively large batch of unlabeled examples, whose uncertainty was measured through the best-versus-second-best (BvSB) strategy. The diversity of each sample was then measured between the limited labeled training data and previously selected unlabeled samples, and similarity is measured among the previously selected samples. Finally, a tradeoff parameter is introduced to control the balance between informative and representative samples, and these samples are then used to construct a powerful ELM classifier. Extensive experiments were conducted using benchmark and multiclass motor imagery EEG datasets to evaluate the efficacy of the proposed method. Experimental results show that the performance of the new algorithm exceeds or matches those of several state-of-the-art active learning algorithms. It is thereby shown that the proposed method improves classifier performance and reduces the need for training samples in BCI applications.

Feature Extraction of Surface Electromyography Based on Improved Small-World Leaky Echo State Network.

Surface electromyography (sEMG) is an electrophysiological reflection of skeletal muscle contractile activity that can directly reflect neuromuscular activity. It has been a matter of research to investigate feature extraction methods of sEMG signals. In this letter, we propose a feature extraction method of sEMG signals based on the improved small-world leaky echo state network (ISWLESN). The reservoir of leaky echo state network (LESN) is connected by a random network. First, we improved the reservoir of the echo state network (ESN) by these networks and used edge-added probability to improve these networks. That idea enhances the adaptability of the reservoir, the generalization ability, and the stability of ESN. Then we obtained the output weight of the network through training and used it as features. We recorded the sEMG signals during different activities: falling, walking, sitting, squatting, going upstairs, and going downstairs. Afterward, we extracted corresponding features by ISWLESN and used principal component analysis for dimension reduction. At the end, scatter plot, the class separability index, and the Davies-Bouldin index were used to assess the performance of features. The results showed that the ISWLESN clustering performance was better than those of LESN and ESN. By support vector machine, it was also revealed that the performance of ISWLESN for classifying the activities was better than those of ESN and LESN.