Visualisation of cardiac fibrillation plays a very considerable role in cardiophysiological study and clinical applications. One of the ways to obtain the image of these phenomena is the registration of mechanical displacement fields reflecting the track from electrical activity. In this work, we read these fields using cross-correlation analysis from the video of open pig's epicardium at the start of fibrillation recorded with electrocardiogram. However, the quality of obtained displacement fields remains low due to the weak pixels heterogeneity of the frames. It disables to see more clearly such interesting phenomena as mechanical vortexes that underline the mechanical dysfunction of fibrillation. The applying of chemical or mechanical markers to solve this problem can affect the course of natural processes and falsify the results. Therefore, we developed a novel scheme of an unsupervised deep neural network that is based on the state-of-art positional coding technology for a multilayer perceptron. This network enables to generate a couple of frames with a more heterogeneous pixel texture, that is more suitable for cross-correlation analysis methods, from two consecutive frames. The novel network scheme was tested on synthetic pairs of images with different texture heterogeneity and frequency of displacement fields and also it was compared with different filters on our cardiac tissue image dataset. The testing showed that the displacement fields obtained with our method are closer to the ground truth than those which were computed only with the cross-correlation analysis in low contrast images case where filtering is impossible. Moreover, our model showed the best results comparing with the one of the popular filter CLAHE on our dataset. As a result, using our approach, it was possible to register more clearly a mechanical vortex on the epicardium at the start of fibrillation continuously for several milliseconds for the first time.
Rotating nonlinear waves of excitation in the heart cause dangerous cardiac arrhythmias. Frequently, ventricular arrhythmias occur as a result of myocardial infarction and are associated with rotation of the waves around a post-infarction scar. In this paper, we perform a detailed in silico analysis of scroll waves in an anatomical model of the human ventricles with a generic model of the infarction scar surrounded by the gray zone with modified properties of the myocardial tissue. Our model includes a realistic description of the heart shape, anisotropy of cardiac tissue and a detailed description of the electrical activity in human ventricular cells by a TP06 ionic model. We vary the size of the scar and gray zone and analyze the dependence of the rotation period on the injury dimensions. Two main regimes of wave scrolling are observed: the scar rotation, when the wave rotates around the scar, and the gray zone rotation, when the wave rotates around the boundary of the gray zone and normal tissue. The transition from the gray zone to the scar rotation occurs for the width of gray zone above 10–20 mm, depending on the perimeter of the scar. We compare our results with simulations in 2D and show that 3D anisotropy reduces the period of rotation. We finally use a model with a realistic shape of the scar and show that our approach predicts correctly the period of the arrhythmia.
Waves of electrical excitation rotating around an obstacle is one of the important mechanisms of dangerous cardiac arrhythmias occurring in the heart damaged by a post-infarction scar. Such a scar is also surrounded by the region of heterogeneity called a gray zone. In this paper, we perform the first comprehensive numerical study of various regimes of wave rotation around an obstacle surrounded by a gray zone. We use the TP06 cellular ionic model for human cardiomyocytes and study how the period and the pattern of wave rotation depend on the radius of a circular obstacle and the width of a circular gray zone. Our main conclusions are the following. The wave rotation regimes can be subdivided into three main classes: (1) functional rotation, (2) scar rotation and the newly found (3) gray zone rotation regimes. In the scar rotation regime, the wave rotates around the obstacle, while in the gray zone regime, the wave rotates around the gray zone. As a result, the period of rotation is determined by the perimeter of the scar, or gray zone perimeter correspondingly. The transition from the scar to the gray rotation regimes can be determined from the minimal period principle, formulated in this paper. We have also observed additional regimes associated with two types of dynamical instabilities which may affect or not affect the period of rotation. The results of this study can help to identify the factors determining the period of arrhythmias in post-infarction patients.
Waves of electrical excitation rotating around an obstacle is one of the important mechanisms of dangerous cardiac arrhythmias occurring in the heart damaged by post-infarction scar. Such a scar also has a border zone around it, which has electrophysiological properties different from the rest of normal myocardial tissue. Spatial patterns of wave rotation in the presence of such tissue heterogeneity are poorly studied. In this paper we perform a comprehensive numerical study of various regimes of rotation of a wave in a plane layer of the ventricular tissue around an obstacle surrounded by a gray zone. We use a TP06 cellular ionic model which reproduces the electrophysiological properties of cardiomyocytes in the left ventricle of human heart. We vary the extent of obstacle and gray zone and study the pattern of wave rotation and its period. We observed different regimes of wave rotation that can be subdivided into several classes: (1) functional rotation and (2) scar rotation regimes, which were identified in the previous studies, and new (3) gray zone rotation regime: where the wave instead of rotation around the obstacle, rotates around the gray zone (an area of tissue heterogeneity) itself. For each class, the period of rotation is determined by different factors, which we discuss and quantify. We also found that due to regional pathological remodeling of myocardial tissue, we can obtain additional regimes associated with dynamical instabilities of two types which may affect or not affect the period of rotation.
E intense movement of the heart and the presence of blood on surface, the application of the necessary small markers is rather difficult, and the use of luminous chemicals would harm physiological functioning. Moreover, these videos contain motion artifacts that complicate further analysis with Particle Image Velocimetry. In this paper, an image preprocessing algorithm was proposed. It is based on approximate tracking individual fragments using the Mean Squared Error for the matrix. The result is binary images where small points are built instead of each fragment. In this study, the proposed algorithm showed better results in comparison with the most suitable filtering methods for specific frames, namely, the Sobel filter and the Canny edge detector. This can be partially explained by the higher density of vector fields due to the absence of unreliable vectors. Thus, the proposed method, unlike others, allows to get vector fields with visible vortex-like mechanical movements.
Popularity of using deep network architectures consisting of multilayer perceptrons (MLPs) has increased recently. The main reason for the phenomenon is the emergence of state-of-the-art technologies that overcome the problem of MLP underfitting in generating graphical information: Gaussian Fourier Function Transformation (GFFT) of MLP input and SineLayer included. Specifically, fully unsupervised MLP-based schemes combining the work of two MLP, grid generator and image generator, have been developed for measuring displacement fields and optical flow in images. However, because of the novelty of these technologies, there is little research dedicated to studying the optimal architecture for such schemes, as well as comparing the methods of SineLayer and GFFT specifically for the task of measuring displacement fields in 2D images. Therefore, the main advantage of this research is fitting the most optimal architecture of MLP in the scheme grid generator - image generator. The results showed that the best variant is an architecture with grid generator and image generator with 4 hidden layers including SineLayers. The testing on synthetic images revealed that the final model has advantages in computing displacement fields in images with gradient-changing pixel values compared to the state-of-the-art method (Unsupervised DICNet and ALDIC). Furthermore, it showed promising results in realistic images of the heart’s surface with additional elements, providing displacement fields without any sharp jumps in displacement MSE. Thus, it can be concluded that the use of MLP-based models for measuring displacement fields in 2D images has potential for the further research and development of novel methods for similar tasks.