칠복음(七福飮)이 Glucose Oxidase에 의해 손상(損傷)된 대뇌피질(大腦皮質) 신경세포(神經細胞)에 미치는 영향(影響)

최공한,박승택,류도곤,최민호,엄상섭,허진영,강성도,고정수,서의석,성은경,조남수,이춘우,황일택,선성규,류영수,Choi, Kong-Han,Park, Seung-Taeck,Ryu, Do-Gon,Choi, Min-Ho,Um, Sang-Sub,Hea, Jin-Young,Kang, Sung-Do,Go, Jeong-Soo,Sou, Eui-Suk,Sung, Yeun-Ky 대한동의생리학회 1999 동의생리학회지 Vol.14 No.2

As the average life span have been lengthened and the rate of senile population have been raised, chronic degenerative diseases incident to aging has been increased rapidly and become a social problem. With this social background, recently, the facts that oxygen radicals(OR) have toxic effects on Central Nervous System and Peripheral Nervous System and cause neuropathy such as Parkinson's Disease, Alzheimer Disease have been turned out, and accordingly lots of studies on the mechanism of the toxic effects of OR on nerves, the diseases caused by OR and the approaches to curing the diseases have been made. The purpose of this study is to examine the toxic effects caused by Glucose Oxidase(GO) and the effects of herbal extracts such as Chilbokyeum(CBY) on the treatment of the toxic effects. For this purpose, experiments with the cultured cell from the cerebrums of new born mice were done. The results of these experiments were as follows. 1. GO, a oxygen radical, decreased the survival rate of the cultured cells on NR assay and MTT assay 2. GO, a oxygen radical, increased lipid peroxidation and the amount of LDH. 3. CBY have efficacy of decreasing lipid peroxidation. 4. CBY have efficacy of decreasing the amount of LOH. From the above results, It is concluded that Chilbokyeum has marked efficacy as a treatment for the damages caused in the GO-mediated oxidative process. And Chilbokyeum is thought to have certain pharmacological effects on controlling over aging and treating Dementia. Further clinical study of this pharmacological effects of Chilbokyeum should be complemented.

석창포(石菖蒲) 전탕액(煎湯液)이 Glucose Oxidase에 의해 손상(損傷)된 대뇌피질(大腦皮質) 신경세포(神經細胞)에 미치는 영향(影響)

최공한,박승택,류도곤,최민호,허진영,강성도,고정수,양상철,성은경,조남수,이춘우,서의석,류영수,Choi, Kong-Han,Park, Seung-Taeck,Ryu, Do-Gon,Choi, Min-Ho,Hea, Jin-Young,Kang, Sung-Do,Go, Jeong-Soo,Yang, Sang-cheal,Sung, Yeun-Kyung,Cho, Nam-Su,L 대한동의생리학회 1999 동의생리학회지 Vol.14 No.2

The purpose of this study is to examine the toxic effects caused by Glucose Oxidase(GO) and the effects of herbal extracts such as Acori Rhizoma(AR) on the treatment of the toxic effects. For this purpose, experiments with the cultured cell from the cerebrums of new born mice were done. The results of these experiments were as follows. 1. GO, a oxygen radical, decreased the survival rate of the cultured cells on NR assay and MTT assay. 2. GO, a oxygen radical, decreased the amount of neurofilaments and total protein. 3. AR have efficacy of increasing the amount of neurofilament. 4. AR have efficacy of increasing the amount of total protein. From the above results, It is concluded that AR has marked efficacy as a treatment for the damages caused in the GO-mediated oxidative process. And AR is thought to have certain pharmacological effects on controlling over aging. Further clinical study of this pharmacological effects of AR should be complemented.

광대역 잡음제거를 위한 신경망 적응잡음제거기 설계

최한고,이상재,채창현 金烏工科大學校 産業技術開發硏究院 1999 産業技術開發硏究 Vol.15 No.-

This paper describes a nonlinear adaptive noise canceller(ANC) using neural networks as the nonlinear adaptive filter to make up for drawback of the conventional noise canceler with the linear adaptive filter. Performance of the proposed noise canceller is evaluated using both periodic and nonstationary signals contaminated with broadband noise, assuming that the noise of primary and reference inputs is uncorrelated. The proposed ANC is compared with the adaptive noise canceller of the linearfilter for the relative comparison of noise rejection performance. Experimental results show that the neural network based ANC outperforms the linear ANC with respect to mean square error, and the recurrent neural network adaptive filter performs the best. It is verified that the neural network adaptive filter, having highly nonlinear dynamics by virtue of the distributed nonlinearities, can be used as an effective adaptive filter for the rejection of broadband noise.

뇌파신호 디스플레이를 위한 데시메이션 기법들의 비교연구

최한고 國立金烏工科大學校附設生産技術硏究所 1994 産業技術開發硏究 Vol.10 No.-

The display of biological signals in raster scan monitors often involves a multirate sampling operation which consists of decimation and interpolation. All digitized electroencephalographic(EEG) samples cannot be displayed due to predefined formats. For instance, the EEG signal of 10 to 30 seconds(2,500 to 7,500 samples at 250[Hz] sampling frequency) must be displayed in the computer screen to keep the aspect ratio of the paper polygraph output. Since the current affordable display technology plots at most 2,000 pixels per row, some signal samples need to be discarded. This paper studies methods to perform this operation characterizing them from the signal processing view-point and compares the display quality among several decimations.

전원의 간섭잡음 억제를 위한 적응잡음 제거기법

최한고 國立金烏工科大學校附設生産技術硏究所 1993 産業技術開發硏究 Vol.9 No.-

Spectrum and timing analysis of biomedical signals have been used as effective diagnosis or remedy tools for pathological diseases. The most common problem in bioelectric recordings is the presence of interferences from electrical power sources, i.e., 60[Hz] and its harmonics. Reduction of these interferences from the recorded signals is essential to detect a particular activity of the signal or to reduce the effect of spectral overlap of both the signals. This paper describes the cancellation of the electrical interferences from respiratory electromyographic (EMG) activity using an adaptive noise calceler and some features in processing this technique. The adaptive noise calcellation technique was found to successfully reduce the interferences.

신경망을 이용한 적응잡음제거기 설계

곽우혁,최한고,박원우 金烏工科大學校 1998 論文集 Vol.19 No.-

This paper describes a nonlinear adaptive noise canceller using neural networks as the nonlinear adaptive filter, which is suitable for extracting periodic signals and nonstationary signals contaminated with broadband noise. Two neural networks. feedforward (FNN) and recurrent neural networks (RNN), are used as nonlinear adaptive filters. In this paper, it assumes that the noise of the primary input in the adaptive noise canceller is not the same characteristic as that of the reference input. Performance of the proposed noise canceller is evaluated using peripdic sawtooth signal and speech signal with white noise, and than compared with conventional adaptive noise canceller of linear structure. Experimental results show that the neural network based noise canceller outperforms the linear noise canceller. It is also verified that the RNN adaptive filer performs best with respect to noise rejection of mean square error.

GTO 게이트 드라이버를 위한 2차측 제어 ZVS-PWM 컨버터

김학성,최한고,이상재 金烏工科大學校 1994 論文集 Vol.15 No.-

전원장치의 입출력간의 전기적 절연은 전력 트랜지스포모로 제어신호는 보통 optocoupler를 이용한다. 그러나 고압절연(>10KV)인 경우는 제어신호의 절연은 트랜스포머나 광케이블을 이용하게 되는데 이들은 직류신호를 직접 전송할 수 없으므로, 출력전압을 주파수로 변환하여 전달하게 되므로 변환시간 만큼의 시간지연이 생기게되어 전원장치의 응답특성과 제어성이 나빠지게 된다. 본 논문에서는 ZVS-PWM컨버터를 고전압 절연이 요구되는 GTO 게이트 드라이브의 전원데 적용하여 온라인으로 제어하는 것과 같이 빠른 응답특성과 안전성 및 제어성을 얻기위한 2차측 제어회로를 제안하여 그 유용성을 시뮬레이션을 통하여 확인하였다.

퍼지 논리를 이용한 적응필터의 시변 수렴인자제어에 관한 연구

채석,유경종,최한고 金烏工科大學校 1995 論文集 Vol.16 No.-

The LMS(Least Mean Square) algorithm is widely used to control the adaptive filter in an adaptive noise canceler. The convergence factor in the algorithm is usually determined empirically or based on the power of input signals. This paper proposes the time-varying control of the convergence factor using fuzzy-logic. The proposed is evaluated by applying both test and respiratory myographic signals to two structures of the adaptive filter, FIR and linear combiner structures. Simulation results show that the new method outperforms fixed Finite impulse Response values of the convergence rate and the optimal method[6] in terms of the tradeoff between the convergence speed and the attenuation of the system error.

개선된 리커런트 신경망을 이용한 등화기

박정규,서동신,최한고 國立金烏工科大學校附設生産技術硏究所 1997 産業技術開發硏究 Vol.13 No.-

An adaptive equalizer in digital communication has been used to reduce intersymbol interference(ISI) and noise via radio and telephone channels. The problem of equalization can be viewed as a classification problem, which seeks to classify an observed channel output into one of the finite symbol points in the data. This paper presents the nonlinear adaptive equalizer using the modified recurrent neural networks(RNNs). The RNN structure used in this paper differs from Jordan's and Elman's RNNs with respect to feedback paths. Comparing with the conventional LMS and the feedforward neural network equalizers, the results of the RNN based equalizer show superier performance in terms of mean square error.

신경망과 TDL 구조에 의한 신호예측

이호섭,김태원,최한고,이상재 金烏工科大學校 1997 論文集 Vol.18 No.-

Neural networks have the potential in the adaptive prediction of nonstationary time series by virtue of the distributed nonlinearities and the learnings ability. This paper describes the adaptive prediction using the cascade combination of nonlinear and linear modulus. Recurrent neural networks (RNNs) and conventional tapped-delay-line (TDL) filter are used as the nonlinear and linear models respectively. For the relative comparison of prediction performance, the proposed predictor was compared with the RNN-based predictor. Simulation results show that the proposed predictor with smaller number of system weights performs better than the RNN-based predictor. We also found that it was well applicable for the prediction of the time series which contains large variations of signal amplitude.