RESUMEN
Colorectal cancer (CRC) is one of the most common types of cancer among men and women. The grading of dysplasia and the detection of adenocarcinoma are important clinical tasks in the diagnosis of CRC and shape the patients' follow-up plans. This study evaluated the feasibility of deep learning models for the classification of colorectal lesions into four classes: benign, low-grade dysplasia, high-grade dysplasia, and adenocarcinoma. To this end, a deep neural network was developed on a training set of 655 whole slide images of digitized colorectal resection slides from a tertiary medical institution; and the network was evaluated on an internal test set of 234 slides, as well as on an external test set of 606 adenocarcinoma slides from The Cancer Genome Atlas database. The model achieved an overall accuracy, sensitivity, and specificity of 95.5%, 91.0%, and 97.1%, respectively, on the internal test set, and an accuracy and sensitivity of 98.5% for adenocarcinoma detection task on the external test set. Results suggest that such deep learning models can potentially assist pathologists in grading colorectal dysplasia, detecting adenocarcinoma, prescreening, and prioritizing the reviewing of suspicious cases to improve the turnaround time for patients with a high risk of CRC. Furthermore, the high sensitivity on the external test set suggests the model's generalizability in detecting colorectal adenocarcinoma on whole slide images across different institutions.
Asunto(s)
Adenocarcinoma , Neoplasias Colorrectales , Aprendizaje Profundo , Masculino , Humanos , Femenino , Redes Neurales de la Computación , Adenocarcinoma/diagnóstico , Adenocarcinoma/patología , Patólogos , Hiperplasia , Neoplasias Colorrectales/diagnósticoRESUMEN
Bipedal robots have gained increasing attention for their human-like mobility which allows them to work in various human-scale environments. However, their inherent instability makes it difficult to control their balance while they are physically interacting with the environment. This study proposes a novel balance controller for bipedal robots based on a behavior cloning model as one of the machine learning techniques. The behavior cloning model employs two deep neural networks (DNNs) trained on human-operated balancing data, so that the trained model can predict the desired wrench required to maintain the balance of the bipedal robot. Based on the prediction of the desired wrench, the joint torques for both legs are calculated using robot dynamics. The performance of the developed balance controller was validated with a bipedal lower-body robotic system through simulation and experimental tests by providing random perturbations in the frontal plane. The developed balance controller demonstrated superior performance with respect to resistance to balance loss compared to the conventional balance control method, while generating a smoother balancing movement for the robot.
RESUMEN
Though neurotransmitters are essential elements in neuronal signal transduction, techniques for in vivo analysis are still limited. Here, we describe an organic electrochemical transistor array (OECT-array) technique for monitoring catecholamine neurotransmitters (CA-NTs) in rat brains. The OECT-array is an active sensor with intrinsic amplification capability, allowing real-time and direct readout of transient CA-NT release with a sensitivity of nanomolar range and a temporal resolution of several milliseconds. The device has a working voltage lower than half of that typically used in a prevalent cyclic voltammetry measurement, and operates continuously in vivo for hours without significant signal drift, which is inaccessible for existing methods. With the OECT-array, we demonstrate simultaneous mapping of evoked dopamine release at multiple striatal brain regions in different physiological scenarios, and reveal a complex cross-talk between the mesolimbic and the nigrostriatal pathways, which is heterogeneously affected by the reciprocal innervation between ventral tegmental area and substantia nigra pars compacta.
Cells in the nervous system pass messages using a combination of electrical and chemical signals. When an electrical impulse reaches the end of one cell, it triggers the release of chemicals called neurotransmitters, which pass the message along. Neurotransmitters can be either activating or inhibitory, determining whether the next cell fires its own electrical signal or remains silent. Currently, researchers lack effective methods for measuring neurotransmitters directly. Instead, methods mainly focus on electrical recordings, which can only tell when cells are active. One new approach is to use miniature devices called organic electrochemical transistors. Transistors are common circuit board components that can switch or amplify electrical signals. Organic electrochemical transistors combine these standard components with a semi-conductive material and a flexible membrane. When they interact with certain biological molecules, they release electrons, inducing a voltage. This allows organic electrochemical transistors to detect and measure neurotransmitter release. So far, the technology has been shown to work in tissue isolated from a brain, but no-one has used it to detect neurotransmitters inside a living brain. Xie, Wang et al. now present a new device that can detect the release of the neurotransmitter, dopamine, in real-time in living rats. The device is a miniature microarray of transistors fixed to a blade-shaped film. Xie, Wang et al. implanted this device into the brain of an anaesthetised rat and then stimulated nearby brain cells using an electrode. The device was able to detect the release of the neurotransmitter dopamine, despite there being a range of chemicals released inside the brain. It was sensitive to tiny amounts of the neurotransmitter and could distinguish bursts that were only milliseconds apart. Finally, Xie, Wang et al. also implanted the array across two connected brain areas to show that it was possible to watch different brain regions at the same time. This is the first time that transistor arrays have measured neurotransmitter release in a living brain. The new device works at low voltage, so can track brain cell activity for hours, opening the way for brand new neuroscience experiments. In the future, adaptations could extend the technology even further. More sensors could give higher resolution results, different materials could detect different neurotransmitters, and larger arrays could map larger brain areas.