Do Dogs Learn More Quickly from Verbal or Visual Signals?

Source: Creative Commons License CC0

One of the continuing debates among dog trainers and the various individuals who work or compete with trained dogs has to do with the best way for humans to communicate with canines. Specifically, one of the most frequently asked questions boils down to whether it is more effective to tell a dog what you want it to do by using a verbally spoken command, or whether it is better to give the command visually using a hand signal. Science has been very slow in resolving this question, although there is a bit of data that suggests that if dogs already know both the voice command and the hand signa,l you will tend to get a more reliable performance by using the visual signal in most instances (click here for more about that).

In a recent study which appeared in the journal Scientific Reports, a team of researchers headed by Ashley Prichard of the Psychology Department at Emory University in Atlanta, Georgia, decided to determine whether verbal or visual signals work best during the actual process of teaching a dog something new. In this case, they set out to look directly at what happens in the dog's brain when it is being taught a new association using different types of signals. What they were looking at is a form of fast neural learning as it shows up in changes in the response of parts of the canine brain.

The technique used in this study was pioneered around four or five years ago by Gregory Berns (also at Emory University). He was the first scientist to apply to dogs a very high tech measurement procedure that has proven to be quite useful in broadening our understanding of what goes on in the human brain. The technique is functional magnetic resonance imaging (fMRI). It is a measurement procedure that looks at the level of activity in specific brain structures by detecting changes in the blood flow and oxygen levels, which can then be measured through changes in the magnetic field around the head. All that you need to do this is a million dollar MRI machine, a battery of high-powered computers for analysis, and a team of skilled scientists and technicians who can interpret the data. You also need a dog that will lie quietly for 9 or 10 minutes at a time in the bore of the machine while the measurements are being taken. This latter requirement is not easy because MRI machines make a lot of noise (such as whirring gear sounds along with loud clanks and bangs) and such noises might be expected to startle a dog and cause him to move. Because of this problem, two or more months of training may be required for each of the dogs to learn to hold still in order to be tested in fMRI studies. However, this is worth the effort since using the fMRI allows the scientists to directly observe what is happening in the dog's brain during the learning process.

In this study, the actual relationships that the dog had to learn were quite simple. Specifically, all that the researchers were teaching the dog was to discriminate which of two different signals was associated with a reward and which was not. In this study, the learning stimuli could involve smell, vision, or hearing. The test stimuli used were two different scents (one of which smelled like banana while the other is a fruity scent), or two different visual stimuli (a plastic pineapple versus a pink flamingo), or two different spoken words (the nonsense words Callooh compared to Frabjous). The setup was straightforward; the dog's owner would present his pet with a pair of stimuli (one at a time). One of these would always be rewarded with a bit of hot dog, and the other would not. In each session, only one pair of training stimuli would be used for training.

For the purpose of this experiment, these scientists focused on three parts of the brain (the caudate nucleus, amygdala, and parietotemporal cortex), all of which are known to be involved with learning and rewards. For example, activity in the caudate nucleus can be triggered just by the anticipation that something pleasant is about to happen. The dog's brain quickly masters the concept that one of each pair of items is rewarded and prepares for this pleasant event with a quick and vigorous neural response. It is this anticipatory response that is the evidence of learning that the researchers are looking for.

The analysis of fMRI data is complicated, but in the end, the results proved to be quite clear. The dogs learned the difference between the scent stimuli most quickly, followed closely by the difference between the visual stimuli. The difference between the rewarded versus the non-rewarded member of the pair of verbally spoken stimuli took longest for the canine brain to grasp. Such data would clearly suggest that if you are trying to teach a dog something new, you would be better off using visual signals rather than talking to him.

Given this pattern of data, one must ask why is it the case that in dog obedience classes and dog training programs many instructors continue to insist mostly on the use of spoken words as commands when teaching dogs? These researchers conclude that "Our results suggest that the human inclination for verbal communication appears to be based on human preferences, rather than the dog’s innate aptitude." They further go on to say that "pet and working dog training programs would likely become more productive, with accelerated learning rates for the dog, if commands were introduced via hand signals or other physical modes of communication."

Copyright SC Psychological Enterprises Ltd. May not be reprinted, reposted without permission