Approaches to Design-Based Stereology
In contrast to two-dimensional image analysis, the goal of modern design-based stereology is to provide an unbiased estimate of the absolute number or size of measured objects within a three-dimensional region of interest, as opposed to a relative value (e.g., number of objects per unit area, i.e., density). This approach allows valid comparisons to be made within and among studies, while avoiding the “reference trap” pitfall, in which the derived answer is based on the interplay of multiple variables (e.g., density is a function of both the number of objects and the volume they occupy). However, the ability of a particular stereological probe to obtain accurate estimates is dependent on the spatial parameters selected for sampling, especially if the objects to be measured are not distributed uniformly throughout the prepared specimen. Consequently, it is common practice to conduct a brief pilot study using control specimens prior to the main investigation, in order to establish appropriate sampling parameters. Special algorithms in software such as MBF’s Stereo Investigator allow the user to oversample and re-sample control specimens to determine optimal parameters for use in the definitive study.