Development of a protocol for studying object recognition memory in the dog.

1. Dogs had considerable difficulty learning a delayed-non-matching-to sample task at a short delay (approximately 5 seconds) for an extended period (900 trials). Only 3 of 19 dogs met the learning criterion. 2. Acquisition on the DNMS task was markedly improved when a pause was introduced on presentation of the stimulus objects, when the objects were approximately 30 cm from the dog; eleven of 16 dogs learned the task within 600 trials. 3. Dogs learned the task more rapidly at 20 and 30 second delays than at 10-second delays. This indicates a transfer of learning. 4. Dogs that did learn the task were able to perform at accuracy greater than 85% at delays of 150 and 200 seconds. At a 5-minute delay, performance was at 75%. 5. When the animals were switched to a repeated object paradigm, accuracy markedly declined. 6. The improved performance produced by introduction of the pause is attributable to: (1) presenting the object at a distance longer than the dogs' near point, and (2) allowing increased processing time.