| Using drugs or other methods to make the central nervous system and peripheral nerve network appears reversible inhibition is called anesthesia. Patients with total or partial loss of sensation will not feel pain at the time of surgery. Therefore, anesthesia is an essential key in the course of modern clinical operation. But surgery and anesthesia are risky. Physical condition of the patient, type of surgery and the experience of medical units and conditions are risk factors. Improper anesthesia may cause intraoperative awareness, postoperative pain,postoperative nausea and so on, what make things worse is that death caused by improper anesthesia in severe cases. So that controlling the risk and achieving exact anesthesia are essential, especially in the clinical. At abroad, monitoring depth of anesthesia is mainly dependent on EEG now, and the limitations of their products are still exist. However, the reality in China is that there is no mature independent product so far. That's why this paper proposes the key technology based on multi-channel EEG in depth of anesthesia monitoring system.Firstly, the research backgrounds and significances of accurate anesthesia are explained,including characteristics of EEG, common methods of processing EEG, relationship between EEG and the depth of anesthesia. Secondly, the key performance requirements of the hardware part are also presented. It emphasis on the handling of interference signals and hardware requirement analysis. The key part of the circuit is made up of signal processing,amplification, AD sampling, CPU controlling, impedance detection, rapid baseline reset, etc.The performance of the hardware of the system is tested by the third party organization and it meets the design targets well. Besides, software for PC and for modules were redesigned. The function of PC modulus include waveform display, parameters of the calculation show, patient information input, data playback and interaction with module software, etc. And it is very convenient to utilize the My SQL database to manage patient information and clinical data.The software of modulus is based on the STM32F4 microcontrollers, and it transplants u COSII embedded systems to achieve the deployment of the task and implement data collection, detect electrode impedance, interact with PC platform and so on. In addition, when ADI(Anesthesia Depth Index) algorithm for filtering the raw EEG, wavelet de-noising is used to obtain better quality EEG. Double spectral domain being used compared to the previous time domain, frequency domain of ADI parameters. The correlation between sub-parameter ‘synchslow' and the depth of anesthesia was greater in a highly conscious state. What's more, BP neural network model is established to train and validate a large number of clinical data, sub-parameters are fused to the depth of anesthesia index ADIThe calculated value of ADI algorithm were compared with the value provided by referring device BIS and the calculated parameters of ADI algorithm are at a good level. However,since the BIS sensor is expensive, ADI special sensors was designed in this paper to conduct a preliminary exploration of the sensor. Components of the ADI sensors were described in detail, and ADI sensor is superior to ECG sensor in the correlation degree and the standard deviation, which can obtain a better EEG data and consistent with the BIS sensor.Finally, a large number of clinical data collected provides the feasibility to establish the depth of anesthesia clinical databases. With BIS as a reference to analyze the Bland-Altman diagram and validate ADI system clinically, the vast majority of the points are within confidence interval. After more than 60 cases of clinical data for analysis, the average deviation of ADI and BIS is less than 6, while standard deviation is less than 7 and correlation greater than 0.75.The study has laid a good foundation for the subsequent ADI algorithm and commercialization of ADI. |
PDF Full Text Request