日本生理学会大会発表要旨集
日本生理学会大会発表要旨集
セッションID: 1P1-028
会議情報
ニューラルネットワークによるシミュレーションモデル:入力条件を変化させたときの動脈圧の反応
*煙山 健仁平川 晴久晝間 恵西田 育弘
著者情報
会議録・要旨集 フリー

詳細
抄録
We have been trying to calculate of arterial pressure (AP) on simulation model by a neural network (NN) algorithm without approximate equations, to develop a new approach to estimate the relationship between AP and other factors (renal sympathetic nerve activity (RSNA), heart rate (HR) and respiration rate). Previously, we confirmed the simulated AP consisted with measured AP by the NN model. In this study, we examined whether the simulated AP would consist with measured AP when measured values of hypertensive rats were input to the learned NN by those of normotensive rats. The AP simulation model was developed using Neural Network Toolbox of MATLAB (The Mathworks, Inc.). The back propagation was selected as a learning algorithm of the layered NN. AP, HR, RSNA and respiration rate were obtained from conscious chronically instrumented rats. Those were used for the learning of NN algorithm to establish the relationship between AP and other factors. When measured values obtained from hypertensive rats were input to the learned NN by those from normotensive rats, the simulated AP were lower than measured AP. We confirmed that the relationship between AP and other factors of hypertensive rats were different from those of normotensive rats by the NN model. These results suggest that AP of hypertensive rats was not only dependent on the relationship between HR, RSNA and respiration rate. [J Physiol Sci. 2006;56 Suppl:S131]
著者関連情報
© 2006 日本生理学会
前の記事 次の記事
feedback
Top