ABSTRACT
Motor imagery base brain-computer interface (BCI) is an appropriate solution for stroke patient to rehabilitate and communicate with external world. For such applications speculating whether the subjects are doing motor imagery is our primary mission. So the problem turns into how to precisely classify the two tasks, motor imagery and idle state, by using the subjects' electroencephalographic (EEG) signals. Feature extraction is a factor that significantly affects the classification result. Based on the concept of Continuous Wavelet Transform, we proposed a wavelet-liked feature extraction method for motor imagery discrimination. And to compensate the problem that the feature varies between subjects, we use the subjects' own EEG signals as the mother wavelet. After determining the feature vector, we choose Bayes linear discriminant analysis (LDA) as our classifier. The BCI competition III dataset IVa is used to evaluate the classification performance. Comparing with variance and fast Fourier transform (FFT) methods in feature extraction, 2.02% and 16.96% improvement in classification accuracy are obtained in this work respectively.
Subject(s)
Algorithms , Brain-Computer Interfaces , Imagery, Psychotherapy , Motor Activity , Wavelet Analysis , Bayes Theorem , Discriminant Analysis , Electrodes , Electroencephalography , Humans , Signal Processing, Computer-AssistedABSTRACT
fMRI was used to investigate brain organization for reading in Chinese. Subjects were shown two-character Chinese words. A control task was used to eliminate the non-linguistic visual and motor confounds. Results show that naming of Chinese logographs is characterized by left-lateralized neuronal networks for the processing of orthographic, phonological, and semantic attributes. The orchestration of the middle frontal cortex, superior temporal cortex, superior parietal cortex, basal temporal area and extrastriate cortices of the left hemisphere may manifest the particularity of the central representation of simple word naming in Chinese.
Subject(s)
Cerebral Cortex/physiology , Functional Laterality/physiology , Magnetic Resonance Imaging , Reading , Adult , Female , Frontal Lobe/physiology , Humans , Male , Parietal Lobe/physiology , Space Perception/physiology , Temporal Lobe/physiologyABSTRACT
This paper presents a system for reconstructing a four-dimensional (4D) heart-beating image from transesophageal echocardiographic (TEE) data acquired with a rotational approach. The system consists of the necessary processing modules for two-dimensional (2D) echocardiogram reformation and 3D/4D-image reconstruction. These include the modules of image decoding, image re-coordinating, and three-dimensional (3D) volume rendering. The system is implemented under PC platform with Windows 95 operating system (with Intel Pentium-166 CPU, 64 MB RAM on board, and 2.0 GB hard disk capacity). It takes 6 min to reconstruct a 4D echocardiographic data set. The resultant 2D/3D/4D echocardiographic image provide the tools for investigating the phenomenon of heart beating, exploring the heart structure, and reformatting the 2D echocardiograms in an arbitrary plane. The functions provided by the system can be applied for further studies, such as 3D cardiac shape analysis, cardiac function measurement, and so forth.
Subject(s)
Computer Systems , Echocardiography, Four-Dimensional/methods , Echocardiography, Transesophageal/methods , Image Processing, Computer-Assisted/methods , Data Display , Humans , Programming Languages , SoftwareABSTRACT
A method for three-dimensional shape analysis of left ventricle (LV) is presented in this article. The method uses three-dimensional transesophageal echocardiography (TEE) as the source to derive the 3D wire-frame model and the related shape descriptors. The shape descriptors developed in this article include regional surface changing (RSC), global surface curvature (GSC), surface distance (SD), normalized surface distance (ND), and effective radius (ER) of the endocardial surface. Based on these shape descriptors, the shape of LV could be sketched in both static and dynamic manner. The results show that the new approach provides a robust but easy method to quantify regional and global LV shape from 2D and 3D echocardiograms.
Subject(s)
Echocardiography, Three-Dimensional , Echocardiography, Transesophageal , Heart Ventricles/diagnostic imaging , Image Processing, Computer-Assisted/methods , Algorithms , Analysis of Variance , Endocardium/diagnostic imaging , Feasibility Studies , Humans , Image Enhancement/methods , Observer Variation , Phantoms, ImagingABSTRACT
A unified approach to managing multimedia medical databases in a telemedicine system is proposed. In order to manage, search, and display patient information more efficiently, we define a patient information package (PIP) as a concise data set of a patient's medical information from each visit. By means of PIP's, both patient-oriented and problem-oriented query strategies, which are most frequently used in daily clinical practice and medical education, can be accommodated. We also provide a unified methodology for accessing various types of patient medical records as well as design two types of user interfaces, high-quality data display and web-based interface, for different medical service purposes. The PIP-based management of databases has been successfully implemented between the National Taiwan University (NTUH), Taipei, and the Chinshan health care center, Chinshan, Taiwan, for teleconsultation, telediagnosis, and tele-education.
