Search

Adjuvant nab-paclitaxel plus gemcitabine vs gemcitabine alone for resected pancreatic ductal adenocarcinoma: A single center experience in China

Efficient navigation through space encompasses the ability to maintain a representation of self-location relative to the external environment in real behavioral time. Beyond the here and now, recalling a past experience or imagining a future scenario can be rather rapid, but experienced as if it were occurring in real time. Interestingly, damage to the hippocampus or its interactions with the prefrontal cortex impairs navigational decisions, memory retrieval, and future planning. How could the hippocampal-prefrontal system flexibly support these distinct timescale functions? To investigate the underlying physiological mechanisms, we continuously tracked the same hippocampal and prefrontal ensembles in rats throughout learning of a spatial alternation task. We found flexible expression of neuronal sequences at multiple timescales in both regions, including slow behavioral-timescale sequences (~seconds), as well as two types of fast sequences (~100-200 ms) as theta sequences during online periods and replay sequences during offline periods. The focus of my dissertation was to characterize and assess the function roles of these sequences, and to consider how these multiple-timescale neural codes may function as neuronal substrates to link memory to navigational decisions., This dissertation first examines hippocampal-prefrontal replay in awake versus sleep states, and further discusses our findings in the context of earlier studies towards a framework of how sleep and awake replay may work together to support learning and memory. Second, it investigates the learning dynamics of hippocampal-prefrontal replay, and how such dynamics contribute to different stages of learning to support our ability of retrospective evaluation and prospective planning. Third, it interrogates the functional roles of theta sequences in both the hippocampus and prefrontal cortex, and considers the idea of how replay and theta sequences may contribute to complementary parts of a coordinated mechanism for learning and memory. Finally, it provides a synthesis of these observations, and suggests that these multiple-timescale neural codes in the hippocampal-prefrontal system contribute to flexible use of memories for future planning during navigational decisions.

Supplemental material, sj-docx-1-jet-10.1177_15266028221125581 for Network Meta-Analysis of All Available Regimens Based on Drug-Coated Balloon Angioplasty and Laser Atherectomy for Femoropopliteal In-Stent Restenosis by Chaowen Yu, Wenbo Tang, Tiancai Ren, Qiwei Chen, Ran Lu and Yong Gao in Journal of Endovascular Therapy

This paper presents the evolution of completion sand control designs for the Peng Lai field in Bohai Bay, offshore China. A range of sand control designs from open hole gravel packs to open hole standalone screens to cased hole high rate water packs and cased hole frac-packs have been tried over time. Evaluation of these various designs indicates that single-trip multi-zone cased hole frac-pack designs provide the most effective method for developing the thick, low net-to-gross, unconsolidated sandstone reservoirs in this waterflood field. The paper describes the completion design and evaluation process used in the field and shows how field challenges led to innovative frac-pack downhole equipment designs. This paper highlights completion practices and results from a range of well types used to develop the large Peng Lai field, offshore China. Development challenges and productivity and reliability results for alternative completion designs are discussed on the basis of field performance. The resulting preferred design, consisting of cased hole frac-packs using large-bore and small-bore single-trip multi-zone equipment are then discussed in further detail with emphasis on development of an innovative, fit-for-purpose, small-bore design. The installation challenges, solutions and successes experienced during the design and installation process and the actual performance and benefits of the new system are also discussed in this paper. Over the course of Peng Lai field development, a range of alternative completion sand control designs have been tried and evaluated. Cased hole frac-packs provide the most effective completion method for the thick, low net-to-gross, sandstone reservoirs. To enable effective completions in sidetracked wells, a 7" single-trip multi-zone frac pack tool has been developed. This small-bore system has been used to complete 13 wells with 67 zones in the field to date, saving over 43 rig days and $8MM. This paper presents reliability and productivity comparisons for a range of alternative completion sand control designs in a single, large field environment. A new 7" single-trip, fit-for-purpose design has been developed for the field and successfully implemented in multiple well campaigns. Installation time data and cost savings are provided.

Recently, deep convolutional neural networks (DCNNs) have shown outstanding performance in dense classification problems such as semantic segmentation. While existing semantic segmentation networks work well for the 2D image, the research for 3D neuron image segmentation is insufficient especially in terms of feature fusion. In this paper, we propose an approach, named Multi-Path 3D U-Net, for 3D image segmentation that leverages the power of a features fusion in encoder-decoder framework, trained in an end-to-end manner, to predict pixel-level labels in volumetric images. In Multi-Path 3D U-Net, We change the features that used for skip connection by fusing high-layer and low-layer features. In this way, the features from the high-layer can increase the semantic information and the features from the low-layer can increase the global information. We demonstrate the effectiveness of the proposed model on SYNTANEI datasets.

In the brain, the nerve fibers of adjacent neurons are intertwined. If the three-dimensional neuron image is directly tracked and reconstructed, it is easy to be interfered by nerve fibers and other noises. Therefore, the basic work of reconstructing or tracing 3D neuron structure is to segment the target neuron from neuron images containing nerve fiber entanglement. Today, deep learning has a wide range of applications in the field of medical image segmentation, and the most representative segmentation model is U-Net. However, the traditional U-Net model has some shortcomings, for example, the jump connection directly transmits the features to the corresponding up-sampling layer without processing, which may cause edge misjudgment. In this article, we propose an improved method based on 3D U-Net. We redesign the jump connection, add the residual path, enhance the low-resolution information, extract the context information through the hole convolution and splice it to the corresponding up-sampling layer, and add an upper jump connection to supplement the loss of deep learning Detailed information. Compared the proposed model with other networks, we come to conclusion that our proposed model significantly improved the accuracy of neuron segmentation.