The naming of neurons: applications of taxonomic theory to the study of cellular populations.

For many purposes, biologists must study large brains through groups of similar neurons, since these populations - not individual cells - are the smallest units for which exact counterparts can be recognized unequivocally across a series of brains. One who surveys singel neurons, by whatever techniques, may discern major aspects of a tissue's organization by classifying the elements studied, thereby performing an exercise in taxonomy at the cellular level. The discovery of neuronal types is best achieved by imitating the naturalist who seeks new biological species: a large sample of cells is gathered by a regular, widely effective method, and an effort is made to understand the biases in the sampling procedure; a numerous and diverse set of features is observed for each neuron encountered; and the cell sets recognized are described in agreement with the polythetic concept of natural groups. The resulting multidimensional population descriptions, the most useful of which include the temporal information available through electrophysiologic recording, may be quite powerful for testing circuit hypotheses about the large nervous system.