Introduction

Classical neuropathology implements gross and microscopic methods to determine diagnosis and neuroanatomical localization of pathology. Diagnostic standards for dementing illnesses have been refined as semi-quantitative grading systems for dementia pathology, allowing rough clinical-pathological correlations. These staging systems include the Consortium to Establish a Registry for Alzheimer's Disease (CERAD) and the Braak and Braak criteria for Alzheimer's disease (AD), and consensus neuropathological criteria for dementia with Lewy bodies (DLB) (Braak & Braak, 1991; Mirra et al., 1991; McKeith et al., 1996). For more detailed quantitative analysis of the neuropathology of dementia, sensitive morphometric methods have been developed. The dominant method in recent years is stereology, which applies a series of rules to overcome biases in counting objects (e.g. neurons) on a slide (Sterio, 1984; Gundersen et al., 1988a,b). Stereological approaches to the assessment of anatomical volumes, neurite length, synapse counts, neuron counts and neuron size have utilized such varied resources as gross brains, MRI scans, light microscopy, confocal microscopy, and electron microscopy (Geinisman et al., 1996; Everall et al., 1999; Peterson, 1999; Roberts et al., 2000). This chapter reviews principles of stereology as applied to anatomical volume measurements and neuron counting at the light microscope level, with a particular emphasis on AD. Other aspects of stereology, such as using the rotator or nucleator for particle volume and cycloid intercepts for surface area, can be found in recent reviews and texts (Gundersen et al., 1988a; Howard & Reed, 1998).

Profile-counting methods

Since it is inefficient to count all the neurons in a particular brain region, methods have been developed to estimate neuron number from a sampling of microscopic fields.