Structure and function of gustatory neurons in the nucleus of the solitary tract. I. A classification of neurons based on morphological features

Prior investigations in other laboratories have provided convincing evidence that the neurons of the rostral nucleus of the solitary tract (rNST) can be grouped according to their physiological response properties or morphologic features. The present study is based on the premise that the response properties of gustatory neurons are related to, and perhaps governed by, their morphology and, connectivity. In this first phase of our ongoing investigation of structure‐function relationships in the rNST of the rat, we have used intracellular injection of neurobiotin to label individual physiologically characterized gustatory neurons. A total of 63 taste‐sensitive neurons were successfully labeled and subjected to three‐dimensional quantitative and qualitative analysis. A cluster analysis using six morphologic features (total cell volume, soma area, mean segment length, swelling density, spine density, and number of primary dendrites) was used to identify six cell groups. Subsequent analyses of variance and posthoc comparisons verified that each of these six groups differed from all others with respect to at least one variable, so each group was “typified” by at least one of the six morphologic features. Neurons in group A were found to be the smallest neurons in the sample. The cells in group B had small somata and exhibited the highest swelling density of any group. Group C neurons were distinguished by dendrites with long, spine‐free branches. These dendrites were significantly longer than those of any other group except Group F. The neurons in group D had more primary dendrites; than any other group. Group E neurons possessed dendrites with the lowest swelling density but the most spines of any group. The cells in group F were the largest neurons in our sample and possessed the largest somata of any group. Thus, overall cell size and density of dendritic spines and swellings were found to be particularly important variables in this classification scheme. Our preliminary results suggest that the number and density of dendritic spines (as well as other morphologic features) may be related to a given neuron's most effective stimulus, indicating that it will indeed be possible to use the criteria established in the present investigation to derive structure‐function relationships for gustatory neurons in the rNST. © 1994 Wiley‐Liss, Inc.