Two assays of working memory in companion dogs: The holeboard and disappearing object tasks

Abstract

Variation in executive function and age-related cognitive decline may underlie the emergence of behaviour and welfare problems in dogs. A better understanding of such links, and of dog cognition in general, will be facilitated by the development of cognitive tasks that can be readily implemented, including with publicly-owned dogs that are available for relatively short testing periods. Working memory is a key component of cognitive and executive function that is often measured using tests such as delayed-non-match-to-sample or radial-arm maze which require extensive training and testing. Here we successfully adapt the Holeboard Task to measure working and reference memory in dogs, and show that another test of working memory, the Disappearing Object Task, can be performed in a single day. Working memory (p = 0.002) and reference memory (p < 0.001) scores in a 16-hole Holeboard Task (four holes baited) increased across sessions, with reference memory scores falling steeply as expected when the configuration of baited buckets changed. In the Disappearing Object Task dogs were able to successfully locate an object displaced behind one of four visual barriers, and their ability to do this fell as the memory retention interval (0 s, 30 s, 60 s, 120 s, 240 s) between hiding and locating the object increased (p < 0.001). Holeboard and Disappearing Object working memory measures were not correlated, possibly due to differences in the motivational context and exact learning demands of the tasks. In summary we show that the Holeboard Task can be adapted for use in dogs and that the Disappearing Object Task can be implemented in a single day. The latter task may be particularly useful for working memory studies of dogs owned by the public where prolonged access is often infeasible, and the three-day Holeboard Task is faster to implement than other commonly used laboratory-based tasks.