Benignus V A, Boyes W K, Hudnell H K, Frey C M, Svendsgaard D J
U.S. Environmental Protection Agency, Human Studies Division, Research Triangle Park, NC 27711.
Neurosci Biobehav Rev. 1991 Spring;15(1):165-71. doi: 10.1016/s0149-7634(05)80110-1.
Cross-species extrapolation will be defined as prediction from one species to another without empirical vetification. Cross-species mapping (CSM) is the same except empirical vetification is performed. CSM may be viewed as validation of methods for extrapolation. Algorithms for CSM may originate from theory, from empirical observations or a combination of the two. Regardless of their origins, CSM algorithms must be explicated and confidence intervals given around their predictions. This paper offers a quantitative method for constructing CSM equations which is useful in evaluation of the CSM and as an aid in the design of new experiments in CSM and extrapolation. The method requires fitting mathematical models for the physiological or behavioral phenomena to be mapped across species. A CSM equation can then be derived from the models in each species and approximate confidence limits may be obtained for predictions from the equation. The method is useful even when the models in the two species differ in form, implying differences in physiology or behavioral principles between species. The method proposed has a number of remaining uncertainties and possible problems.
跨物种外推将被定义为在没有实证验证的情况下从一个物种到另一个物种的预测。跨物种映射（CSM）与之相同，只是进行了实证验证。CSM可被视为外推方法的验证。CSM算法可能源于理论、实证观察或两者的结合。无论其来源如何，CSM算法都必须明确阐述，并给出其预测的置信区间。本文提供了一种构建CSM方程的定量方法，该方法在评估CSM以及辅助设计CSM和外推的新实验方面很有用。该方法需要为跨物种映射的生理或行为现象拟合数学模型。然后可以从每个物种的模型中导出CSM方程，并为该方程的预测获得近似置信限。即使两个物种的模型在形式上不同，这意味着物种之间在生理或行为原理上存在差异，该方法也很有用。所提出的方法仍存在一些不确定性和可能的问题。
