In the radiotherapy of nasopharyngeal carcinoma (NPC), magnetic resonance imaging (MRI) is widely used to delineate tumor area more accurately. While MRI offers the higher soft tissue contrast, patient positioning and couch correction based on bony image fusion of computed tomography (CT) is also necessary. There is thus an urgent need to obtain a high image contrast between bone and soft tissue to facilitate target delineation and patient positioning for NPC radiotherapy. In this paper, our aim is to develop a novel image conversion between the CT and MRI modalities to obtain clear bone and soft tissue images simultaneously, here called bone-enhanced MRI (BeMRI).Thirty-five patients were retrospectively selected for this study. All patients underwent clinical CT simulation and 1.5T MRI within the same week in Shenzhen Second People's Hospital. To synthesize BeMRI, two deep learning networks, U-Net and CycleGAN, were constructed to transform MRI to synthetic CT (sCT) images. Each network used 28 patients' images as the training set, while the remaining 7 patients were used as the test set (~1/5 of all datasets). The bone structure from the sCT was then extracted by the threshold-based method and embedded in the corresponding part of the MRI image to generate the BeMRI image. To evaluate the performance of these networks, the following metrics were applied: mean absolute error (MAE), structural similarity index (SSIM), and peak signal-to-noise ratio (PSNR).In our experiments, both deep learning models achieved good performance and were able to effectively extract bone structure from MRI. Specifically, the supervised U-Net model achieved the best results with the lowest overall average MAE of 125.55 (P<0.05) and produced the highest SSIM of 0.89 and PSNR of 23.84. These results indicate that BeMRI can display bone structure in higher contrast than conventional MRI.A new image modality BeMRI, which is a composite image of CT and MRI, was proposed. With high image contrast of both bone structure and soft tissues, BeMRI will facilitate tumor localization and patient positioning and eliminate the need to frequently check between separate MRI and CT images during NPC radiotherapy.
This paper presents a novel Radial Basis Function (RBF) neural network model based on Artificial Immune principle, termed AI-based RBF, to estimate the regional head tissue conductivity. In this model, immune learning algorithm is used for determining the number and location of the centers of the hidden layer by regarding the input data of network as antigens, and the centers of the hidden layer as antibodies. The least square algorithm is adopted for achieving the weights of the output layer. With a 2-D concentric circular model of 3 layers, the higher precision and less computation time by this strategy are obtained than those by RBF model
Support vector machines (SVMs) are learning algorithms derived from statistical learning theory, and originally designed to solve binary classification problems. How to effectively extend SVMs for multiclass classification problems is still an ongoing research issue. In this paper, a sphere-shaped SVM for multiclass problems is presented. Compared with the classical plane-shaped SVMs, the number of convex quadratic programming problems and the number of variables in each programming are smaller. Such SVM classifier is applied to the electroencephalogram (EEG) source localization problem, and the multiplicity of source models is determined according to the potentials recorded on the scalp. Experimental results indicate that the sphere-shaped SVM based classifier is an effective and promising approach for this task.
Here, the tests about the study of Chinese semantic are presented. The experiments are performed on 9 healthy university students by using the scalp EEG and by using the pictures of Chinese words and idioms as stimuli. The mapping of the power density spectrums are adopted to detect the cognition-related region in the working memory including the memory and the retrieval. The non-linear regression method is available to investigate the interactions in these brain areas and the statistical measurement, ANOVA is utilized to get the significance of interdependencies differences between the working memory and before memory. It is concluded that not only the power density spectrums of gamma oscillations are significantly strengthened, but also the interdependencies in the prefrontal brain region are increased during these memory tasks on all subjects. It is indicated that it is possible to exist the strong feature modulations of attention in the local brain regions during the short-time memory of Chinese semantics.
The work here presents an abnormal EEG simulation and an analysis for the abnormal spikes in the simulation by using the wavelet method. The simulation is derived from the electrophysiological model of an excitable neuron being in a disorder process. The spike wave and the multi-spike wave of the EEG morphology are reconstructed by step changes in the concentration of the intracellular calcium ions ([Ca] i ). In the further work, when the concentration of [Ca] i is sufficiently large, the multi-spike wave can also be reconstructed and the spikes of the potentials are analyzed by the multi-layer wavelet method. The work will be helpful to understand how the EEG morphology is formed from the microcosmic viewpoint
Many electrophysiological experiments have shown that epileptic seizures often originate from the synchronous activities of abnormally excitable neurons. The dynamic process of epilepsy is very complex, and characterized by a seemingly rapid and dramatic birth of new oscillations, essentially leading to a propagation and amplification of the original aberrant activity. It is very difficult to thoroughly understand the mechanism from a theoretical standpoint, however some special work can prove helpful. Here we present a theoretical framework to investigate chaos and complexity in the synchrony of excitable neurons in an effort to study the collective oscillations within a neural network. As endogenous rhythms, oscillations arise because most cellular processes contain feedback. The Chay model of excitable neurons is chosen because the model describes the abnormal process, where spiking can be transformed into bursting via bifurcation. In our study, the Chay model is regarded as an abnormal oscillator and coupled via a resistor representing the effect of gap junctions (electrical synapses). In this paper, we present some models developed from the original Chay model, for the synchrony of two cells and a 2D neural network. Lyapunov exponent and phase portrait are utilized to evaluate the chaotic dynamics. Finally, approximate entropy is utilized to measure its complexity. Our results show that the synchrony of abnormal oscillations can occur when the coupling strength of the gap junction is sufficiently large. It is also found that the concentration of Ca 2+ ions does not synchronize. In the 2D network, approximate entropies of different oscillations with strong coupling strength are greater than those with weak coupling strength. It is indicated that synchronous neurons have greater ability to produce new oscillations than asynchronous ones. This work shows that nonlinear analytical methods may prove useful in elucidating the mechanisms of pathologic conditions, where new oscillations are born and propagated, such as in epilepsy.
Mental fatigue induced by long time mental work can cause deterioration in task performance and increase the risk of accidents. Recently, electroencephalogram (EEG)-based monitoring of mental fatigue has received increasing attention in the field of brain-computer interfaces (BCI). This study aims to employ EEG signals to measure the mental fatigue level by estimating reaction time (RT) in a psychomotor vigilance task (PVT). In a 36-hour sleep deprivation experiment, EEG data from 18 subjects were recorded every four hours in nine blocks, each consisting of three tasks: a 6-minute PVT task and two 3-minute resting states (eyes closed and eyes open). The mean RT in the PVT task showed a generally increasing trend during the 36-hour awake period, reflecting the increase of fatigue over time. For each task, multiple EEG features were extracted and selected to better estimate RT using a multiple linear regression (MLR) method. The correlation between predicted RT and actual RT was evaluated using a leave-one-subject-out (LOSO) validation strategy. After parameter optimization, EEG data from the PVT task obtained a mean correlation coefficient of $0.81 \pm 0.16$ across all subjects. Resting-state EEG data showed lower correlations (eyes-closed: $0.65 \pm 0.20$, eyes-open: $0.50 \pm 0.30)$ partially due to the involvement of shorter data lengths. These results demonstrate the feasibility and robustness of the EEG-based fatigue monitoring method, which could be potential for applications in operational environments.
Estimating head tissue conductivity for each layer is a high dimensional, non-linear and ill-posed problem which is part of Electrical Impedance Tomography (EIT) inverse problem. Traditional methods have many difficulties in resolving this problem. Support Vector Machine (SVM) based on Statistical Learning Theory (SLT) is a new kind of learning method including Support Vector Classification (SVC) and Support Vector Regression (SVR). A new method using SVR is proposed to solve the problem in multi-input and multi-output system named Multi-SVM (MSVM). Tissue conductivity for each layer in 2-D head model is estimated effectively by MSVM. Compared with wavelet neural network method, MSVM not only obtains higher accuracy of learning, it also has greater generalization ability and faster computing speed as our experiment demonstrates
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)