Search for: All records

Background The novel coronavirus SARS-CoV-2 and its associated disease, COVID-19, have caused worldwide disruption, leading countries to take drastic measures to address the progression of the disease. As SARS-CoV-2 continues to spread, hospitals are struggling to allocate resources to patients who are most at risk. In this context, it has become important to develop models that can accurately predict the severity of infection of hospitalized patients to help guide triage, planning, and resource allocation. Objective The aim of this study was to develop accurate models to predict the mortality of hospitalized patients with COVID-19 using basic demographics and easily obtainable laboratory data. Methods We performed a retrospective study of 375 hospitalized patients with COVID-19 in Wuhan, China. The patients were randomly split into derivation and validation cohorts. Regularized logistic regression and support vector machine classifiers were trained on the derivation cohort, and accuracy metrics (F1 scores) were computed on the validation cohort. Two types of models were developed: the first type used laboratory findings from the entire length of the patient’s hospital stay, and the second type used laboratory findings that were obtained no later than 12 hours after admission. The models were further validated on a multicenter external cohort of 542 patients. Results Of the 375 patients with COVID-19, 174 (46.4%) died of the infection. The study cohort was composed of 224/375 men (59.7%) and 151/375 women (40.3%), with a mean age of 58.83 years (SD 16.46). The models developed using data from throughout the patients’ length of stay demonstrated accuracies as high as 97%, whereas the models with admission laboratory variables possessed accuracies of up to 93%. The latter models predicted patient outcomes an average of 11.5 days in advance. Key variables such as lactate dehydrogenase, high-sensitivity C-reactive protein, and percentage of lymphocytes in the blood were indicated by the models. In line with previous studies, age was also found to be an important variable in predicting mortality. In particular, the mean age of patients who survived COVID-19 infection (50.23 years, SD 15.02) was significantly lower than the mean age of patients who died of the infection (68.75 years, SD 11.83; P<.001). Conclusions Machine learning models can be successfully employed to accurately predict outcomes of patients with COVID-19. Our models achieved high accuracies and could predict outcomes more than one week in advance; this promising result suggests that these models can be highly useful for resource allocation in hospitals. 

Nervous systems sense, communicate, compute, and actuate movement, using distributed hardware with tradeoffs in speed and accuracy. The resulting sensorimotor control is nevertheless remarkably fast and accurate due to highly effective layered architectures. However, such architectures have received little attention in neuroscience due to the lack of theory that connects the system and hardware level speed-accuracy tradeoffs. In this paper, we present a theoretical framework that connects the speed-accuracy tradeoffs of sensorimotor control and neurophysiology. We characterize how the component SATs in spiking neuron communication and their sensory and muscle endpoints constrain the system SATs in both stochastic and deterministic models. The results show that appropriate speed -accuracy diversity at the neurons/muscles levels allow nervous systems to improve the speed and accuracy in control performance despite using slow or inaccurate hardware. Then, we characterize the fundamental limits of layered control systems and show that appropriate diversity in planning and reaction layers leads to both fast and accurate system despite being composed of slow or inaccurate layers. We term these phenomena “Diversity Sweet Spots.” The theory presented here is illustrated in a companion paper, which introduces simple demos and a new inexpensive and easy-to-use experimental platform. 

This paper describes several surprisingly rich but simple demos and a new experimental platform for human sensorimotor control research and also controls education. The platform safely simulates a canonical sensorimotor task of riding a mountain bike down a steep, twisting, bumpy trail using a standard display and inexpensive off-the-shelf gaming steering wheel with a force feedback motor. We use the platform to verify our theory, presented in a companion paper. The theory tells how component hardware speed-accuracy tradeoffs (SATs) in control loops impose corresponding SATs at the system level and how effective architectures mitigate the deleterious impact of hardware SATs through layering and “diversity sweet spots” (DSSs). Specifically, we measure the impacts on system performance of delays, quantization, and uncertainties in sensorimotor control loops, both within the subject's nervous system and added externally via software in the platform. This provides a remarkably rich test of the theory, which is consistent with all preliminary data. Moreover, as the theory predicted, subjects effectively multiplex specific higher layer planning/tracking of the trail using vision with lower layer rejection of unseen bump disturbances using reflexes. In contrast, humans multitask badly on tasks that do not naturally distribute across layers (e.g. texting and driving). The platform is cheap to build and easy to program for both research and education purposes, yet verifies our theory, which is aimed at closing a crucial gap between neurophysiology and sensorimotor control. The platform can be downloaded at https://github.com/Doyle-Lab/WheelCon. 

Significance Nervous systems use highly effective layered architectures in the sensorimotor control system to minimize the harmful effects of delay and inaccuracy in biological components. To study what makes effective architectures, we develop a theoretical framework that connects the component speed–accuracy trade-offs (SATs) with system SATs and characterizes the system performance of a layered control system. We show that diversity in layers (e.g., planning and reflex) allows fast and accurate sensorimotor control, even when each layer uses slow or inaccurate components. We term such phenomena “diversity-enabled sweet spots (DESSs).” DESSs explain and link the extreme heterogeneities in axon sizes and numbers and the resulting robust performance in sensorimotor control.