Poststroke rehabilitation using repetitive transcranial magnetic stimulation during pregnancy: A case report.

BACKGROUND: Repetitive transcranial magnetic stimulation (rTMS) is a form of magnetic stimulation therapy used to treat depression, migraine, and motor function impairment in patients with stroke. As there is little research on the effects of rTMS in pregnant women, it is not widely used in these patients. This case report aimed to demonstrate the safety of rTMS in pregnant patients. CASE SUMMARY: After much consideration, we applied rTMS to treat recent stroke and hemiplegia in a 34-year-old pregnant woman. The patient received 45 sessions of low-frequency treatment over the course of 10 wk. We closely monitored the mother and fetus for potential side effects; the results showed significant improvement in the patient's motor function, with no harmful effects on the mother or fetus during pregnancy or after delivery. The patient's fine motor and walking functions improved after treatment. This case is the first instance of a stroke patient treated with rTMS during pregnancy. CONCLUSION: This case demonstrates that rTMS could be used to improve motor function recovery in stroke patients during pregnancy.

Low Complexity Gradient Computation Techniques to Accelerate Deep Neural Network Training.

Deep neural network (DNN) training is an iterative process of updating network weights, called gradient computation, where (mini-batch) stochastic gradient descent (SGD) algorithm is generally used. Since SGD inherently allows gradient computations with noise, the proper approximation of computing weight gradients within SGD noise can be a promising technique to save energy/time consumptions during DNN training. This article proposes two novel techniques to reduce the computational complexity of the gradient computations for the acceleration of SGD-based DNN training. First, considering that the output predictions of a network (confidence) change with training inputs, the relation between the confidence and the magnitude of the weight gradient can be exploited to skip the gradient computations without seriously sacrificing the accuracy, especially for high confidence inputs. Second, the angle diversity-based approximations of intermediate activations for weight gradient calculation are also presented. Based on the fact that the angle diversity of gradients is small (highly uncorrelated) in the early training epoch, the bit precision of activations can be reduced to 2-/4-/8-bit depending on the resulting angle error between the original gradient and quantized gradient. The simulations show that the proposed approach can skip up to 75.83% of gradient computations with negligible accuracy degradation for CIFAR-10 dataset using ResNet-20. Hardware implementation results using 65-nm CMOS technology also show that the proposed training accelerator achieves up to 1.69× energy efficiency compared with other training accelerators.

Main results of the Korea National Hospital Discharge In-depth Injury Survey, 2004-2016.

OBJECTIVES: The purpose of this study was to estimate the incidence of injuries and to identify their causes by classifying injuries according to various categories including age, sex, mechanism of injury, body parts injured, and place of injury. METHODS: This study used data from the Korea National Hospital Discharge In-depth Injury Survey (KNHDIS) from 2004 to 2016. The KNHDIS is conducted annually by the Korea Centers for Disease Control and Prevention, and its survey population includes all hospitalized patients discharged from medical institutions that have 100 or more beds, such as hospitals, general hospitals, and secondary community health centers. The number of injured cases is weighted and estimated using the mid-year estimated population of each year. RESULTS: The injury discharge rate steadily increased since 2004 (1,505 per 100,000 population in 2004, 2,007 per 100,000 population in 2016) and most injuries were unintentional (annual average of 94.7%). On average, during the 13-year study period, the injury rate for males was 1.5 times as high as for females. The 2 main causes of injury were consistently traffic accidents and falls. Notably, the rate of injuries resulting from falls rose by 1.7-fold from 463 to 792 per 100,000 people, and exceeded the rate of traffic accidents in 2016. CONCLUSIONS: The incidence of injuries steadily increased after the survey was first conducted, whereas mortality resulting from injuries mostly remained unchanged. This suggests that effective strategies and interventions should be reinforced to reduce unintentional injuries.