目的直接动脉血压(arterial blood pressure,ABP)连续监测是侵入式的,传统袖带式的间接血压测量法无法实现连续监测。既往利用光学体积描记术(photoplethysmography,PPG)实现了连续无创血压监测,但其为收缩压和舒张压的离散值,而非ABP...目的直接动脉血压(arterial blood pressure,ABP)连续监测是侵入式的,传统袖带式的间接血压测量法无法实现连续监测。既往利用光学体积描记术(photoplethysmography,PPG)实现了连续无创血压监测,但其为收缩压和舒张压的离散值,而非ABP波的连续值,本研究期望基于卷积神经网络-长短期记忆神经网络(CNN-LSTM)利用PPG信号波重建ABP波信号,实现连续无创血压监测。方法构建CNN-LSTM混合神经网络模型,利用重症监护医学信息集(medical information mart for intensive care,MIMIC)中的PPG与ABP波同步记录信号数据,将PPG信号波经预处理降噪、归一化、滑窗分割后输入该模型,重建与之同步对应的ABP波信号。结果使用窗口长度312的CNN-LSTM神经网络时,重建ABP值与实际ABP值间误差最小,平均绝对误差(mean absolute error,MAE)和均方根误差(root mean square error,RMSE)分别为2.79 mmHg和4.24 mmHg,余弦相似度最大,重建ABP值与实际ABP值一致性和相关性情况良好,符合美国医疗器械促进协会(Association for the Advancement of Medical Instrumentation,AAMI)标准。结论CNN-LSTM混合神经网络可利用PPG信号波重建ABP波信号,实现连续无创血压监测。展开更多
Objective To assess the possibility of using arterial pressure waveform or pulse oximetry plethysmographic waveform variation to estimate the pulmonary arterial wedge pressure (PAWP) Methods Fourteen American Socie...Objective To assess the possibility of using arterial pressure waveform or pulse oximetry plethysmographic waveform variation to estimate the pulmonary arterial wedge pressure (PAWP) Methods Fourteen American Society of Anesthesiologists grade Ⅰ-Ⅱ patients aged 33-69 years and weighing 62 0±9 5 kg scheduled for elective abdominal tumor surgery were studied Their hemoglobin exceeded 120 g/L and hematocrit exceeded 35% Pre operative acute hypervolemic hemodilution was applied immediately after general anesthestic induction and tracheal intubation PAWP, systolic pressure variation (SPV), delta down (dDown), SPV plet , dDown plet and other hemodynamic parameters were measured and recorded when total fluid volume (crystalloid and colloid) infused reached 10 ml/kg and 20 ml/kg and again at the end of the operation Central venous pressure was maintained at 10-12 mm Hg during operation Systolic blood pressure at the end of Valsalva maneuver (airway pressure was kept at 22 mm Hg) and the systolic pressure before the Valsalva manoeuvre during apnea were used to calculate arterial pressure ratio (APR) Results APR, SPV, dDown, SPV plet and dDown plet all correlated well with PAWP ( r =0 717, -0 695, -0 680, -0 522 and -0 624 respectively, P <0 01) There was a closer linear correlation between APR and PAWP than between the other parameters The regression equation was PAWP (mm Hg)=0 207×APR (%)-0 382 Conclusion During positive pressure mechanical ventilation, APR, SPV, dDown, SPV plet and dDown plet can be used to estimate PAWP effectively展开更多
文摘目的直接动脉血压(arterial blood pressure,ABP)连续监测是侵入式的,传统袖带式的间接血压测量法无法实现连续监测。既往利用光学体积描记术(photoplethysmography,PPG)实现了连续无创血压监测,但其为收缩压和舒张压的离散值,而非ABP波的连续值,本研究期望基于卷积神经网络-长短期记忆神经网络(CNN-LSTM)利用PPG信号波重建ABP波信号,实现连续无创血压监测。方法构建CNN-LSTM混合神经网络模型,利用重症监护医学信息集(medical information mart for intensive care,MIMIC)中的PPG与ABP波同步记录信号数据,将PPG信号波经预处理降噪、归一化、滑窗分割后输入该模型,重建与之同步对应的ABP波信号。结果使用窗口长度312的CNN-LSTM神经网络时,重建ABP值与实际ABP值间误差最小,平均绝对误差(mean absolute error,MAE)和均方根误差(root mean square error,RMSE)分别为2.79 mmHg和4.24 mmHg,余弦相似度最大,重建ABP值与实际ABP值一致性和相关性情况良好,符合美国医疗器械促进协会(Association for the Advancement of Medical Instrumentation,AAMI)标准。结论CNN-LSTM混合神经网络可利用PPG信号波重建ABP波信号,实现连续无创血压监测。
文摘提出一种检测动脉血压(ABP)波起始点位置和峰值位置的迭代算法。该方法应用斜坡和函数处理原始信号,以迭代的方法标识每个波动周期的起始位置和峰值位置。迭代计算时,考虑样本数据值的统计特性、数据序列间距关系。在对Massachusetts General Hospital/Marquette Foundation(MGH/MF)Waveform Database的数据进行算法测试后表明:该迭代算法具有快速、有效性和强壮性。
文摘Objective To assess the possibility of using arterial pressure waveform or pulse oximetry plethysmographic waveform variation to estimate the pulmonary arterial wedge pressure (PAWP) Methods Fourteen American Society of Anesthesiologists grade Ⅰ-Ⅱ patients aged 33-69 years and weighing 62 0±9 5 kg scheduled for elective abdominal tumor surgery were studied Their hemoglobin exceeded 120 g/L and hematocrit exceeded 35% Pre operative acute hypervolemic hemodilution was applied immediately after general anesthestic induction and tracheal intubation PAWP, systolic pressure variation (SPV), delta down (dDown), SPV plet , dDown plet and other hemodynamic parameters were measured and recorded when total fluid volume (crystalloid and colloid) infused reached 10 ml/kg and 20 ml/kg and again at the end of the operation Central venous pressure was maintained at 10-12 mm Hg during operation Systolic blood pressure at the end of Valsalva maneuver (airway pressure was kept at 22 mm Hg) and the systolic pressure before the Valsalva manoeuvre during apnea were used to calculate arterial pressure ratio (APR) Results APR, SPV, dDown, SPV plet and dDown plet all correlated well with PAWP ( r =0 717, -0 695, -0 680, -0 522 and -0 624 respectively, P <0 01) There was a closer linear correlation between APR and PAWP than between the other parameters The regression equation was PAWP (mm Hg)=0 207×APR (%)-0 382 Conclusion During positive pressure mechanical ventilation, APR, SPV, dDown, SPV plet and dDown plet can be used to estimate PAWP effectively
