No abstract

Distal astrocytic processes have a complex morphology, reminiscent of branchlets and leaflets. Astrocytic branchlets are rod-like processes containing mitochondria and endoplasmic reticulum, capable of generating inositol-3-phosphate (IP3)-dependent Ca2+ signals. Leaflets are small and flat processes that protrude from branchlets and fill the space between synapses. Here we use three-dimensional (3D) reconstructions from serial section electron microscopy (EM) of rat CA1 hippocampal neuropil to determine the astrocytic coverage of dendritic spines, shafts and axonal boutons. The distance to the maximum of the astrocyte volume fraction (VF) correlated with the size of the spine when calculated from the center of mass of the postsynaptic density (PSD) or from the edge of the PSD, but not from the spine surface. This suggests that the astrocytic coverage of small and larger spines is similar in hippocampal neuropil. Diffusion simulations showed that such synaptic microenvironment favors glutamate spillover and extrasynaptic receptor activation at smaller spines. We used complexity and entropy measures to characterize astrocytic branchlets and leaflets. The 2D projections of astrocytic branchlets had smaller spatial complexity and entropy than leaflets, consistent with the higher structural complexity and less organized distribution of leaflets. The VF of astrocytic leaflets was highest around dendritic spines, lower around axonal boutons and lowest around dendritic shafts. In contrast, the VF of astrocytic branchlets was similarly low around these three neuronal compartments. Taken together, these results suggest that astrocytic leaflets preferentially contact synapses as opposed to the dendritic shaft, an arrangement that might favor neurotransmitter spillover and extrasynaptic receptor activation along dendritic shafts.

2[0000−0002−2880−2887] , Evgeny Vasiliev 1[0000−0002−7949−1919] , and Vadim Turlapov 1[0000−0001−8484−0565]Abstract. MRI analysis takes central position in brain tumor diagnosis and treatment, thus it's precise evaluation is crucially important. However, it's 3D nature imposes several challenges, so the analysis is often performed on 2D projections that reduces the complexity, but increases bias. On the other hand, time consuming 3D evaluation, like, segmentation, is able to provide precise estimation of a number of valuable spatial characteristics, giving us understanding about the course of the disease. Recent studies, focusing on the segmentation task, report superior performance of Deep Learning methods compared to classical computer vision algorithms. But still, it remains a challenging problem. In this paper we present deep cascaded approach for automatic brain tumor segmentation. Similar to recent methods for object detection, our implementation is based on neural networks; we propose modifications to the 3D UNet architecture and augmentation strategy to efficiently handle multimodal MRI input, besides this we introduce approach to enhance segmentation quality with context obtained from models of the same topology operating on downscaled data. We evaluate presented approach on BraTS 2018 dataset and discuss results.

No abstract

No abstract