Why do animals with more mass tend to have brains with so much more mass when it seems like a similar mass brain should be able to do the job?

Question

Why does a wolf have a brain so much bigger than that of a poodle, when a poodle's brain is big enough to do the job, I would have thought. Likewise, if tigers are not smarter than house cats, why do they have brains that are so much bigger? Brain tissue is metabolically expensive, and so you'd think there would be a strong selection pressure to have the smallest brain that could do the job.

I'm familiar with the idea that a large body needs a large brain to control it, but I've never understood why that should be the case. And since poodles are modified wolves, it should be easy for wolves to quickly evolve poodle brains in nature. Has anyone tried to create a large breed of dog with a small brain, to see whether it's possible, and what problems if any the animal has in controlling it's body, for example?

So my question is: Why do animals with more mass tend to have brains with so much more mass when it seems like a similar mass brain should be able to do the job?

Answer 1

I think that you can find a good answer to your question in a more careful evaluation of the hypothetical that you pose:

since poodles are modified wolves, it should be easy for wolves to quickly evolve poodle brains in nature.

How easy is it, actually, for a wolf to evolve into a poodle vs. evolving only the wolf's brain into a poodle's brain?

The answer can be found in the science of evolutionary developmental biology (evo-devo), a good popular introduction to which can be found in the book Endless Forms Most Beautiful.

In general, evolution of scale and morphology operates by controlling the differentiation of tissues in the early development of a body-plan via Hox genes and other related regulatory elements. When understanding their effects on an organism as a whole, in many cases, each such gene can be thought of as a variable that shifts or distorts a body coordinate system. The more fundamental such a gene is to the body plan, the more widespread its effects will be and with a greater degree of side effects. This means that sometimes it's actually much easier to select for a large complex of linked changes than for a single specific change.

In fact, it turns out that for many animals the overall scale is one of the easiest variables to control! Consider, after all, how much you yourself changed in scale during your own growth from birth to adulthood, while maintaining viability and system integration the whole time. Many animals can thus readily become bigger by running the growth program more or smaller by running it less, and these scale changes are often also associated with neoteny. It is, in fact, pretty easy to turn a wolf into a chihuahua, and a chihuahua is pretty close to a wolf pup.

To change only the brain of the mature animal, however, is actually harder, requiring a more specific selection signal. Brains appear to have gotten bigger with respect to body under selection pressure that led to novel architectural features giving specific highly-competitive capabilities. Selection pressure for a small brain, however, would likely be about metabolic burden instead, and there are a lot of potential strategies for reducing metabolic burden, at the same time that many of the variables affecting brain size are likely to be entangled with other critical features as well, such as the jaw.

Bottom line: it's evolutionarily a lot harder to shrink just the brain than to shrink the whole animal.

Answer 2

Well, it's worth noting that the number of neurons doesn't scale linearly with the brain size; in particular for two dogs analyzed by Herculano-Houzel et al. (2017), the number of neurons in the cortex scaled the least:

Both dog individuals examined (a 7.45 kg mixed-breed and a 32 kg golden retriever) had larger brains than the cat (brain mass in dogs, 58.4 and 114.5 g, respectively; cat, 34.8 g), and also more brain neurons than the cat (dogs, 1.8 and 2.6 billion neurons, respectively; cat, 1.2 billion neurons). The same applies to the cerebral cortex of the dogs, at 46.2 g with 429 million neurons and 84.8 g with 623 million neurons, against 24.2 g with 250 million neurons in the cat. Strikingly, although the cerebral cortex of the golden retriever was almost twice as large as the cortex of the smaller dog, it only had 46% more neurons than the smaller dog cortex (as expected from the non-linear scaling of cortical mass with number of cortical neurons [of carnivora in general]);

And the authors of this paper also write that this unsurprising because cortex neurons are apparently the most expensive to scale up in numbers:

We have previously shown that the metabolic cost of the brain is proportional to its number of neurons, regardless of brain size, and that neurons in the cerebral cortex cost on average 10 times as much energy as neurons in the cerebellum (Herculano-Houzel, 2011). Thus, cerebral cortical neurons are expected to be both more vulnerable to caloric shortage than other brain neurons, and to also contribute more to decreasing total metabolic cost when their numbers are reduced than the loss of other neuronal populations would.

And seemingly being a bigger carnivore with somewhat more cortex neurons might actually make one a little bit smarter. The paper notes in support of that that:

Cognitive performance in carnivorans was recently addressed specifically by Benson-Amram et al. (2015). Across these species, even though brain size relative to body mass is a significant predictor of success in opening a puzzle box, species with larger absolute brain volumes also tended to be better than others at opening the puzzle box (Benson-Amram et al., 2015).

(This paper might have been somewhat sensitive to the method used for testing, like the scaling of the box. It actually found that [unsurprisingly to me] bears were champions at opening boxes... but Herculano-Houzel's paper actually found that bears don't have that many cortex neurons... "the brown bear has [...] 251 million [cortex] neurons, which is only about as many as the house cat, even though the brown bear cortex had a nearly 10-fold larger mass". So something doesn't quite click here.)

Anyway, your hypothesis that a cat and a lion (or actually a bear) have the exact same cognitive abilities might not be correct, when generalized. (Lions have as many cortex neurons as dogs, on average, 500 million, according to Herculano-Houzel.)

There is also a 2019 paper by Horschler et al. on dogs specifically that finds something similar based on more data on dogs:

Using citizen science data on more than 7000 purebred dogs from 74 breeds, and controlling for genetic relatedness between breeds, we identify strong relationships between estimated absolute brain weight and breed differences in cognition. Specifically, larger-brained breeds performed significantly better on measures of short-term memory and self-control. However, the relationships between estimated brain weight and other cognitive measures varied widely, supporting domain-specific accounts of cognitive evolution.

I haven't looked at their methods section yet.