The effect of domestication on inhibitory control: wolves and dogs compared

Abstract

Inhibitory control i.e. blocking an impulsive or prepotent response in favour of a more appropriate alternative, has been suggested to play an important role in cooperative behaviour. Interestingly, while dogs and wolves show a similar social organization, they differ in their intraspecific cooperation tendencies in that wolves rely more heavily on group coordination in regard to hunting and pup-rearing compared to dogs. Hence, based on the 'canine cooperation' hypothesis wolves should show better inhibitory control than dogs. On the other hand, through the domestication process, dogs may have been selected for cooperative tendencies towards humans and/or a less reactive temperament, which may in turn have affected their inhibitory control abilities. Hence, based on the latter hypothesis, we would expect dogs to show a higher performance in tasks requiring inhibitory control. To test the predictive value of these alternative hypotheses, in the current study two tasks; the 'cylinder task' and the 'detour task', which are designed to assess inhibitory control, were used to evaluate the performance of identically raised pack dogs and wolves. Results from the cylinder task showed a significantly poorer performance in wolves than identically-raised pack dogs (and showed that pack-dogs performed similarly to pet dogs with different training experiences), however contrary results emerged in the detour task, with wolves showing a shorter latency to success and less perseverative behaviour at the fence. Results are discussed in relation to previous studies using these paradigms and in terms of the validity of these two methods in assessing inhibitory control.

Fig 1. Wolves’ and pack-dogs’ performance in the detour task.

Mean duration (in seconds) and SE of time spent close to the fence in front of the food by wolves and dogs in the detour task. *p<0.01.

Fig 2. The opaque cylinder used in training trials.

The exact same apparatus, but transparent, was used in test trials.

Fig 3. Wolves’ and pack-dogs’ performance in the cylinder task.

Mean number (and 95% confidence interval) of trials to criterion in the training phase, and of correct trials in the test phase, for wolves and dogs in the cylinder task. *p<0.01.

Fig 4. Pet- and pack- dog’s performance in the cylinder task.

Mean number (and 95% confidence interval) of trials to criterion in the training phase, and of correct trials in the test phase, for trained and untrained pet dogs and dogs tested with no prior familiarization with the opaque cylinder.