Odor–visual and visual–visual matching to sample with dogs

Abstract

Matching-to-sample (MTS) procedures are commonly employed to study learning and behavior, but they have rarely been used with dogs. One MTS procedure that might have applied and experimental value is an odor–visual matching procedure in which selection of a specific visual stimulus is reinforced following exposure to a specific odor. We evaluated such a procedure using an automated odor delivery apparatus and touchscreen apparatus for selection of arbitrary shapes. In Experiment 1, four dogs performed with above-chance accuracy when working with two odors, but only one dog met criteria to progress to training with a third odor. With the goal of improving performance, Experiment 2 involved the same procedures but added an error-correction procedure and an errorless-learning procedure. The errorless-learning procedure may have been counterproductive, as dogs performed with chance accuracy in the MTS task and failed to complete all steps of the errorless-learning procedure. In Experiment 3, we evaluated dogs' ability to perform a visual–visual matching task based closely on the procedures used in Experiment 2. This investigation produced similar findings, but two of four dogs performed with above-chance accuracy, suggesting that visual–visual matching may be established more readily than odor–visual matching.