Search

The defense meteorological satellite program- operational line-scan system nighttime stable light (NTL) data have been widely used to evaluate the intensity of human activities. However, the sensor's defects lead to the saturation phenomenon, which greatly limits the reliability of the research results based on NTL data. Thus, this article has attempted to propose a new spectral index - the points of interest, road network and EVI adjusted NTL index (PREANTLI) to effectively correct saturated pixels. To evaluate the desaturation effect, the PREANTLI was compared with existing saturation correction data across three aspects: the capacity to display differences in light intensities; the similarity with National Polar-orbiting Partnership/Visible Infrared Imaging Radiometer Suite (NPP/VIIRS) data; and the capacity to estimate the gross domestic product (GDP) and electric power consumption (EPC). The results showed that the PREANTLI can more easily identify light intensity differences than other indexes. The PREANTLI presents a strong linear correlation with the NPP/VIIRS data and socioeconomic statistics (GDP and EPC) at provincial and municipal scales. It is worth noting that, on the pixel scale, the correlation between PREANTLI and NPP/VIIRS (R2 = 0.63) is far higher than that of the other two existing saturation correction index whose R2 are both below 0.45. Thus, the PREANTLI can be considered as a reasonable index that is not only easy to calculate but can also better alleviate light intensity saturation and emphasize light differences within a city than other indexes.

Autonomous vehicles (AVs) are expected to greatly redefine the future of transportation. However, before people fully realize the benefits of autonomous vehicles, there are still major engineering challenges to be solved. One of the challenges is to build models that reliably predict the movement of the vehicle and its surrounding objects. In this paper, we proposed our ML policy to fully control a Self Driving Vehicle (SDV). The policy is a CNN architecture based on ResNet50 which is invoked by the SDV to obtain the next command to execute. In each step, we predict several different trajectories and their probabilities to assist us in decision-making. Compared with VGG16 and ResNet34, the simulation results demonstrate that our model based on ResN et50 improves the performance by 2.23% and 22.5%, respectively. It also shows that ResNet achieves better performance than VGG in the aspect of motion prediction. What's more, increasing the depth of the network can further improve the performance of the network.

Cancer is a disease that has not been completely overcome in the modern medical field, and skin cancer is one of them. Melanoma in skin cancer is highly lethal. Once the disease is not found in time, it is undoubtedly a disaster for patients. Therefore, early visual detection of melanoma has become a growing concern for researchers. We proposed an online model fusion method based on VGG and DenseNet, which can effectively improve the generalization and performance of the single model. In addition, we analyzed the relationship between the impact of melanoma on different regions of human body, and used data augmentation to effectively augment the dataset to further improve the performance. The experimental results indicate that the AUC-ROC score of data augmentation and online model fusion method on ISIC 2020 challenge dataset is 93.0%, which is higher than that of the single model VGG16 or DenseNet201.

Sleep stage classification is a technique for analyzing sleep quality. Manual sleep staging is time-consuming and laborious. In this paper, we propose an automatic sleep stage classification model combining feature learning and sequence learning, which extract features with convolutional neural network(CNN) and learn the sequence transition rule through multi-layer long short term memory(LSTM) architecture with attention mechanism. In addition, we also noticed that most of the misclassified samples locate in transition period. Therefore, multi-label classification scheme is introduced to provide more label information, so as to improve the classification performance of transition period. We evaluate on two public datasets (Sleep EDF Expanded and Physionet2018), where our framework reaches macro F1-score of 79.7 and 79.8, respectively. The proposed network achieves the state-of-the-art classification performance on Sleep EDF Expanded dataset and sets new benchmark on Physionet2018 dataset.

Epilepsy is a common brain disease that has serious negative effects on patients. Electroencephalogram (EEG) is widely used for detecting epileptic signal. Because epilepsy EEG data set is generally small, in this paper, we use a shallow convolutional neural network (CNN) to classify the raw data. To improve the performance of the model, we propose a layerwise pre-training mechanism. In our experiment, we validate the effectiveness of the method on the public epilepsy EEG data set provided by University of Bonn.

Meditation aims to improve individuals’ core psychological capacities, such as attentional and emotional self-regulation. In this study, we experimented to explore the utility of using permutation entropy (PE) features on electroencephalogram (EEG) data to distinguish among states in human meditation, attention, and relaxation. Twenty advanced yogis with above three years yoga experience and twenty nonmeditators were recruited in our experiment. The whole experiment contained 7 trials. During each trial, subjects were requested to maintain meditation, attention, and relaxation state for three minutes in random order, respectively, after the corresponding voice prompted. In the meantime, 30-channel EEG data were collected in real-time. We calculated PE features from different EEG frequency bands and selected the most correlative features using Fisher’s ratio technique. After that, we fed the feature vectors into a three-class SVM classifier. Twenty yogis and twenty non-meditators achieved average offline classification accuracies of 74.31% and 62.16%, respectively. These results indicated that PE features could be useful in discriminating meditation state and might bring new insight in building meditation-related Brain-Computer Interfaces (BCIs).

This paper presents a new asynchronous hybrid brain-computer interface (BCI) system that integrates a speller, a web browser, an e-mail client, and a file explorer using electroencephalographic (EEG) and electrooculography (EOG) signals. More specifically, an EOG-based button selection method, which requires the user to blink his/her eyes synchronously with the target button's flashes during button selection, is first presented. Next, we propose a mouse control method by combining EEG and EOG signals, in which the left-/right-hand motor imagery (MI)-related EEG is used to control the horizontal movement of the mouse and the blink-related EOG is used to control the vertical movement of the mouse and to select/reject a target. These two methods are further combined to develop the integrated hybrid BCI system. With the hybrid BCI, users can input text, access the internet, communicate with others via e-mail, and manage files in their computer using only EEG and EOG without any body movements. Ten healthy subjects participated in a comprehensive online experiment, and superior performance was achieved compared with our previously developed P300- and MI-based BCI and some other asynchronous BCIs, therefore demonstrating the system's effectiveness.

In this paper, a noise-immune method with light intensity modulation was proposed by using a Mach-Zehnder fiber interferometer. Meanwhile, the fiber interferometer was reused as an etalon to make the sensor configuration more compact. Compared with direct absorption spectroscopy (DAS), the optical path for measuring the absorption spectrum and the optical path for measuring wavenumber were combined. It combined the advantages of measuring the absolute absorbance profile from the DAS method with noise suppression ability of the amplitude modulation. Experiments showed that the new method is more robust than the DAS method in case of background light interference. It provided a new way to detect the actual temperature in a harsh environment, such as backgrounds interrupted by fluctuating thermal radiations.

Weighted ℓ1 minimization (WL1M) is a general and powerful framework for reconstructing sparse signals from underdetermined measurements. The performance improvement of WL1M owes to the incorporation of additional structural priors of signals by means of its weights. However, the selection of weights relies on hand-crafted designs in existing works, so that high-order structural priors of signals are hard to be captured. This paper proposes a data-driven method, namely RBM-WL1M, to alleviate this situation. In the RBM-WL1M, restricted Boltzmann machines (RBMs) are employed to learn the prior distribution of the signals from training data; furthermore, utilizing the RBM, high frequency support set and non-zero probabilities for each of the entries in signals can be estimated effectively, which are used to appropriately select the weights. In our experiments, the proposed framework demonstrates superior performance over several state-of-the-art CS methods on the Physikalisch-Technische Bundesanstalt(PTB) Diagnostic ECG Data set.

Epileptic seizures are caused by excessive, synchronized activity of large groups of neurons. In human electroencephalograph (EEG), they are reflected by multiple epileptic characteristic waves. Based on the idea of template matching, this paper presents a patient-specific approach for the automatic detection of epileptic seizures. In our method, a set of wavelets are constructed based on the epileptic characteristic waves extracted from training EEG signals, and then continuous wavelet transform (CWT) is performed on the recorded EEG. The coefficients of CWT reflect the similarity of the recorded EEG and the epileptic characteristic waveforms and thus can be used to detect if the epileptic characteristic waveforms exist in the EEG. After applying data fusion to the CWT coefficient matrices corresponding to the multiple constructed wavelets, the boundaries of seizures can be determined. In the experiment, our constructed wavelets performed better in the detection of epileptic characteristic waves compared to the Daubechies wavelet. We analyzed the EEG of 10 patients and our method detected 32 out of 34 seizures and declared five false detections. Therefore, our method is promising for the automatic detection of epileptic seizures and the real-time monitoring of patients’ EEG signal.

Deep convolutional neural networks have been used to classify environmental sound recently. The classification system with high performance often requires a large well-labled dataset. The cost of tagging audio segments correctly and completely is quite high thus the deep learning models need to have high generalization ability if a weakly-tagged dataset is used. An algorithm named Weighting Filters algorithm(WF) which can be considered as an improved algorithm based on Dropout is proposed in this paper to enhance the generalization ability of models. To implement the Weighting Filters algorithm, an extra layer trained by backpropagation algorithm is introduced to produce a series of weighted filters. The simulation results show that the Weighting Filters algorithm is an effective way to improve the generalization ability of the model. Further more, a deep convolutional neural network using weighting filters algorithm is proposed for the applications of environmental sound classification. The main contributions of this paper are as follows: First, we proposed an effective algorithm WF based on Dropout, and secondly, we proposed a CNN-based framework using WF(CNN-WF) for environmental sound classification. The results obtained on ESC-50 demonstrate that the CNN-based framework we proposed has considerable performance for environmental sound classification.

Cosmic radiation phenomena such as Solar Particle Events cause high radiation flux lasting from hours to days, thus increasing the probability of Single-Event Upsets (SEUs) for several orders of magnitude. In space applications it is necessary, therefore, to monitor the SEU rate in order to ensure timely detection of high radiation levels and efficient protection of radiation-sensitive circuits. This work proposes an approach combining the SEU monitoring and data storage functions in the same on-chip Static Random Access Memory (SRAM) module, with negligible cost and overheads compared to traditional stand-alone SEU monitors. Furthermore, it also enables the detection of permanent faults in SRAM. The proposed monitor is intended to be further integrated into a highly dependable and self-adaptive multiprocessing platform in which it will drive the selection of the multiprocessor operating modes. Thus, a dynamic trade-off between reliability, performance and power consumption in real-time can be achieved.

Single event transients (SETs) stand out as one of the major causes of soft errors in nanoscale CMOS integrated circuits. To reduce the need for exhaustive circuit simulations in the design of radiation-hard integrated circuits, the cost-effective approaches for characterization and modeling of SET generation effects in standard logic cells are required. In this work, a SPICE-based methodology for characterization of SET generation effects, employing two different SET current models, is presented. Based on the acquired simulation results, the empirical models for the two main SET generation metrics (SET critical charge and SET pulse width) are derived. The SET generation models and the respective model parameters are intended to be used as inputs for the higher-level analysis of SET effects in digital circuits designed with the characterized standard logic cells. By storing the model parameters for each gate in the look-up table, instead of storing the raw data obtained from SPICE simulations, the amount of characterization data can be significantly reduced, allowing to speed up the subsequent SET analysis of a complex circuit.

Many brain decoding algorithms are based on the assumption that the training and test data are in the same feature space and have the same distribution. However, in many practical applications, the statistical distribution of brain signal varies across subjects as well as sessions, restricting the transferability of training data or training models between different subjects or sessions. In this study, we focus on this problem in magnetoencephalography (MEG) signal. By reviewing the transfer learning literatures that have achieved satisfactory results in brain decoding, we proposed a new method for MEG decoding across subjects. First, a new kernel is derived by establishing a connection between the Riemannian geometry and the sample covariance matrix (SCM), which is then applied to MEG transfer decoding by using an improved multi-task learning framework. The experiments on an MEG dataset of face vs. scramble decoding task demonstrated that the proposed approach is superior to other comparable methods.

Since the spatial covariance matrices of EEG lie on a Riemannian manifold, it is potential to exploit Riemannian geometry to decode the motor imagery EEG signal. However, high dimensionality on Riemannian manifold limits the application of existing techniques. In this paper, taking advantage of the locality preserving on Riemannian manifold, we propose a novel dimensionality reduction method, named local isometric embedding (LIE), to learn a low-dimensional embedding from Riemannian manifold. Further with a support vector machine (SVM) classifier performed on the embedding, we obtain a efficient decoding method for motor imagery classification. Experimental evaluation on Dataset IIa of BCI Competition IV reveals that the proposed method outperforms other competing methods.

This paper investigates the applicability of CMOS decoupling cells for mitigating the Single Event Transient (SET) effects in standard combinational gates. The concept is based on the insertion of two decoupling cells between the gate's output and the power/ground terminals. To verify the proposed hardening approach, extensive SPICE simulations have been performed with standard combinational cells designed in IHP's 130 nm bulk CMOS technology. Obtained simulation results have shown that the insertion of decoupling cells results in the increase of the gate's critical charge, thus reducing the gate's soft error rate (SER). Moreover, the decoupling cells facilitate the suppression of SET pulses propagating through the gate. It has been shown that the decoupling cells may be a competitive alternative to gate upsizing and gate duplication for hardening the gates with lower critical charge and multiple (3 or 4) inputs, as well as for filtering the short SET pulses induced by low-LET particles.

Common spatial pattern (CSP) is a spatial filtering algorithm which can be used in motor imagery-based braincomputer interface (BCI) for extracting spatial feature of multivariate signals. But CSP algorithm has inherent drawback, which is that the estimation of the covariance matrices is sensitive to noise, and when few data are available, CSP algorithm is very likely to overfit while it is very time consuming to collect large amount of training data from each task. In this paper, we propose a multitask learning method to extract discriminative subspace shared between subjects and regularize CSP away from the orthogonal complement of this subspace. We compared our method with the standard CSP algorithm on three publicly available datasets of BCI competitions, and an significant performance gain was observed, which therefore demonstrated the advantage of the proposed method.

It is a long-standing challenge to accurately estimate the locations and extents of brain activities from electroencephalography and magnetoencephalography (E/MEG). In this work, a new fully data-driven source imaging method, Source Imaging based on Spatio-Temporal Decompositions (SI-STD), which is built upon a Bayesian framework, is proposed to address this issue. SI-STD decomposes the source matrix as a linear combination of several unknown temporal basis functions (TBFs). The prior of TBFs is set in the empirical Bayesian manner. Additionally, the spatial prior of sources is constructed using the Destrieux atlas and local smoothness constraint to model the globally sparse and locally smooth character of sources. Through variational Bayesian inference techniques, the TBFs and corresponding weight matrix are learned from E/MEG simultaneously. The results of numerical experiments with simulated and experimental E/MEG data show the superior performance of SI-STD in reconstructing extended sources compared with the ?2-norm constrained methods. By virtue of the spatio-temporal decompositions of source matrix, SI-STD also produces better estimations of spatial and temporal profiles than the spatial-only constraint method.

In this paper, an EEG-based brain-computer-interface (BCI) system used for automatic sleep stage classification was proposed. After wavelet de-noising, thirteen features, including spectral and statistical analysis of single-channel EEG signal, were extracted for each epoch. Support Vector Machine was employed to classify four states (Awake/ Light Sleep/ Slow Wave Sleep/ Rapid Eye Movement). Two experiments involving ten recordings from Sleep-EDF Database and the other three healthy subjects in our laboratory were conducted to validate our method and BCI system. For the former, the overall accuracy and kappa coefficient has achieved 86.82% and 0.83 of Fpz-Cz channel, respectively (86.01% and 0.81 for Pz-Oz channel, respectively). For the latter, the overall accuracy and kappa coefficient has achieved 81.30% and 0.73, respectively. The experimental outcomes validate the superior of the proposed method in accuracy.

Currently, the clinical diagnosis of patients with disorders of consciousness (DOC) is based on behavioral assessments, which rely on the patients' behavioral responses to external stimulation. However, high rates of clinical misdiagnosis occurred, because many patients cannot provide efficient physical movements during the evaluation. Brain-computer interfaces (BCIs) provide an alternative approach to assist clinical assessments in patients with DOC. This study proposed a three-dimensional (3D) BCI based on stereo stimuli to assist the clinical object recognition for patients with DOC. Subjects were instructed to selectively focus on the target object cued by instruction. Four healthy subjects participated in two experiment sessions corresponding to the two-dimensional (2D) and 3D modes respectively. The results indicated that the 3D BCI system outperformed the 2D BCI system. This 3D B-CI system was then applied to assess the object recognition of a patient in the minimally conscious state (MCS). The patient cannot recognize objects in the JFK CRS-R assessment before the experiment, but got a score of 75% in the BCI experiment. The result demonstrated his residual awareness and object recognition ability.

In this paper, we present a method for hardening memory and sequential cells against soft errors. The effect of the ionizing particle on the bulk potential is exploited to prevent the induced SET from propagating in the circuit. The proposed method requires a minimum number of extra transistors. The solution is applied to D Flip-Flop design, and alpha and heavy-ions test results are presented.

In this study, we propose a new web browser based on a hybrid brain computer interface (BCI) combining electroencephalographic (EEG) and electrooculography (EOG) signals. Specifically, the user can control the horizontal movement of the mouse by imagining left/right hand motion, and control the vertical movement of the mouse, select/reject a target, or input text in an edit box by blinking eyes in synchrony with the flashes of the corresponding buttons on the GUI. Based on mouse control, target selection and text input, the user can open a web page of interest, select an intended target in the web and read the page content. An online experiment was conducted involving five healthy subjects. The experimental results demonstrated the effectiveness of the proposed method.

Virtual reality (VR) technology allows users to communicate or interact with a virtual artifact in an immersive simulated environment. Most of the existing virtual reality platform uses standard input devices such as a keyboard and mouse, or through multimodal devices such as a wired glove. In this paper, a novel electroencephalo-graph (EEG) based virtual driving (EEG-VD) prototype is proposed, investigated, and designed. Users can enjoy driving a virtual car in a simulated and realistic city only through brain signals. User's EEG signals are obtained, analyzed, and recognized into control commands, which are used to control a virtual car in a three-dimensional (3D) city environment. The paradigm display of P300 in the proposed EEG-VD system is embedded in the display of VR, to realize closely combination of VR and P300. This guarantees the verisimilitude of the system. In addition, asynchronous BCI paradigm based on P300 signals used in the proposed EEG-VD system realizes a high degree of freedom, which can offer appropriate number of control instructions. In order to verify the effectiveness and performance of the designed EEG-VD system, eight volunteers are asked to drive the car in the 3D city. The experiment result shows that the accuracy of P300 recognition rate is larger than 90%. Most of the users (62.5%) can drive the car without any collisions in the first virtual environment, and all users can drive the car without any collisions in the second virtual environment. The designed system can not only be used as a tool to train a subject of BCI but also be used as a game for enjoying life. Potentially, the system can promote the research of BCI applications.

Visual fixation is an item of the Coma Recovery Scale-Revised (CRS-R), it is difficult to be detected by clinicians using the behavioral scales because of fluctuations of arousal level and the presence of motor impairment in disorders of consciousness (DOC) patients. Brain-computer interfaces (BCIs), which directly detect brain response without any behavioral expression, can be used to evaluate a patient's response to the stimuli. In this study, we explored the application of a 4-choice visual BCI in assisting visual fixation assessment of DOC patients. The results obtained from three patients indicated that one locked-in syndrome (LIS) patient showed visual fixation with the BCI assessment while exhibit higher visual function with the CRS-R assessment. The other two patients did not show any response to visual fixation with BCI assessment. The result revealed that EEG based BCI could detect brain response to the evaluation item of CRS-R in DOC patients. Therefore, the proposed BCI may provide a promising method for more objective result and thus assist CRS-R behavioral assessment.

The aim of this study is to determine the potential prognostic value of using Brain-computer Interface (BCI) to identify patients with disorder of consciousness (DOC), who show potential for recovery. A retrospective study involved 51 patients with DOC were conducted. Each patient conducted in a BCI experiment to detect awareness and received a 3-months follow-up. The BCI accuracies were correlated to patient outcomes according to the Coma Recovery Scale Revised (CRS-R). The statistical tests showed that patients with significant accuracies higher than the chance level showed greater improvement in CRS-R scores after 3 months compared to patients without significant accuracies. The sensitivity and specificity of the BCI accuracies in determining individual's consciousness recovery after 3 months are 67.7% and 90% respectively. A highly significant relationship between BCI accuracies and subsequent recovery was thus found. The BCI is suggested as an important tool to assess information-processing capacities that can predict the likelihood of recovery in patients with disorder of consciousness.

In this paper, a novel electroencephalographic (EEG) based mind controlled virtual-human obstacle-avoidance platform (EEG-MC-VHOAP) is designed to improve brain computer interface (BCI) systems and offer a new game. With the EEG-MC-VHOAP, subjects can use their brain signals to control a virtual human to have a training of avoiding obstacles in a three dimensional (3D) environment. The EEG-MC-VHOAP is composed of an EEG based BCI subsystem and a 3D virtual-human subsystem. In the EEG-based BCI subsystem, a self-adaptive bayesian linear discriminant analysis(SA-BLDA) is adopted to classify the P300 signals, and is then transformed into four control commands. The control commands are used to control the virtual-human to walk forward, walk in a crouch, turn left and turn right. Three subjects were asked to attend the testing with the EEG-MC-VHOAP. All subjects accessed to a 100% train accuracy for repeating flashing 4–6 trails, a less than 10% average collision rate, and a higher than 80% online accuracy. Both the training and online testing results demonstrate the effectiveness of the proposed EEG-MC-VHOAP.

In this paper, we study the performance guarantee of weighted l 1 -constrained quadratic programming in recovering the support of a sparse signal from a few linear measurements. A new sufficient condition for the success of weighted l 1 -constrained quadratic programming is derived. Further, we demonstrate its applications in two typical weighted l 1 models: Firstly, a theoretical result on the modified-BPDN can be generated directly, which indicates that if a partial support information is available and more than half of the information is accurate, then modified-BPDN relaxes the sufficient condition of BPDN ensuring the support recovery for any sparse signals. Secondly, the proposed condition offers a criterion to compute the optimal weights of weighted l 1 minimization for nonuniform sparse models. Some simulations are carried out to validate its effectiveness.

In this paper, an EEG-based brain-computer interface (BCI) system used for emotion recognition is proposed to detect two basic emotional states (happiness and sadness). Selection of frequency bands plays a vital role in distinguishing brain patterns associated with emotions. This paper explores a new method to select suitable subject-specific frequency bands instead of using fixed frequency bands for the emotion recognition. Common spatial pattern and support vector machine were employed to classify two emotional states. Two experiments involving six subjects were conducted to validate our method and BCI system. An average online accuracy of 74.17% for two classes was achieved. The data analysis results demonstrated that the proposed method based on subject-specific frequency bands outperformed the method based on the fixed frequency bands in terms of accuracy.

P300, which is usually evoked by visual stimulus, is widely used in electroencephalography (EEG) based brain computer interface (BCI) studies. As an application-oriented BCI study, the P300 speller would inevitably be used in outdoor environments. However, the visual stimulation effect might be interfered by the reflections in outdoor environment. This paper attempted to improve the outdoor P300 speller performance by using polarizer. Six subjects participated our two experiments, each experiment was conducted three times under three different conditions. The results demonstrated that a suitable polarizer can indeed improve the performance of outdoor P300 speller.

Single Event Transients (SETs) are a growing concern in advanced integrated circuits yet techniques to accurately characterize the cross-section and pulse width of SETs are less mature than those for measuring SEUs. We present four circuits for measuring SETs, an analysis of their capabilities and the subtleties in their implementation. Post-layout circuit simulation results are presented for a test-chip implemented in 28 nm FSDOI technology and integrating these detectors.

In this paper, we presented two hybrid brain computer interfaces (BCIs), one combing motor imagery (MI) and P300 and another combing P300 and steady state visual evoked potential (SSVEP), and their applications. An important issue in BCI research is multidimensional control. Potential applications include BCI controlled computer mouse, document and email processing, web browser, wheelchair and neuroprosthesis. The challenge for EEG-based multidimensional control is to obtain multiple independent control signals from the noisy EEG data. For this purpose, hybrid BCIs may yield better performance than BCIs those use only one type of brain pattern. In this project, we first developed a hybrid system for 2-D cursor control. In our system, two independent signals based on MI and P300 were produced from EEG for the vertical and horizontal movement control of the cursor respectively, and the cursor can be moved from an arbitrary initial position to a randomly given target position. Furthermore, a hybrid feature was extracted for selecting the target-of-interest and rejecting the target-of-no-interest, as fast and accurate as possible. In this way, a BCI mouse was implemented. Then an internet browser and a mail client were developed based on the BCI mouse. Moreover, we extended this hybrid BCI system to control a virtual car/wheelchair. On the other hand, we also developed a P300 and SSVEP-based hybrid BCI not only to improve the classification performance, but also to validate the possibility of clinical applications, e.g., detection of residual cognitive function and covert awareness in patients with disorders of consciousness.

In recent years, various applications which apply the hybrid brain computer interfaces (BCIs) have been studied. In this paper, we present a hybrid BCI system to operate the explorer with P300 and motor imagery, which is mainly composed of a BCI mouse, a BCI speller and an explorer. Through this system, the user can access to his computer and manipulate (open, close, copy, paste, delete) files such as documents, pictures, music, movies and so on. The system has been tested with 5 subjects, and the experimental results show that the explorer can be successfully operated according to subjects' intention with only a small number of mistakes.

Common Spatial Patterns (CSP) is among the most popular spatial filtering methods used in motor imagery based brain-computer interfaces (BCIs). In this paper, a novel algorithm termed spatio-temporally regularized common spatial patterns (STR-CSP) is proposed to enhance CSP such that discriminative spatio-temporal EEG patterns can be extracted. Unlike most previous CSP-based spatio-temporal filtering algorithms in which spatial and temporal filters are optimized in a sequential manner, STR-CSP optimizes spatio-temporal filters simultaneously under a unified criterion with analytical solutions. Moreover, the algorithm enables us to alleviate the over-fitting problem by L2-norm regularization. Successful application to the analysis of a BCI competition EEG data set comprising 18 subsets demonstrates the superiority of STR-CSP over contemporary methods.