Strategies Used by Pet Dogs for Solving Olfaction-Based Problems at Various Distances

Abstract

The olfactory acuity of domestic dogs has been well established through numerous studies on trained canines, however whether untrained dogs spontaneously utilize this ability for problem solving is less clear. In the present paper we report two studies that examine what strategies family dogs use in two types of olfaction-based problems as well as their success at various distances. In Study 1, thirty dogs were tasked with distinguishing a target, either their covered owner (Exp 1) or baited food (Exp 2), from three visually identical choices at distances of 0m (touching distance), 1m, and 3m. There were nine consecutive trials for each target. We found that in Exp 1 the dogs successfully chose their owners over strangers at 0m and 1m, but not at 3m, where they used a win-stay strategy instead. In Exp 2 the dogs were only successful in choosing the baited pot at 0m. They used the win-stay strategy at 1m, but chose randomly at 3m. In Study 2, a different group of dogs was tested with their owners (Exp 1) and baited food (Exp 2) at just the 3m distance with two possible targets in 10-10 trials. In Exp 1 the dogs' overall performance was at chance level; however, when analyzed by trial, we noticed that despite tending to find their owners on the first trial, they generally switched to a win-stay strategy in subsequent trials, only to return to correctly choosing their owners based on olfaction in the later trials. In Exp 2, the dogs chose randomly throughout. We also found that dogs who relied on visual information in the warm-up trials were less successful in the olfaction-based test. Our results suggest that despite their ability to successfully collect information through olfaction, family dogs often prioritize other strategies to solve basic choice tasks.

Fig 1. Testing Room.

Setup of the testing room with the three conditions (0m, 1m, 3m distances), and three positions (A, B, C). The experimenter (E) and the dog entered the room through the side door during the warm up trials, and through the main door (“Enter”) during the test trials. E and dog stood in the waiting area for 15s at the start of each trial.

Fig 2. Finding Owner at Three Distances.

Median value of choices in each condition that were either successful (correct), or identical to the owner’s position during the previous trial (win-stay). Chance level for success (1 correct out of 3) is 33.3%; *** P<0.001, * P<0.05. Chance level for win-stay is 0.899; ## P<0.01.

Fig 3. Finding Food at Three Distances.

Median value of choices in each condition that were either successful (correct), or identical to the position of the baited food during the previous trial (win-stay). Chance level for success (1 correct out of 3) is 33.3%; *** P<0.001. Chance level for win-stay is 0.899; # P<0.05.

Fig 4. Learning Across Owner Trials.

Average number of correct choices in the two halves of the owner trials. **P = 0.01; ## P<0.01. Dashed line represents chance level (2.5 out of 5).

Fig 5. Choices by Trial for Owner Average accuracy and use of win-stay strategy across various owner trials.

Dashed line represents chance level for correct choices. Note that in the 2^nd trial the target was never positioned to the same place as it was in the 1^st. From the 3^rd trial onward a win-stay choice (choosing the position the owner was located in the previous trial) could be either correct or incorrect.