Computer based three-dimensional reconstruction and quantitative analysis of neurons and dendritic bundles in the cerebral cortex

Abstract

The cerebral cortex is the layer of gray matter that covers the cerebral hemispheres. It is just a few millimeters thick and consists of a complex network of neurons, glial cells and capillaries. It is generally accepted that quantitative data that describe the occurrence and distribution of cortical neurons is fundamental for elaborating valid models of cortical circuits. Despite this fact, there is only meager and varying information about the number of neurons in different parts of the nervous system and there are controversies about such basic figures as the total number of neurons in the human cerebral cortex and the number of neurons under 1 mm2 of cortical surface in different areas.One of the most influential ideas about the cortical organization is the compartmentalization of cortical cells and their connections into vertical modules. With the exception of cortical barrels, it has been difficult to find structural correlates to the physiological modules. However, there is in the cortex one vertically oriented structural element, which has been proposed to be the modular core component: the dendritic bundle. A dendritic bundle is defined as a non-random group of at least 3 apical dendrites of layer V pyramidal cells. The steric and quantitative relationships of the dendritic bundles and the surrounding neurons are not fully clarified.This thesis is a part of a joint medical-engineering (image analysis) project aiming at the development of a system for computer-generated to-scale quantitative and qualitative three-dimensional reconstructions of nervous tissue in general and the cerebral cortex in particular. As we are concerned with the ultrastructural level we have implemented preparatory- and program procedures that allow reconstructions based on series of both light- and electron microscopic images. During my work in this joint-project I have focused on two issues of the analysis of the cerebral cortex; aspects of quantitative analysis of neurons and the use of an aligned stack of images of serial sections for the reconstruction of dendrites.For the quantitative analysis we have developed a method for stereological analysis of the cerebral cortex. The method is based on aldehyde-fixed cortex which was sectioned into 50-mm-thick sections. The sections were stained with Richardson´s solution, which prevented shrinkage and stained the neurons and left glial cells unstained (Paper I). A computer-based unbiased counting method was implemented by taking advantage of the stereological procedure referred to as the disector . When applying the method for quantitative analysis of neurons on the rat cerebral cortex, our data showed that there were significant differences between primary motor, primary somatosensory and secondary visual cortex (areas Fr1, HL and Oc2) in both the number of neurons under a certain area of cortical surface and in the neuronal packing density (Paper II). These results challenge the uniformity concept proposed by Rockel et al. (1980). When reconstructing the laminar pattern of rat primary motor cortex, our analysis revealed, in contrast to the current view, the presence of a genuine layer IV consisting of densely packed small neurons (Paper III).The effects of growth hormone deficiency on the cortical volume and the gross organization of primary somatosensory and motor cortex in the mutant little mouse were also investigated (Paper IV). The little strain produces, in its homozygous state (lit/lit), characteristics similar to those seen in human isolated growth hormone deficiency. Our results show that the volume and the neuron density of the little mouse was about 89% and 115% in comparison to that of the controls, respectively. No significant difference in the size of the neuronal cellbodies was found. Our results indicate that there is a reduction in the space between the neuronal somas due to the growth hormone deficiency, in the primary somatosensory and motor cortex of the little mouse.When using the aligned stack of images of serial sections for the reconstruction of dendrites, we found that cinematographic display of the image stack, gave the impression of "traveling through" the tissue and facilitated the possibility to follow structures extending more or less perpendicular to the sectioning plane (Paper V). This method was then used to study the steric relationship of dendritic bundles of layer V pyramidal cells and the surrounding neurons, in the primary somatosensory cortex of the rat (Paper VI). The occurrence of the bundles in tangential sections through layer IV was about 193 bundles/mm2 and the average number of neurons per bundle was estimated to approximately 600. However, when calculating Voronoi-diagrams, the number of neurons, which with this mathematical technique will be ascribed to each of the reconstructed dendritic bundles, varied between 200 and 1000. The possibility that the dendritic bundles are centers in cortical cell modules is discussed.