Department of Clinical and Biomedical Engineering, Rikshospitalet, Oslo University Hospital, Oslo, Norway.
Ann Biomed Eng. 2013 May;41(5):1074-83. doi: 10.1007/s10439-013-0743-4. Epub 2013 Jan 17.
Skin conductance (SC) has previously been reported to correlate strongly with sweat rate (Swr) within subjects, but weakly between subjects. Using a new solution for simultaneous recording of SC, skin susceptance (SS) and skin potential (SP) at the same skin site, the aim of this study was to assess how accurately sweat production can be estimated based on combining these electrical properties of skin. In 40 subjects, SC, SS, SP and Swr by skin water loss was measured during relaxation and mental stress. SC and Swr had high intraindividual correlations (median r = 0.77). Stepwise multilinear regression with bootstrap validation lead to a sweating estimation model based on the sum of SC increases, the SP area under the curve and the SS area under the curve, yielding an interindividual accuracy of R(2) = 0.73, rmse = 12.9%, limits of agreement of +27.6, -30.4% and an intraclass correlation coefficient of 0.84. Bootstrapping of training and test-sets gave median rmse = 15.4%, median R(2) = 0.66. The model was also validated for intraindividual variability. The results show that estimation of sweating is significantly improved by the addition of SS and SP measurement.
皮肤电导 (SC) 先前已被报道与受试者内部的出汗率 (Swr) 有很强的相关性，但与受试者之间的相关性较弱。本研究使用一种新的解决方案，可同时记录同一皮肤部位的皮肤传导率 (SS) 和皮肤电位 (SP)，旨在评估基于这些皮肤电特性的组合，汗液生成量的估计值的准确度。在 40 名受试者中，在放松和心理应激期间测量了 SC、SS、SP 和皮肤失水的 Swr。SC 和 Swr 具有很高的个体内相关性（中位数 r = 0.77）。逐步多元线性回归和 bootstrap 验证导致了一种基于 SC 增加总和、SP 曲线下面积和 SS 曲线下面积的出汗估计模型，得出个体间准确性 R(2) = 0.73，rmse = 12.9%，一致性界限为+27.6%、-30.4%，内类相关系数为 0.84。训练集和测试集的 bootstrap 分析得出的中位数 rmse = 15.4%，中位数 R(2) = 0.66。该模型还针对个体内变异性进行了验证。结果表明，通过添加 SS 和 SP 测量，出汗的估计值显著提高。
