In principle, digital acquisition of cell-count data from serially-sectioned immunocytochemical material is a straightforward enterprise. First, a serial brain section is magnified by use of a microscope interfaced to a computer. Then, using appropriate hardware and software, a digital image is captured, and cellular profiles of interest are segmented from background objects according to mean grayscale intensity and pixel area. Ideally, the cells of interest would be uniformly distinguishable from other objects or areas of the image, with respect to grayscale intensity and size. However, due to non-uniformity in background staining of neuropil, immunocytochemical material often departs markedly from this ideal situation. As a consequence, determining grayscale intensity and cell size cutoff values which separate cells of interest from background becomes laborious and arbitrary. This problem can be diminished by increasing the magnification of the digitized image, which increases the figure-ground resolution of the image. However, high-magnification images make tissue navigation difficult and require that multiple images be captured. This paper describes a two focal plane procedure for obtaining cell counts from nuclear-stained immunocytochemistry material. This procedure allows the capturing and cell counting of relatively low-magnification images with high digital figure-ground resolution.
Copyright 1998 Elsevier Science B. V. All rights reserved.