Unsupervised heterogeneous domain adaptation for EEG classification.

Objective.Domain adaptation has been recognized as a potent solution to the challenge of limited training data for electroencephalography (EEG) classification tasks. Existing studies primarily focus on homogeneous environments, however, the heterogeneous properties of EEG data arising from device diversity cannot be overlooked. This motivates the development of heterogeneous domain adaptation methods that can fully exploit the knowledge from an auxiliary heterogeneous domain for EEG classification.Approach.In this article, we propose a novel model named informative representation fusion (IRF) to tackle the problem of unsupervised heterogeneous domain adaptation in the context of EEG data. In IRF, we consider different perspectives of data, i.e. independent identically distributed (iid) and non-iid, to learn different representations. Specifically, from the non-iid perspective, IRF models high-order correlations among data by hypergraphs and develops hypergraph encoders to obtain data representations of each domain. From the non-iid perspective, by applying multi-layer perceptron networks to the source and target domain data, we achieve another type of representation for both domains. Subsequently, an attention mechanism is used to fuse these two types of representations to yield informative features. To learn transferable representations, the maximum mean discrepancy is utilized to align the distributions of the source and target domains based on the fused features.Main results.Experimental results on several real-world datasets demonstrate the effectiveness of the proposed model.Significance.This article handles an EEG classification situation where the source and target EEG data lie in different spaces, and what's more, under an unsupervised learning setting. This situation is practical in the real world but barely studied in the literature. The proposed model achieves high classification accuracy, and this study is important for the commercial applications of EEG-based BCIs.

Reendothelialization of Acellular Adipose Flaps under Mimetic Physiological Dynamic Conditions.

The extensive soft-tissue defects resulting from trauma and tumors pose a prevalent challenge in clinical practice, characterized by a high incidence rate. Autologous tissue flap transplantation, considered the gold standard for treatment, is associated with various drawbacks, including the sacrifice of donor sources, postoperative complications, and limitations in surgical techniques, thereby impeding its widespread applicability. The emergence of tissue-engineered skin flaps, notably the acellular adipose flap (AAF), offers potential alternative solutions. However, a critical concern confronting large-scale tissue-engineered skin flaps currently revolves around the reendothelialization of internal vascular networks. In our study, we have developed an AAF utilizing perfusion decellularization, demonstrating excellent physical properties. Cytocompatibility experiments have confirmed its cellular safety, and cell adhesion experiments have revealed spatial specificity in facilitating endothelial cells adhesion within the adipose flap scaffold. Using a novel mimetic physiological fluid shear stress setting, endothelial cells were dynamically inoculated and cultured within the acellular vascular network of the pedicled AAF in our research. Histological and gene expression analyses have shown that the mimetic physiological fluid dynamic model significantly enhanced the reendothelialization of the AAF. This innovative platform of acellular adipose biomaterials combined with hydrodynamics may offer valuable insights for the design and manufacturing of 3D vascularized tissue constructs, which can be applied to the repair of extensive soft-tissue defects.

Beta rebound reduces subsequent movement preparation time by modulating of GABAA inhibition.

Post-movement beta synchronization is an increase of beta power relative to baseline, which commonly used to represent the status quo of the motor system. However, its functional role to the subsequent voluntary motor output and potential electrophysiological significance remain largely unknown. Here, we examined the reaction time of a Go/No-Go task of index finger tapping which performed at the phases of power baseline and post-movement beta synchronization peak induced by index finger abduction movements at different speeds (ballistic/self-paced) in 13 healthy subjects. We found a correlation between the post-movement beta synchronization and reaction time that larger post-movement beta synchronization prolonged the reaction time during Go trials. To probe the electrophysiological significance of post-movement beta synchronization, we assessed intracortical inhibitory measures probably involving GABAB (long-interval intracortical inhibition) and GABAA (short-interval intracortical inhibition) receptors in beta baseline and post-movement beta synchronization peak induced by index finger abduction movements at different speeds. We found that short-interval intracortical inhibition but not long-interval intracortical inhibition increased in post-movement beta synchronization peak compared with that in the power baseline, and was negatively correlated with the change of post-movement beta synchronization peak value. These novel findings indicate that the post-movement beta synchronization is related to forward model updating, with high beta rebound predicting longer time for the preparation of subsequent movement by inhibitory neural pathways of GABAA.

Altered functional connectivity in the hippocampal and striatal systems after motor sequence learning consolidation in medial temporal lobe epilepsy individuals.

Both the hippocampal and striatal systems participate in motor sequence learning (MSL) in healthy subjects, and the prominent role of the hippocampal system in sleep-related consolidation has been demonstrated. However, some pathological states may change the functional dominance between these two systems in MSL consolidation. To better understand the functional performance within these two systems under the pathological condition of hippocampal impairment, we compared the functional differences after consolidation between patients with left medial temporal lobe epilepsy (LmTLE) and healthy control subjects (HCs). We assessed participants' performance on the finger-tapping task (FTT) during acquisition (on day 1) and after consolidation during sleep (on day 2). All participants underwent an MRI scan (T1 and resting state) before each FTT. We found that the LmTLE group showed performance deficits in offline consolidation compared to the HC group. The LmTLE group exhibited structural changes, such as decreased gray matter volume (GMV) in the left hippocampus and increased GMV in the right putamen (striatum). Our results also revealed that whereas the main effect of consolidation was observed in the hippocampus-related functional connection in the HC group, it was only evident in the striatum-related functional loop in the LmTLE group. Our findings indicated that LmTLE patients may rely more on the striatal system for offline consolidation because of structural impairments in the hippocampus. Additionally, this compensatory mechanism may not fully substitute for the role of the impaired hippocampus itself.NEW & NOTEWORTHY Motor sequence learning (MSL) relies on both the hippocampal and striatal systems, but whether functional performance is altered after MSL consolidation when the hippocampus is impaired remains unknown. Our results indicated that whereas the main effect of consolidation was observed in the hippocampus-related functional connection in the healthy control (HC) group, it was only evident in the striatum-related functional loop in the left medial temporal lobe epilepsy (LmTLE) group.

Diagnosis Framework for Probable Alzheimer's Disease and Mild Cognitive Impairment Based on Multi-Dimensional Emotion Features.

BACKGROUND: Emotion and cognition are intercorrelated. Impaired emotion is common in populations with Alzheimer's disease (AD) and mild cognitive impairment (MCI), showing promises as an early detection approach. OBJECTIVE: We aim to develop a novel automatic classification tool based on emotion features and machine learning. METHODS: Older adults aged 60 years or over were recruited among residents in the long-term care facilities and the community. Participants included healthy control participants with normal cognition (HC, n = 26), patients with MCI (n = 23), and patients with probable AD (n = 30). Participants watched emotional film clips while multi-dimensional emotion data were collected, including mental features of Self-Assessment Manikin (SAM), physiological features of electrodermal activity (EDA), and facial expressions. Emotional features of EDA and facial expression were abstracted by using continuous decomposition analysis and EomNet, respectively. Bidirectional long short-term memory (Bi-LSTM) was used to train classification model. Hybrid fusion was used, including early feature fusion and late decision fusion. Data from 79 participants were utilized into deep machine learning analysis and hybrid fusion method. RESULTS: By combining multiple emotion features, the model's performance of AUC value was highest in classification between HC and probable AD (AUC = 0.92), intermediate between MCI and probable AD (AUC = 0.88), and lowest between HC and MCI (AUC = 0.82). CONCLUSIONS: Our method demonstrated an excellent predictive power to differentiate HC/MCI/AD by fusion of multiple emotion features. The proposed model provides a cost-effective and automated method that can assist in detecting probable AD and MCI from normal aging.

Cortical Contributions to Imagined Power Grip Task: An EEG-Triggered TMS Study.

Previous studies have demonstrated that motor imagery leads to desynchronization in the alpha rhythm within the contralateral primary motor cortex. However, the underlying electrophysiological mechanisms responsible for this desynchronization during motor imagery remain unclear. To examine this question, we conducted an investigation using EEG in combination with noninvasive transcranial magnetic stimulation (TMS) during index finger abduction (ABD) and power grip imaginations. The TMS was administered employing diverse coil orientations to selectively stimulate corticospinal axons, aiming to target both early and late synaptic inputs to corticospinal neurons. TMS was triggered based on the alpha power levels, categorized in 20th percentile bins, derived from the individual alpha power distribution during the imagined tasks of ABD and power grip. Our analysis revealed negative correlations between alpha power and motor evoked potential (MEP) amplitude, as well as positive correlations with MEP latency across all coil orientations for each imagined task. Furthermore, we conducted functional network analysis in the alpha band to explore network connectivity during imagined index finger abduction and power grip tasks. Our findings indicate that network connections were denser in the fronto-parietal area during imagined ABD compared to power grip conditions. Moreover, the functional network properties demonstrated potential for effectively classifying between these two imagined tasks. These results provide functional evidence supporting the hypothesis that alpha oscillations may play a role in suppressing MEP amplitude and latency during imagined power grip. We propose that imagined ABD and power grip tasks may activate different populations and densities of axons at the cortical level.

A novel noninvasive brain-computer interface by imagining isometric force levels.

Physiological circuits differ across increasing isometric force levels during unilateral contraction. Therefore, we first explored the possibility of predicting the force level based on electroencephalogram (EEG) activity recorded during a single trial of unilateral 5% or 40% of maximal isometric voluntary contraction (MVC) in right-hand grip imagination. Nine healthy subjects were involved in this study. The subjects were required to randomly perform 20 trials for each force level while imagining a right-hand grip. We proposed the use of common spatial patterns (CSPs) and coherence between EEG signals as features in a support vector machine for force level prediction. The results showed that the force levels could be predicted through single-trial EEGs while imagining the grip (mean accuracy = 81.4 ± 13.29%). Additionally, we tested the possibility of online control of a ball game using the above paradigm through unilateral grip imagination at different force levels (i.e., 5% of MVC imagination and 40% of MVC imagination for right-hand movement control). Subjects played the ball games effectively by controlling direction with our novel BCI system (n = 9, mean accuracy = 76.67 ± 9.35%). Data analysis validated the use of our BCI system in the online control of a ball game. This information may provide additional commands for the control of robots by users through combinations with other traditional brain-computer interfaces, e.g., different limb imaginations.

Transferable multi-modal fusion in knee angles and gait phases for their continuous prediction.

Objective.The gait phase and joint angle are two essential and complementary components of kinematics during normal walking, whose accurate prediction is critical for lower-limb rehabilitation, such as controlling the exoskeleton robots. Multi-modal signals have been used to promote the prediction performance of the gait phase or joint angle separately, but it is still few reports to examine how these signals can be used to predict both simultaneously.Approach.To address this problem, we propose a new method named transferable multi-modal fusion (TMMF) to perform a continuous prediction of knee angles and corresponding gait phases by fusing multi-modal signals. Specifically, TMMF consists of a multi-modal signal fusion block, a time series feature extractor, a regressor, and a classifier. The multi-modal signal fusion block leverages the maximum mean discrepancy to reduce the distribution discrepancy across different modals in the latent space, achieving the goal of transferable multi-modal fusion. Subsequently, by using the long short-term memory-based network, we obtain the feature representation from time series data to predict the knee angles and gait phases simultaneously. To validate our proposal, we design an experimental paradigm with random walking and resting to collect data containing multi-modal biomedical signals from electromyography, gyroscopes, and virtual reality.Main results.Comprehensive experiments on our constructed dataset demonstrate the effectiveness of the proposed method. TMMF achieves a root mean square error of0.090±0.022s in knee angle prediction and a precision of83.7±7.7% in gait phase prediction.Significance.We demonstrate the feasibility and validity of using TMMF to predict lower-limb kinematics continuously from multi-modal biomedical signals. This proposed method represents application potential in predicting the motor intent of patients with different pathologies.

Effect of high-frequency repetitive transcranial magnetic stimulation over M1 for consciousness recovery after traumatic brain injury.

BACKGROUND: The brain area stimulated during repetitive transcranial magnetic stimulation (rTMS) treatment is important in altered states of consciousness. However, the functional contribution of the M1 region during the treatment of high-frequency rTMS remains unclear. OBJECTIVE: The aim of this study was to examine the clinical [the Glasgow coma scale (GCS) and the coma recovery scale-revised (CRS-R)] and neurophysiological (EEG reactivity and SSEP) responses in vegetative state (VS) patients following traumatic brain injury (TBI) before and after a protocol of high-frequency rTMS over the M1 region. METHODS: Ninety-nine patients in a VS following TBI were recruited so that their clinical and neurophysiological responses could be evaluated in this study. These patients were randomly allocated into three experimental groups: rTMS over the M1 region (test group; n = 33), rTMS over the left dorsolateral prefrontal cortex (DLPFC) (control group; n = 33) and placebo rTMS over the M1 region (placebo group; n = 33). Each rTMS treatment lasted 20 min and was carried out once a day. The duration of this protocol was a month with 20 treatments (5 times per week) occurring with that time. RESULTS: We found that the clinical and neurophysiological responses improved after treatment in the test, control, and placebo groups; the improvement was highest in the test group compared to that in the control and placebo groups. CONCLUSIONS: Our results demonstrate an effective method of high-frequency rTMS over the M1 region for consciousness recovery after severe brain injury.

Graph-Based Class-Imbalance Learning With Label Enhancement.

Class imbalance is a common issue in the community of machine learning and data mining. The class-imbalance distribution can make most classical classification algorithms neglect the significance of the minority class and tend toward the majority class. In this article, we propose a label enhancement method to solve the class-imbalance problem in a graph manner, which estimates the numerical label and trains the inductive model simultaneously. It gives a new perspective on the class-imbalance learning based on the numerical label rather than the original logical label. We also present an iterative optimization algorithm and analyze the computation complexity and its convergence. To demonstrate the superiority of the proposed method, several single-label and multilabel datasets are applied in the experiments. The experimental results show that the proposed method achieves a promising performance and outperforms some state-of-the-art single-label and multilabel class-imbalance learning methods.

Multi-Source Transfer Learning for EEG Classification Based on Domain Adversarial Neural Network.

Electroencephalogram (EEG) classification has attracted great attention in recent years, and many models have been presented for this task. Nevertheless, EEG data vary from subject to subject, which may lead to the performance of a classifier degrades due to individual differences. To collect enough labeled data to model would address the issue, but it is often time-consuming and labor-intensive. In this paper, we propose a new multi-source transfer learning method based on domain adversarial neural network for EEG classification. Specifically, we design a domain adversarial neural network, which includes a feature extractor, a classifier, and a domain discriminator, and therefore reduce the domain shift to achieve the purpose. In addition, a unified multi-source optimization framework is constructed to further improve the performance, and the result for EEG classification is induced by the weighted combination of the predictions from multiple source domains. Experiments on three publicly available EEG datasets validate the advantages of the proposed method.

Novel Approach to Transfer of Coordination Skills to the Paretic Hand by Improving the Efficiency of Cortical Network.

We have previously shown that healthy subjects can transfer coordination skills to the unpracticed hand by performing a unimanual task with the other hand and visualizing a bimanual action using a game-like interactive system. However, whether this system could be used to transfer coordination skills to the paretic hand after stroke and its underlying neural mechanism remain unknown. Here, using a game-like interactive system for visualization during physical practice in an immersive virtual reality environment, we examined coordination skill improvement in the unpracticed/paretic hand after training in 10 healthy subjects and 13 chronic and sub-acute stroke patients. The bimanual movement task was defined as simultaneously drawing non-symmetric three-sided squares (e.g., U and C), while the training strategy was performing a unimanual task with the right/nonparetic hand and visualizing a bimanual action. We found large decreases in the intra-hand temporal and spatial measures for movement in the unpracticed/paretic hand after training. Furthermore, a substantial reduction in the inter-hand temporal and spatial interference was observed after training. Additionally, we examined the related cortical network evolution using EEG in both the healthy subjects and stroke patients. Our studies show that the cortical network became more efficient after training in the healthy subjects and stroke patients. These results demonstrate that our proposed method could contribute to the transference of coordination skill to the paretic/unpracticed hand by promoting the efficiency of cortical networks.

Gating at cortical level contributes to auditory-motor synchronization during repetitive finger tapping.

Sensory integration contributes to temporal coordination of the movement with external rhythms. How the information flowing of sensory inputs is regulated with increasing tapping rates and its function remains unknown. Here, somatosensory evoked potentials to ulnar nerve stimulation were recorded during auditory-cued repetitive right-index finger tapping at 0.5, 1, 2, 3, and 4 Hz in 13 healthy subjects. We found that sensory inputs were suppressed at subcortical level (represented by P14) and primary somatosensory cortex (S1, represented by N20/P25) during repetitive tapping. This suppression was decreased in S1 but not in subcortical level during fast repetitive tapping (2, 3, and 4 Hz) compared with slow repetitive tapping (0.5 and 1 Hz). Furthermore, we assessed the ability to analyze temporal information in S1 by measuring the somatosensory temporal discrimination threshold (STDT). STDT increased during fast repetitive tapping compared with slow repetitive tapping, which was negatively correlated with the task performance of phase shift and positively correlated with the peak-to-peak amplitude (% of resting) in S1 but not in subcortical level. These novel findings indicate that the increased sensory input (lower sensory gating) in S1 may lead to greater temporal uncertainty for sensorimotor integration dereasing the performance of repetitive movement during increasing tapping rates.

Fast Multilabel Feature Selection via Global Relevance and Redundancy Optimization.

Information theoretical-based methods have attracted a great attention in recent years and gained promising results for multilabel feature selection (MLFS). Nevertheless, most of the existing methods consider a heuristic way to the grid search of important features, and they may also suffer from the issue of fully utilizing labeling information. Thus, they are probable to deliver a suboptimal result with heavy computational burden. In this article, we propose a general optimization framework global relevance and redundancy optimization (GRRO) to solve the learning problem. The main technical contribution in GRRO is a formulation for MLFS while feature relevance, label relevance (i.e., label correlation), and feature redundancy are taken into account, which can avoid repetitive entropy calculations to obtain a global optimal solution efficiently. To further improve the efficiency, we extend GRRO to filter out inessential labels and features, thus facilitating fast MLFS. We call the extension as GRROfast, in which the key insights are twofold: 1) promising labels and related relevant features are investigated to reduce ineffective calculations in terms of features, even labels and 2) the framework of GRRO is reconstructed to generate the optimal result with an ensemble. Moreover, our proposed algorithms have an excellent mechanism for exploiting the inherent properties of multilabel data; specifically, we provide a formulation to enhance the proposal with label-specific features. Extensive experiments clearly reveal the effectiveness and efficiency of our proposed algorithms.

Subject adaptation convolutional neural network for EEG-based motor imagery classification.

Objective.Deep transfer learning has been widely used to address the nonstationarity of electroencephalogram (EEG) data during motor imagery (MI) classification. However, previous deep learning approaches suffer from limited classification accuracy because the temporal and spatial features cannot be effectively extracted.Approach.Here, we propose a novel end-to-end deep subject adaptation convolutional neural network (SACNN) to handle the problem of EEG-based MI classification. Our proposed model jointly optimizes three modules, i.e. a feature extractor, a classifier, and a subject adapter. Specifically, the feature extractor simultaneously extracts the temporal and spatial features from the raw EEG data using a parallel multiscale convolution network. In addition, we design a subject adapter to reduce the feature distribution shift between the source and target subjects by using the maximum mean discrepancy. By minimizing the classification loss and the distribution discrepancy, the model is able to extract the temporal-spatial features to the prediction of a new subject.Main results.Extensive experiments are carried out on three EEG-based MI datasets, i.e. brain-computer interface (BCI) competition IV dataset IIb, BCI competition III dataset IVa, and BCI competition IV dataset I, and the average accuracy reaches to 86.42%, 81.71% and 79.35% on the three datasets respectively. Furthermore, the statistical analysis also indicates the significant performance improvement of SACNN.Significance.This paper reveals the importance of the temporal-spatial features on EEG-based MI classification task. Our proposed SACNN model can make fully use of the temporal-spatial information to achieve the purpose.

Behavior-Dependent Corticocortical Contributions to Imagined Grasping: a BCI-triggered TMS study.

Previous studies have indicated that corticocortical neural mechanisms differ during various grasping behaviors. However, the literature rarely considers corticocortical contributions to various imagined grasping behaviors. To address this question, we examine their mechanisms by transcranial magnetic stimulation (TMS) triggered when detecting event-related desynchronization during right-hand grasping behavior imagination through a brain-computer interface (BCI) system. Based on the BCI system, we designed two experiments. In Experiment 1, we explored differences in motor evoked potentials (MEPs) between power grip and resting conditions. In Experiment 2, we used the three TMS coil orientations (lateral-medial (LM), posterior-anterior (PA), and anterior-posterior (AP) directions) over the primary motor cortex to elicit MEPs during imagined index finger abduction, precision grip, and power grip. We found that larger MEP amplitudes and shorter latencies were obtained in imagined power grip than in resting.We also detected lower MEP amplitudes during imagined power grip, while MEP amplitudes remained similar across imagined precision grip and index finger abduction in each TMS coil orientation. Differences in AP-LM latency were longer when subjects imagined a power grip compared with precision grip and index finger abduction. Based on our results, higher cortical excitability may be achieved when humans imagine precision grip and index finger abduction. Our results suggests that higher cortical excitability may be achieved when humans imagine precision grip and index finger abduction. We also propose that preferential recruitment of late synaptic inputs to corticospinal neurons may occur when humans imagine a power grip.

A novel upper-limb tracking system in a virtual environment for stroke rehabilitation.

BACKGROUND: The transfer of the behaviors of a human's upper limbs to an avatar is widely used in the field of virtual reality rehabilitation. To perform the transfer, movement tracking technology is required. Traditionally, wearable tracking devices are used for tracking; however, these devices are expensive and cumbersome. Recently, non-wearable upper-limb tracking solutions have been proposed, which are less expensive and more comfortable. However, most products cannot track the upper limbs, including the arms and all the fingers at the same time, which limits the limb parts for tracking in a virtual environment and may lead to a limited rehabilitation effect. METHODS: In this paper, a novel virtual reality rehabilitation system (VRRS) was developed for upper-limb rehabilitation. The VRRS could track the motion of both upper limbs, integrate fine finger motion and the range of motion of the entire arm and map the motion to an avatar. To test the performance of VRRS, two experiments were designed. In the first experiment, we investigated the effect of VRRS on virtual body ownership, agency and location of the body and usability in 8 healthy participants by comparing it with a partial upper-limb tracking method based on a Leap Motion controller (LP) in the same virtual environments. In the second experiment, we examined the feasibility of VRRS in upper-limb rehabilitation with 27 stroke patients. RESULTS: VRRS improved the users' senses of body ownership, agency, and location of the body. The users preferred using the VRRS to using the LP. In addition, we found that although the upper limb motor function of patients from all groups was improved, the difference between the FM scores tested on the first day and the last day of the experimental group was more significant than that of the control groups. CONCLUSIONS: A VRRS with motion tracking of the upper limbs and avatar control including the arms and all the fingers was developed. It resulted in an improved user experience of embodiment and effectively improved the effects of upper limb rehabilitation in stroke patients. TRIAL REGISTRATION: The study was registered at the First Affiliated Hospital of Jinan University Identifier: KY-2020-036; Date of registration: June 01, 2020.

Movement speed effects on beta-band oscillations in sensorimotor cortex during voluntary activity.

Beta-band oscillations are a dominant feature in the sensorimotor system, which includes movement-related beta desynchronization (MRBD) during the preparation and execution phases of movement and postmovement beta synchronization (PMBS) on movement cessation. Many studies have linked this rhythm to motor functions. However, its associations to the movement speed are still unclear. We make a hypothesis that PMBS will be modulated with increasing of movement speeds. We assessed the MRBD and PMBS during isotonic slower self-paced and ballistic movements with 15 healthy subjects. Furthermore, we conduct an additional control experiment with the isometric contraction with two levels of forces to match those in the isotonic slower self-paced and ballistic movements separately. We found that the amplitude of PMBS but not MRBD in motor cortex is modulated by the speed during voluntary movement. PMBS was positively correlated with movement speed and acceleration through the partial correlation analysis. However, there were no changes in the PMBS and MRBD during the isometric contraction with two levels of forces. These results demonstrate a different function of PMBS and MRBD to the movement speed during voluntary activity and suggest that the movement speed would affect the amplitude of PMBS.NEW & NOTEWORTHY Beta-band oscillations are a dominant feature in the sensorimotor system that associate to the motor function. We found that the movement-related postmovement beta synchronization (PMBS) over the contralateral sensorimotor cortex was positively correlated with the speed of a voluntary movement, but the movement-related beta desynchronization (MRBD) was not. Our results show a differential response of the PMBS and MRBD to the movement speed during voluntary movement.

Alpha Frequency Intervention by Electrical Stimulation to Improve Performance in Mu-Based BCI.

The accuracy of brain-computer interfaces (BCIs) is important for effective communication and control. The mu-based BCI is one of the most widely used systems, of which the related methods to improve users' accuracy are still poorly studied, especially for the BCI illiteracy. Here, we examined a way to enhance mu-based BCI performance by electrically stimulating the ulnar nerve of the contralateral wrist at the alpha frequency (10 Hz) during left- and right-hand motor imagination in two BCI groups (literate and illiterate). We demonstrate that this alpha frequency intervention enhances the classification accuracy between left- and right-hand motor imagery from 66.41% to 81.57% immediately after intervention and to 75.28% two days after intervention in the BCI illiteracy group, while classification accuracy improves from 82.12% to 91.84% immediately after intervention and to 89.03% two days after intervention in the BCI literacy group. However, the classification accuracy did not change before and after the sham intervention (no electrical stimulation). Furthermore, the ERD on the primary sensorimotor cortex during left- or right-hand motor imagery tasks was more visible at the mu-rhythm (8-13 Hz) after alpha frequency intervention. Alpha frequency intervention increases the mu-rhythm power difference between left- and right-hand motor imagery tasks. These results provide evidence that alpha frequency intervention is an effective way to improve BCI performance by regulating the mu-rhythm which might provide a way to reduce BCI illiteracy.

A cosparse analysis model with combined redundant systems for MRI reconstruction.

PURPOSE: Magnetic resonance imaging (MRI) is widely used due to its noninvasive and nonionizing properties. However, MRI requires a long scanning time. In this paper, our goal is to reconstruct a high-quality MR image from its sampled k-space data to accelerate the data acquisition in MRI. METHODS: We propose a cosparse analysis model with combined redundant systems to fully exploit the sparsity of MR images. Two fixed redundant systems are used to characterize different structures, namely, the wavelet tight frame and Gabor frame. An alternating iteration scheme is used for reconstruction with simple implementation and good performance. RESULTS: The proposed method is tested on two MR images under three sampling patterns with sampling ratios ranging from 10% to 60%. The results show that the proposed method outperforms other state-of-the-art MRI reconstruction methods in terms of both subjective visual quality and objective quantitative measurement. For instance, for brain images under random sampling with a ratio of 10%, compared to the other three methods, the proposed method improves the peak signal-to-noise ratio (PSNR) by more than 9 dB. CONCLUSIONS: To better characterize different sparsities of different structures of MRI, a cosparse analysis model combining the wavelet tight frame and Gabor frame is proposed. A partial ℓ 2 norm regularization is leveraged to obtain the optimal solution in a lower dimension. Compared to other state-of-the-art MRI reconstruction methods, the proposed method improves the reconstruction quality of MRI, especially highly undersampled MRI.