Veldhuis J D, Johnson M L
Department of Internal Medicine, University of Virginia Health Sciences Center, Charlottesville 22908.
Biotechniques. 1990 Jun;8(6):634-9.
The availability of increasingly efficient computational systems has made feasible the otherwise burdensome analysis of complex neurobiological data, such as in vivo neuroendocrine glandular secretory activity. Neuroendocrine data sets are typically sparse, noisy and generated by combined processes (such as secretion and metabolic clearance) operating simultaneously over both short and long time spans. The concept of a convolution integral to describe the impact of two or more processes acting jointly has offered an informative mathematical construct with which to dissect (deconvolve) specific quantitative features of in vivo neuroendocrine phenomena. Appropriate computer-based deconvolution algorithms are capable of solving families of 100-300 simultaneous integral equations for a large number of secretion and/or clearance parameters of interest. For example, one application of computer technology allows investigators to deconvolve the number, amplitude and duration of statistically significant underlying secretory episodes of algebraically specifiable waveform and simultaneously estimate subject- and condition-specific neurohormone metabolic clearance rates using all observed data and their experimental variances considered simultaneously. Here, we will provide a definition of selected deconvolution techniques, review their conceptual basis, illustrate their applicability to biological data and discuss new perspectives in the arena of computer-based deconvolution methodologies for evaluating complex biological events.
日益高效的计算系统的出现，使得对复杂神经生物学数据（如体内神经内分泌腺分泌活动）进行原本繁重的分析变得可行。神经内分泌数据集通常稀疏、有噪声，且由在短时间和长时间跨度上同时运行的联合过程（如分泌和代谢清除）生成。卷积积分的概念用于描述两个或多个联合作用过程的影响，它提供了一种有用的数学结构，可用于剖析（反卷积）体内神经内分泌现象的特定定量特征。适当的基于计算机的反卷积算法能够为大量感兴趣的分泌和/或清除参数求解100 - 300个联立积分方程的方程组。例如，计算机技术的一个应用使研究人员能够反卷积具有代数可指定波形的统计显著潜在分泌事件的数量、幅度和持续时间，并使用同时考虑的所有观测数据及其实验方差，同时估计特定受试者和特定条件下的神经激素代谢清除率。在此，我们将给出选定反卷积技术的定义，回顾其概念基础，说明其在生物学数据中的适用性，并讨论基于计算机的反卷积方法在评估复杂生物事件领域的新观点。
