Stereology of nerve cross sections.

Quantitative morphological data from nerve cross sections are of great value in experimental, developmental, and pathological studies. Efficient unbiased stereological techniques to estimate the total numbers and absolute size distributions of both myelinated and unmyelinated axons from a small sample of axons are described and evaluated. Axon numbers are estimated with the fractionator technique. Axon areas are estimated with the 2D-nucleator and the point counting technique for myelinated and unmyelinated axons, respectively. The axon perimeter is estimated by use of a line grid, and the myelin sheath thickness by direct orthogonal measurements in uniform, random locations. The empirically evaluated variance of the number estimates of myelinated axons was approximately approximately 0.05 when approximately 200 axons were counted. The empirical variance of area estimates obtained with a four-way 2D-nucleator was approximately 0.10 and contributed less than 0.5% to the observed total variance of axon areas within an animal. A light microscope modified for stereology with computer assisted stereological test systems and a motorised stage facilitate the acquisition of data for the myelinated axons, which is of primary interest in most studies. The unmyelinated axons are evaluated at the more time consuming electron microscopical level.