Deception
"My Dog Is a Genius Who Can Count"
FMRi's show that the parietotemporal cortex is involved in estimating quantity.
Posted December 21, 2019 Reviewed by Lybi Ma
An essay by science writer Virginia Morell titled "Dog brains have a knack for numbers, much like ours" caught my eye because I'm interested "all things dog" and also because I receive numerous emails containing stories showing that dogs can count or at least estimate quantity fairly accurately. One woman at a local dog park, Cindy, told me she calls her dog Albert, because he's an "Einsteinian genius" with highly evolved math skills." She once told me, "I swear my dog is an 'Einsteinian genius' who can count. He counts treats and always goes to the larger number rather than to the larger pile." Below, you'll see how the research supports her observations.
I'm not sure how far I would go with what Cindy told me, however, in a few demonstrations, Albert, better known as Billy, showed that he had a very strong sense of numbers, or an "approximate number system" (ANS) underlying his ability to rapidly estimate quantity. The ANS, "a system for rapidly assessing the approximate number of items present in an array, appears to be present both across the animal kingdom, and early in human development, with even newborn infants possessing the remarkable ability to discriminate stimuli based on numerosity."
Ms. Morell's essay is an easy-to-read summary of a research paper by Lauren Aulet and five colleagues published in The Royal Society Letters called "Canine sense of quantity: evidence for numerical ratio-dependent activation in parietotemporal cortex." Both pieces are available for free online.
To conduct their research about dogs' visual numerosity, 11 dogs were studied, all of whom had previous experience lying still in fMRI machines. (See "Awake fMRI Reveals How Canine Brains Process Novel Words" for more information about how the dogs used in these sorts of studies do so voluntarily and what's required of them, as well as summaries of other neuroimaging research projects.3) After the dogs were relaxed and motionless in the scanner, they were asked to look at grey dots displayed on a black screen.1
The researchers write, "During scanning, dogs passively viewed dot array stimuli presented on a screen placed in the rear of the scanner. Dogs were presented with alternating stimuli of 2 and 10 (1 : 5 ratio), 4 and 8 (1 : 2 ratio) or 6 and 6 (1 : 1 ratio) dots in a block fMRI design (figure 1)." The array changed every 300 milliseconds.
They discovered that in eight of 11 dogs, block ratio was correlated with increasing activation in parietotemporal lobes, the specific area of which differed among the different breeds. Ms. Morell nicely summarizes the results of this research as follows: "If dogs, like humans and nonhuman primates, have a dedicated brain region for representing quantities, their brains should show more activity there when the number of dots was dissimilar (three small dots versus 10 large ones) than when they were constant (four small dots versus four large dots)."2
Can dogs count and is Billy really an "Einsteinian genius?"
"Taken together, our findings suggest that the ability to represent numerosity and the mechanisms supporting this system are deeply conserved over evolutionary time, perhaps owing to a role in foraging or predation, and persists in a domesticated species."
The data reported in "Canine sense of quantity: evidence for numerical ratio-dependent activation in parietotemporal cortex." show the activity in the parietotemporal regions of the dogs' brains varied as numerical ratio changed, a pattern similar to that previously noted in primates, and eight of the 11 dogs were able to "discriminate visual numerosity" without prior training.
The research also suggests that dogs may be more sensitive to the number of items present rather than their volume, or the space they occupy. This might be useful for dog trainers and recall what Cindy told me about Billy: "I swear, he can count treats and always goes to the larger number rather than to the larger pile." The research I'm discussing supports her "citizen science" observations. Citizen science often mirrors quite a lot about what research has discovered. (See "As Dogs Have Aha! Moments, What's Happening in Their Brains?")
All in all, if you think your dog can count, it's more likely that they're estimating quantity, something they can do fairly well even if it may not be as accurate as humans' ability to do so.
Stand by for forthcoming comparative research on the ways in which noninvasive neuroimaging can help us see how nonhuman animals sense their worlds. It will be interesting to learn not only how similar they are to humans, but also about those brain patterns that are species-specific and reflect adaptations that are unique to a given species or group of species. What an exciting time it is to be interested in the neural bases of cognitive and emotional capacities.
I look forward to future studies and discussions of the sense of number in a wide variety of nonhumans. It's unlikely we're the only animals who have evolved such cognitive abilities and this study adds data that support this suggestion.
References
Notes:
1) Stimuli were 75 dot arrays comprising light grey dots on a black background (800 × 800 pixels). For each numerosity used (2, 4, 6, 8 and 10), stimuli varied in cumulative area (i.e. the total grey on each image). For each numerosity, cumulative area was 10, 20 or 30% of the total stimulus. For each numerosity, 15 unique stimuli were used (five stimuli per cumulative area value). In each stimulus, individual dot size varied up to 30%. Dot location varied randomly. Critically, these controls minimize the influence of non-numerical properties, in order to ensure that the results can be attributed to changes in numerical value [27,28]. In accordance with current estimates of canine visual acuity (approx. 20/75; [29]), all stimuli were analysed to ensure that the inter-dot distances were large enough for dogs to individuate. (Numbers in brackets refer to references in the research essay.)
2) The researchers write, "Crucially, to assess whether dogs, like primates, have number-sensitive regions of cortex, we examined whether activation in the localized ROIs [regions of interest] was ratio-dependent in the held-out data. We found that block ratio was significantly correlated with the level of activation in these regions (t7 = 2.01, p = 0.042, one-sided; figure 2d), consistent with a ratio-dependent effect. These findings suggest that dogs have a visual sense of number subserved by similar parietotemporal mechanisms to those in primates [10,11]." (Numbers in brackets refers to references in the research essay.)
3) Some relevant references about the importance of neuroimaging studies.
Bekoff, Marc. How Dogs View the World: Brain Scans Tell Us What They See.
_____. Awake fMRI Reveals How Canine Brains Process Novel Words.
_____. Jealousy in Dogs: Brain Imaging Shows They're Similar to Us.
_____. What It's Like to Be a Dog. (An interview with Gregory Berns)
_____. Canine Confidential: Why Dogs Do What They Do. Chicago, University of Chicago Press, 2018.
Berns. Gregory. What It's Like to Be a Dog: And Other Adventures in Animal Neuroscience. New York, Basic Boos. 2017.
Gregory Berns: Decoding the dog's mind with awake neuroimaging.
Cook, Peter et al. Jealousy in dogs? Evidence from brain imaging. Animal Sentience, 2018.
