Gao Shangkai

Brain-computer interface based on the high-frequency steady-state visual evoked potential

Low-frequency steady-state visual evoked potentials (SSVEPs) are used as the input signal in the present SSVEP-based brain-computer interface (BCI). This prototype system has a high information transfer rate. On the other hand, it has some limitations including visual fatigue, false positive, and some possibility of causing a seizure. These drawbacks can be largely eliminated when using high-frequency stimulations. In this paper, we study the amplitude versus stimulation frequency response of SSVEPs. The signal-to-noise ratio versus frequency curve suggests that the high-frequency SSVEP (>20Hz) could help to construct a practical BCI system.

An EEG-based cursor control system

Describes an experimental brain-computer interface (BCI) which allows severely disabled persons to move the cursor on the screen. The system is based on the detecting of steady-state visual evoked potential (SSVEP). Four rectangular blocks around the cursor indicate four directions and flash at different frequencies on the screen. Once the subject fixates on one of the blocks, the specific SSVEP is obtained SSVEP signal is processed online to determine which block the eyes are focusing. As a feedback, the cursor can be moved in corresponding directions. Preliminary results are presented and evaluated.

BARKER CODE IN TCD ULTRASOUND SYSTEMS TO IMPROVE THE SENSITIVITY OF EMBOLI DETECTION

In present research, Barker-coded excitation is applied to improve the sensitivity of emboli detection in transcranial Doppler (TCD) ultrasound systems. A 13-chip Barker code with a base pulse sequence that comprises two cycles at 2 MHz is used to compare with a 20-cycle pulse in both simulation studies and in vitro experiments. The results show that the system using coded excitation gains a 2.78-dB improvement in measured embolus-to-blood ratio (MEBR) while keeping a similar signal-to-noise ratio (SNR). The simulation results also indicate that the MEBR of the emboli are related to their velocities-a fast embolus has lower MEBR compared with a slow one with the same backscatter amplitudes, which makes fast emboli more difficult to detect. TCD system using coded excitation is more sensitive for faster and smaller emboli, which is significant for early diagnosis of stroke.

Stimulation frequency extraction in SSVEP-based brain-computer interface

A periodogram-based method is used to extract the stimulation frequency in a brain-computer interface (BCI) based on steady-state visual evoked potential (SSVEP). In the system, tens of buttons illuminated at different frequencies are used to generate deterministic sinusoidal responses or SSVEP at the visual cortex, which are derived from the electroencephalogram (EEG) by a suitable electrode array. Based on the periodogram of a time-series, we test the EEG data for the presence of hidden periodic components, which correspond to SSVEP, and extract the stimulation frequencies. The method performs well in simulation, and is applied successfully to real data.

Construction of a Knowledge Center for Medical Image Processing

This article introduces a medical imaging and image processing knowledge center which serves as an e-learning platform based on network technology. This platform includes contents of medical imaging and medical image processing for the teaching activities, virtual reality tutoring system of human skeleton, medical image database, etc. The purpose of establishing this platform is to recruit and excavate information records which usually associated with the daily data exchanged in teaching and research activities. Then it will gradually form an open professional knowledge database. At present, this platform plays the role of knowledge community in the educational activities of BME program in Tsinghua University

3D navigation of CTVE and correction of MinIP methods in non-invasive diagnostic detection

Navigation is important in Computed Tomography Virtual Endoscopy (CTVE) functions. Existing navigation methods involve planning and re-calculating of a fixed path before the viewpoint flight inside the targeted organs. This includes path construction, centering, smoothing and ulti-branch processing. This paper proposes a 3D navigation method which was achieved by utilizing the compatibility of the 3D navigation and he ray casting 3D rendering method, without the necessity of planning and pre-calculating a fixed path, eliminating the multi-branch problems. In ur 3D navigation method, the viewpoint direction and location are tracked in real time when the viewpoint is inside the organs. At the same time t presents and controls the direction changes and location changes of the viewpoint in x, y, z dimensions. With interactive control of the viewpoint, t can fly in any direction in 3D, not only along a fixed path, thus eliminating multi-branch problems. The viewpoint locations and directions will hange smoothly and will be used to calculate the current scene of CTVE. Accelerated ray casting is used to render 3D scenes, which is compatible ith the 3D navigation method. In Minimum Intensity Projection (MinIP) applications, if the conventional MinIP reconstruction method is used, realistic results cannot be chieved when they are rotated in multi-directions because the Computed Tomography (CT) images always contain empty regions surrounding the issue regions. The conventional MinIP reconstruction algorithm always chalks up a minimum intensity voxel which relates to the empty regions nstead of the tissue regions. To solve such valid voxel searching problems, seed-filling algorithms are used to fill the empty regions of each slice utomatically. The empty voxels are labeled automatically and are avoided in MinIP calculation to gain correct results in all directions. 2006 Published by Elsevier Ltd.

Detection of the short-term preseizure changes in EEG recordings using complexity and synchrony analysis

Abstract An important consideration in epileptic seizure prediction is proving the existence of a pre-seizure state that can be detected using various signal processing algorithms. In the analysis of intracranial electroencephalographic (EEG) recordings of four epilepsy patients, the short-term changes in the measures of complexity and synchrony were detected before the majority of seizure events across the sample patient population. A decrease in complexity and increase in phase synchrony appeared several minutes before seizure onset and the changes were more pronounced in the focal region than in the remote region. This result was also validated statistically using a surrogate data method. * Supported by the Specialized Research Fund for the Doctoral Program of Higher Education (Grant No. 20020003031), NSFC-RGC Joint Research Fund (Grant No. 60318001) and the Key Project of Ministry of Education of the Peoples Republic of China (Grant No. 1041185)

Sleep staging based on the analysis of heart rate variability with hidden Markov models

Summary form only given. The relationship between sleep stages and heart rate variability (HRV) is investigated. A novel method to deduce the sleep stages by the analysis of HRV with hidden Markov models (HMM) is presented. Our method provides a simple means to evaluate the sleep situation, which is useful in many applications.

The forward transfer function of the potential from the cortex to the scalp

Summary form only given. The forward transfer function of potential from the cortical surface to the scalp surface is investigated in this paper. The integral equation of the first kind, which relates cortex to scalp potential, has been derived, and the analytic solution of the point spread function of potential in a three-shelled homogeneous, spherical volume conductor head model has been derived. Simulation results show that the forward transfer coefficient is affected by the ratio of skull to scalp conductivities and the 3 db point of the potential point spread function averages about 1.7 cm under newly suggested value of tissue conductivities founded by recent studies.

The correction of amplitude and phase deviations in ultrasound Doppler blood flowmeter

The amplitude and phase deviations in ultrasound Doppler blood flowmeter would lead to wrong decision in clinical diagnosis. In order to reduce the influence of the deviations, a software method for deviation correction was presented. First, the values of amplitude and phase deviations were obtained by a signal correlation technique, then these values were used to correct the blood flow signal. Simulation results prove that the software method was a promising method.