体表标测系统的导联优选法

Optimal Selection of Electrocardiographic Leads for Body Surface Mapping System

在体表标测系统中合理地选择合适的电极数量和位置,其临床应用的推广意义重大.为此,开发了一种由多导联体表记录选择最优导联位置的新算法.采用基于最小二乘法的多变量线性回归算法确定变换系数,将原始信号与重构势的空间残差均方值作为优化判据,分别用顺序前进法和前进-后退搜索法对奇异值分解(singular value decomposition,SVD)前后的体表势矩阵寻找最优导联组合.结果表明:所得到的4种最优导联位置不尽相同,但分区统计规律相同;同时,优选出的最优导联组合对测试样本的预测最大相对误差均在1%左右,表现出较好的适应性.

Selection of the right number and location of electrodes in body surface potential mapping system is significant in routine clinical practice. A new algorithm was developed to select the optimal electrode positions for multi-lead body surface recording. The transformation coefficients were determined using multiple linear regression by least square method.The criterion was the spatial RMS error to compare original and reconstructed potentials.The sequential forward and forward-backward selections were applied to find the best combination of electrocardiographic leads for the body surface potential matrix before and after singular values decomposition （ SVD ） . The results show that four configurations of the optimal electrode positions are different but have the same statistical rule. At the same time the maximum prediction relative errors by using these configurations are all about 1%, which has better applicability.