Automation of Pattern Discrimination in the Rat

Abstract

Investigation of visual discriminative processes in animals has been facilitated by the use of automatic equipment and improved techniques of behavioral control. Visual generalization gradients and continuous changes in absolute threshold have been measured with new precision (Guttman, 1956; Blough, 1958). As part of a program evaluating the pathological effects of prenatally administered ionizing radiation (Hicks, D'Amato, & Lowe, 1959), the following schedule was used in determining the capacity for pattern vision. Animals were maintained at 70 to 80% of their free-feeding weight and trained to press either of two transparent plastic pigeon keys (Ralph Gerbrands Co.) with their noses in order to obtain a 45-mg. food pellet (P. J. Noyes Co.). After one or two sessions of continuous reinforcement, the keys were transilluminated one at a time and circuitry was arranged so that 15 responses to an unlit key produced only the time-out condition. To avoid superstitious key responses during the time-out period, the occurrence of all such responses reset the 15-second time-out period. With daily sessions of 2-hr. the animals earned from 110 to 200 pellets and performed the brightness discrimination problem with few errors within one to two days. Then the keys were simultaneously transilluminated with patterns (One-plane digital display unit, Model No. 10219, Industrial Electronic Engineers, North Hollywood), one positive (vertical striations) and one negative (horizontal striations). with the same reinforcement schedule in effect. The left. . right position of the positive stimulus followed a pre-arranged random sequence. Fifteen responses to the negative pattern produced time-out and the same left-right arrangement was presented again on the following trial (a correction procedure). The animal sets its own inter-trial interval by its work rate. Control sessions comprised the same schedule without the key stimuli being differential; a square pattern illuminated both keys. We have demonstrated pattern vision in normal rats and rats with cortexes characteristic of irradiation on the 13, 15, 17, or 19th fetal days, at or better than the 90% correct level in three to nine weeks. The occasional rat will not develop the discrimination under these conditions. We consider the arrangement workable but not