The estimation of neuronal population density by a robust distance method

Abstract

A new ‘nearest-neighbour’ or ‘distance’ method of estimating neurone population density is introduced. The method was originally developed for ecological studies but can be imported into histology without significant modification; changes in population density can be estimated by inverting the measure of area per unit cell (the so-called mean area). Its advantages include tests of randomness for the spatial distribution of the cells at issue and a robustness which can tolerate some departure from a random distribution pattern. To illustrate how the method is applied estimates of neurone density, in terms of ‘mean area’ per cell-point, are made on a montage tracing of the human cerebellar dentate nucleus.