The motor imagery-based brain-computer interface (BCI) using electroencephalography (EEG) has been receiving attention from neural engineering researchers and is being applied to various rehabilitation applications. However, the performance degradation caused by motor imagery EEG with very low single-to-noise ratio faces several application issues with the use of a BCI system. In this paper, we propose a novel motor imagery classification scheme based on the continuous wavelet transform and the convolutional neural network. Continuous wavelet transform with three mother wavelets is used to capture a highly informative EEG image by combining time-frequency and electrode location. A convolutional neural network is then designed to both classify motor imagery tasks and reduce computation complexity. The proposed method was validated using two public BCI datasets, BCI competition IV dataset 2b and BCI competition II dataset III. The proposed methods were found to achieve improved classification performance compared with the existing methods, thus showcasing the feasibility of motor imagery BCI.

Abstract. Web services utilize a standard communication infrastructure such as XML and SOAP to communicate through the Internet. Even though Web services are becoming more and more widespread as an emerging technology, it is hard to test Web services because they are distributed applications with numerous aspects of runtime behavior that are different from typical applications. This paper presents a new approach to testing Web services based on EFSM (Extended Finite State Machine). WSDL (Web Services Description Language) file alone does not provide dynamic behavior information. This problem can be overcome by augmenting it with a behavior specification of the service. Rather than domain partitioning or perturbation techniques, we choose EFSM because Web services have control flow as well as data flow like communication protocols. By appending this formal model of EFSM to standard WSDL, we can generate a set of test cases which has a better test coverage than other methods. Moreover, a procedure for deriving an EFSM model from WSDL specification is provided to help a service provider augment the EFSM model describing dynamic behaviors of the Web service. To show the efficacy of our approach, we applied our approach to Parlay-X Web services. In this way, we can test Web services with greater confidence in potential fault detection.

No abstract

No abstract

Next-generation neuroprosthetic limbs will require a reliable long-term neural interface to residual nerves in the peripheral nervous system (PNS). To this end, we have developed novel biocompatible materials and a fabrication technique to create high site-count microelectrodes for stimulating and recording from regenerated peripheral nerves. Our electrodes are based on a biodegradable tyrosine-derived polycarbonate polymer system with suitable degradation and erosion properties and a fabrication technique for deployment of the polymer in a porous, degradable, regenerative, multiluminal, multielectrode conduit. The in vitro properties of the polymer and the electrode were tuned to retain mechanical strength for over 24 days and to completely degrade and erode within 220 days. The fabrication technique resulted in a multiluminal conduit with at least 10 functioning electrodes maintaining recording site impedance in the single-digit kOhm range. Additionally, in vivo results showed that neural signals could be recorded from these devices starting at four weeks postimplantation and that signal strength increased over time. We conclude that our biodegradable regenerative-type neural interface is a good candidate for chronic high fidelity recording electrodes for integration with regenerated peripheral nerves.

No abstract

No abstract

We propose a novel affine projection algorithm (APA) that updates the weights intermittently. While the conventional APA updates the weights at each time instant, the proposed APA performs an intermittent update of the weights through dynamically adjusting the update interval. The adjustment of the update interval is accomplished by comparing the squared output error with a threshold derived from the steady-state mean-squared error. Experimental results show that the proposed algorithm has improved performance in terms of its convergence rate and steady-state error while reducing the number of updates.Index Terms-Adaptive filters, affine projection algorithm (APA), update interval.