Description and Classification of Neuronal Structure in the Frog Retina

The intent of this study is to develop and apply an objective method for the description and classification of structure in a biological system for use in the correlation of structure and function. The biological system of interest is the frog retina, in particular, the neuronal structures which ramify in the inner plexiform layer of the retina. A tree grammar is developed for the description of the j-dimensional structure of Golgi impregnated retinal neurons seen in light microscope examination. This descriptive method includes 1) an optical sectioning procedure for automatic digitization of 3-dimensional light microscope images, 2) various methods for extracting the neuronal image from noise, 1) the extraction from stereo pairs of the primitive elements of the neuronal tree grammar, 4) the use of the grammar as a tool for investigating the occurrence of simple or complex substructures in a neuronal structure, and 5) the use of the grammar as a tool for the quantitative classification of neurons. The Weiner - Lee method for characterizing nonlinear systems is applied to the various cells of the frog retina using a two input, spot-annulus white noise stimulus. These functional studies include Proc1on dye injection of the neurons whose kernels are obtained. Procion dye injection provides a way of linking the functional descriptions to the Golgi-obtained structural descriptors. Correlation of structure and function is approached through comparisons of structural reconstructions and functional identifications. In short this study employs quantitative descriptors and classifiers to correlate structure and function in the frog retina.