- 期刊
The Intelligent Model of a Patient Using Artificial Neural Networks for Inhalational Anaesthesia
利用類神經網路來建立吸入性麻醉的智慧型病人模型
摘要
為了要模擬吸入性麻醉氣體的手術過程，本論文利用類神經網路的學習法則來建立一個智慧型病人模型。此模型共包括四種輸入（即病人年紀、體重、性別和吸入性麻醉氣體濃度）和四種輸出（即病人心跳、血壓、呼出麻醉氣體濃度和腦波）。除了利用代價函數（cost function）外，並也考慮藥物學的特性來評估病人模型的準確性。臨床上以八個病人來做類神經網路的學習，再以十一個病人的數據來測試此病人模型的好壞。結果顯示，本論文所提的建立病人模型方法，其強健性與容忍病人與病人間的差異性均有不錯的效果。
關鍵字
無資料並列摘要
In order to simulate an entire operation during inhalational anaesthesia, a patient model that includes four inputs, which are the patient’s age, weight, gender and anaesthetic agent concentration, and four outputs, which are the heart rate (HR), systolic arterial pressure (SAP), end-tidal anesthetic agents (Etaa), and electroencephalograph (EEG) signals (i.e., the bispectral index), was constructed in this study using artificial neural networks (ANN). The assessment of the performance of patient model presented here is based not only on the minimum cost function, but also on pharmacological reasoning. Eight patients were trained using the weighting function of ANN. After the weights were finalized, eleven more patients were tested and compared with the previous eight patients. The average cost function for the training patients (i.e., 8 patients) and testing patients (i.e., 11 patients) was 0.0834±0.0577 and 0.0563±0.0450, respectively. The results show that this approach can provide a more robust model despite the considerable individual variation in inhalational anaesthesia among patients.