Measuring social attention skills in pet dogs.

In human social settings, dogs need good attention skills to deal with complex social interaction. It has been suggested that deficits in attention skills that compromise the relationship with the owner could be associated with some canine behavioral problems. Methods to assess selective social attention toward the owner are lacking. The aim of this study was to develop a tool to measure social attention skills in pet dogs. Twentyfive privately owned dogs were videorecorded during 2 tasks of increasing difficulty. In task A (N525), the dog observed the owner (O) and an unfamiliar person (S) repeatedly entering the experimental room and leaving through alternate doors over a period of 120 seconds. Task B (N 5 18) had the same protocol, except that people wore hoods that totally covered their heads. At the end of the task, the dogwas let free to move toward the last door used to exit the room by the owner (OD) or the unfamiliar person. Care was taken that dogs did not rely on olfactory cues to choose the direction. Task order and last exit door were randomly assigned. Attention skills were measured as the relative gaze duration at O, the average gaze bout length at O, and the direction chosen at the end of the task. Finally, inattention and activity–impulsivity of the dog were assessed using the questionnaire developed by Vas et al. (2007); the Italian translation version had been properly validated. In task A, dogs’ attention was directed significantly more at O than at S (relative gaze duration 5 72.9% 6 14.5 vs 18.2% 6 11.3; gaze bout length 5 4.2 s 6 1.9 vs 1.1 s 6 0.5, respectively; independent samples t test, P , 0.001). Similar results were obtained for task B (relative gaze duration 5 68.5% 6 12.0 vs 23.7% 6 9.9; gaze bout length 5 3.8 s 6 1.6 vs 1.4 s 6 0.6, respectively; P , 0.001). Nevertheless, within-dog comparisons revealed an increased attention toward S in task B (paired samples t test, P , 0.05), especially when S and O were inside the room (relative gaze duration 5 24.7% 6 10.5 vs 17.5% 6 10.0; P 5 0.022), whereas relative gaze duration at O significantly decreased (P , 0.05). Reduced attention at O in task B was confirmed by the inability to focus on OD and move toward it, which was not the case during task A (binomial test, P , 0.005). Interestingly, the only 4 dogs that did not move toward OD at the end of task A received a significantly higher activity–impulsivity score (3.5 6 0.1 vs 3.0 6 0.8; P , 0.01). Our results confirmed that these tasks could be applied to the study of selective social attention in the dog. Testing more dogs on both tasks might clarify whether increasing difficulty will give us a more sensitive tool that can be applied in different conditions.