A computerized image processing system is described that assists the neurobiologists in analyzing data from 2-DG autoradiography by providing for: (1) Rapid fine-scale digitization of gray levels using a TV camera (2) The recognition of and verification of subtle differences in optical density with the aid of color windows (3) the superimposition of the autoradiographic image upon the histological image, so that the activity seen in the autoradiograph can be accurately assigned to anatomically defined structures (4) The production of numerical data suitable for statistical analysis and line drawings suitable for black on white publication (5) The relating of local gray level to a norm for the image as a whole, so as to remove the variability introduced by variations in section thickness, in the amount of 2-DG seen by the brain during incorporation, in level of anesthesia, etc. If the localized darkening in autoradiographic images is being used as an index of localized functional activity rather than as a measure of metabolism, normalization obviates the need to obtain arterial blood samples. These routines permit anatomically accurate numerical analysis of autoradiographs without any constraints on the experimental situation.