Controlling for dogs' (Canis familiaris) use of nonmnemonic strategies in a spatial working memory task.

Short assessments of spatial working memory (SWM) in dogs are becoming popular evaluations of canine aging and individual differences. In a typical SWM task, an experimenter hides a reward inside of a bucket at a specific stimulus position while the dog watches. Then, following a varying delay interval, the dog is released to choose a bucket. The longest delay at which the dog can successfully choose the bucket containing the reward is considered to reflect the dog's SWM duration. Although past studies were informative, the tasks often lacked a valid measure of SWM due to dogs' ability to use nonmnemonic strategies, such as body orientation or sustained attention, to successfully solve these tasks without relying on working memory mechanisms. Therefore, we designed the first study to assess the internal validity of these tasks by directly comparing dogs' performance on two experimental conditions. We found that dogs performed worse in a control condition in which nonmnemonic strategies were eliminated compared with a typical SWM task condition. In addition, our results indicate a strong relationship between the percentage of delay time that a dog spends orienting its head or body to the correct bucket and performance in a typical SWM task. These findings were the first to show a difference in SWM performance when dogs' use of nonmnemonic strategies was controlled and stress the importance of considering the internal validity of these tasks if used to examine SWM in future work. (PsycInfo Database Record (c) 2021 APA, all rights reserved).