Computer methods in neuroanatomy: determining mutual orientation of whole neuronal arbors

In most neurons orientation can be recognized because their arbors are more or less polarized and/or flattened. These are morphological characteristics of great functional importance. This paper deals with three-dimensional display and mathematical definition of orientation planes and vectors in whole arbors. An orientation plane can be derived from the flattest rectangular prism with which it is possible to enclose the arbor, or may be found by best-fit least square determination (based on all digitized points of the arbor). Both approaches allow description and comparison, in quantitative terms, of the orientation of neurons under various normal, pathological or experimental conditions.