The intelligence of non-bird dinosaurs is one of the most-asked questions about their biology...
Caption: a recent claim posits that T. rex and dinosaurs like it were as intelligent and cerebrally complex as anthropoid primates, and that tool use was a plausible aspect of their behaviour. In this reconstruction, two tyrannosaur individual co-operate, all while using a tree trunk as a tool, in capturing and subduing an ankylosaur. Credit: Andrew S. Minniear.
On the one hand, this interest stems from the well-known fact that most non-bird dinosaurs had proportionally small brains, certainly so relative to mammals like us. On the other, it reflects interest in what extinct animals were like when alive, and how complex, or otherwise, their behaviour might or could have been. The issue of how we determine ‘intelligence’ in any animal (living or extinct) is complex, and what to do when we only have empty endocranial cavities and rare brain endocasts to go on anyway? As a proxy for intelligence, we use approximate brain size – the primary measure being a brain to body size ratio termed encephalisation quotient or EQ (Jerison 1973) – and also what we know about brain structure and complexity.
Caption: a simplified cladogram depicting the relationships of the relevant reptile groups, with diagrammatic representations of their brains at far right. The animals of special interest to our study are shown in the shaded box. On the brains, the pallium (homologous to the mammal cerebral cortex) is shown in green. Note that theropods like tyrannosaurids are not tremendously different in brain form from crocodylomorphs and lepidosaurs. Image: Caspar et al. (2024).
The general consensus on non-bird dinosaurs is that they were most likely on par in cognitive terms with turtles, lizards and crocodylians, though bird-like maniraptorans were likely more similar to birds like emus and ostriches. However, it’s widely recognised that EQ and other measures are only extremely rough guides to intelligence (e.g., Paulina-Carabajal et al. 2023).
On that note, an interesting caveat is the discovery that animal groups differ profoundly in the number of neurons they possess within a given volume of nervous tissue. In a mammal, turtle, and passerine bird of similar brain mass, for example, the total number of neurons contained within the brain varies from 58.8 million in the mammal, to 14.3 million in the turtle, to 164 million in the bird (Olkowicz et al. 2016, Kverková et al. 2022). Within birds, it’s worth noting that members of the clade Telluraves (birds of prey, rollers, songbirds and so on) have a much greater number of neurons than do other bird groups.
Caption: a montage of telluravian birds. This is the clade that includes rollers, hornbills, woodpeckers and kin, all predatory birds including owls, as well as parrots and songbirds. All individuals shown here were photographed in captivity except for the Song thrush Turdus philomelos at lower right. Images: Darren Naish.
Dinosaurs with primate-like neuron counts. What if an increased neuronal density was true of non-bird dinosaurs too? Could they possess unusually ‘neuron-dense’ brains and thus, potentially, possess much greater intelligence than conventionally argued? We were interested specifically in the number of neurons within the telencephalon, this being that part of the brain that contains the olfactory bulbs and tracts as well as the pallium, the region dealing with higher cognitive functions. It’s equivalent to the cerebral cortex in mammals.
Exactly this question was asked and examined by Brazilian neuroscientist Suzana Herculano-Houzel (2022). After working out brain size and volume from fossil endocranial chambers for 29 dinosaur and pterosaur taxa, Herculano-Houzel (2022) used these data to estimate telencephalic neuronal numbers based on established scaling rules. It’s important to note that her conclusions on dinosaurs have not emerged from out of the blue, but are built on previous studies that looked at brain size and neuron count in fossil vertebrates, and on neuron scaling relationships observed across animal groups (e.g., Herculano-Houzel 2017, 2019).
Caption: this nice piece of art, used to present one of Herculano-Houzel’s (2022) primary contentions about dinosaur intelligence, accompanies a YouTube video made specially to promote the research. You can watch it here.
The results: non-bird dinosaurs had exceptionally high telencephalic neuron counts. Big theropods like Acrocanthosaurus and Tyrannosaurus possessed over 2 billion telencephalic neurons, and thus exceeded corvids and were at the anthropoid primate-level (Herculano-Houzel 2022). She went on to suggest that these dinosaurs “had the biological capability to use and craft tools, and develop a culture, like modern birds and primates” (Herculano-Houzel 2022, p. 11), and also – based on previous work correlating neuron density with physiology and longevity – that (to use just one animal she examined as an example) Tyrannosaurus was endothermic, reached sexual maturity at age 5 and lived to between 42 and 49.
The decision to respond. Herculano-Houzel’s study received a large amount of coverage in the global media, and it could be argued that her statements have the potential to influence popular and scientific thought. A group of us – a team interested in the biology of extinct animals, in brain anatomy, and in inferring behaviours like tool use from the fossil record – decided that an academic response to these claims was warranted. It should be noted that – unbeknownst to us – neurobiologist Anton Reiner also decided to publish a response, and it appeared prior to the publication of our own study (Reiner 2023) and definitely complements it in scope and conclusions. UPDATE: it turns out that Reiner’s paper started life well before the publication of Herculano-Houzel (2022) and was not intended to be a response to it. See the comments below for more information.
Caption: Herculano-Houzel’s 2022 article got quite a substantial amount of media coverage, most of it treating her suggestions as fair and scientifically reasonable. Those shown here are from The Daily Mail (left) and The Washington Post.
And thus today sees the publication of a multi-authored response in The Anatomical Record. Titled ‘How smart was T. rex? Testing claims of exceptional cognition in dinosaurs and the application of neuron count estimates in palaeontological research’ (Caspar et al. 2024), the paper was led by Kai Caspar of Heinrich Heine University, Düsseldorf, Cristian Gutierrez-Ibanez of the University of Alberta, Edmonton, and Grant Hurlburt of the Royal Ontario Museum in Toronto. Our authorship includes tyrannosaurid specialists (Thomas R. Holtz and Thomas Carr) and animal tool-use experts (Jennifer Colbourne).
Dinosaur brain size, revisited. One of Herculano-Houzel’s (2022) primary assumptions was that brain size in non-bird theropods was similar, in proportional terms, to that of birds. Because non-bird theropods include animals only distantly related to birds as well as those extremely similar to archaic birds, it follows that (non-bird) theropod anatomy includes innumerable forms of ‘intermediacy’ with respect to a ‘reptilian’ condition versus an ‘avian’ one. In other words… yes, some non-bird theropods were bird-like in brain size. But those less close to birds were not.
Caption: how do non-bird dinosaurs fair in brain-to-body-size ratios, relative to other reptiles? This graph – plotted using brain sizes and body sizes that we consider plausible (Caspar et al. 2024) – shows that non-bird dinosaurs are mostly on the same approximate regression line as living non-bird reptiles ('“Non-avian Sauropsida”), though some are below it. T. rex (black circle) is above it, but still below the avian regression line (“Non-telluravian birds”). Some bird-like non-birds, however, are close to or on the avian regression line. Image: Caspar et al. (2024).
We analysed relative brain size in dinosaurs using established methods that determine how well data points match phylogeny (Garland & Ives 2000), and one of our key conclusions was that non-bird theropods have a predicted brain size about typical for non-bird reptiles, and below that typical for modern birds. Other dinosaur groups (sauropodomorphs and certain ornithischians) have smaller brains than theropods, and – if we go with lower-end estimates – their brains were proportionally smaller than those of living reptiles (Caspar et al. 2024). Hadrosaurs, by the way, overlap similar-sized theropods in brain size.
Caption: when incorporating dinosaur brain volume data, we mostly used data previously compiled and verified by members of our team, Grant Hurlburt especially. For a few specimens, uncertainty over volume meant that we calculated new ones using graphic double integration (GDI). This diagram depicts how GDI was calculated (the example here being the endocast of the T. rex specimen AMNH 5029). Image: Caspar et al. (2024).
Put more succinctly: we found that Herculano-Houzel (2022) substantially over-estimated brain size in non-bird dinosaurs, and the more conservative values found by previous authors (Hurlburt et al. 2013, Morhardt 2016) are more likely to be correct. This is primarily because she assumed that the brain occupied essentially the whole of the endocranial space, yet this is not true for the majority of dinosaur groups (Jerison 1973, Hurlburt et al. 2013, Caspar et al. 2024).
High neuron counts, revisited. What about those estimates of exceptionally higher neuron densities? Herculano-Houzel (2022) assumed that non-bird theropod brains should be similar to those of extant birds in telencephalic neuron density. She then scaled up the neuron counts from the brains of birds to conform to the much larger (in literal terms) brains of extinct dinosaurs, the results being extremely high telencephalic neuronal counts, sometimes in the low billions.
Two things emerge here. Firstly, we’ve seen that Herculano-Houzel (2022) miscalculated brain size. Wouldn’t such a miscalculation result in an artificially inflated neuron count? We re-analysed neuron count relative to estimated brain size, using scaling relationships derived from birds, non-bird reptiles, and mammals… and got very different results. For T. rex (which serves as an exemplar here because its brain anatomy is relatively well known), ‘reptilian scaling’ results in 360 million telencephalic neurons under some volume estimates, and 1.7 billion using ‘avian scaling’ under some volume estimates (Caspar et al. 2024). The estimates based on ‘reptilian scaling’ – which are the ones we should be relying on more than those based on ‘avian scaling’ – don’t put T. rex in an exceptional category at or above the corvid range.
Caption: this scatter plot shows predicted neuron count (in the cognitive part of the brain) as estimated by Herculano-Houzel (2022) (in red) and us (in green). Look at the values predicted for the T. rex specimen AMNH 5029 in particular. Reconstructed with a bird-like neuron count (green circles), our estimates do put T. rex at over 1 billion neurons. But when reconstructed with a non-bird-like neuron count (green squares), we find T. rex to be below 1 billion. Morhardt (2016) estimated neuron counts for the telencephalon alone and her predicted values, shown at lower right, also put T. rex at less than one billion neurons when ‘reptilian scaling’ was used. Image: Caspar et al. (2024).
Secondly, are we sure that high telencephalic neuron counts – on par with those of anthropoid primates – are indicative of behavioural complexity, high intelligence and so on anyway? At this point, I think that many of us have heard about the claimed link between neuron count and intelligence.
What we argue in the paper is that data from across tetrapod diversity reveals that any such proposed link is on shaky ground. Especially high cerebral neuron counts – exceeding those of corvids and most primates – are present in giraffes (1.7 billion) but there’s no indication so far that this is correlated with tool use, complex culture, or anything else linked with intelligence. Dolphins including pilot whales and orcas have absurdly high neocortical neuron counts (exceeding 30 billion) – we humans have 15-20 billion – but there is, as yet, no evidence that they exceed us in whatever intelligence metric you wish to devise (Caspar et al. 2024). You might be wondering at this point if certain animals have high numbers of neurons due to the fact that… well, that they’re big. It has, indeed, generally been thought that this is the case. However, Herculano-Houzel (2011) has pushed back against this, her argument being that it’s absolute brain size and absolute neuron count that’s important. Under this logic, giraffes are secretly super-smart, I guess we just have yet to document it.
Caption: certain living animals have extremely high numbers of neurons in those parts of the brain associated with cognition. In cetaceans, like the orca shown here (I’ll avoid the species-level taxonomy for now…), we can say that this high neuron density is indeed correlated with cultural and behavioural complexity. But what about animals like giraffes? Turns out that they have a really high cerebral neuron count as well. Images: Darren Naish.
Being ‘reptile smart’… even ‘fish smart’… is not a bad thing. In the end, Herculano-Houzel’s (2022) suggestions and speculations, while fascinating, just don’t match the data. The idea that dinosaurs like T. rex might have been capable of cultural transmission or tool manufacture and use – that they could have been ‘monkey smart’, on par with macaques and baboons – should not be considered likely (Caspar et al. 2024).
The evidence from dinosaur brain size shows that non-bird dinosaurs were ‘reptile smart’ – on par with lizards and crocodylians – and not in the same league as big-brained birds or mammals, and not at the primate level. In fact, the especially small brains of sauropodomorphs and certain ornithischians – they’re smaller-brained than living lizards or crocs (Caspar et al. 2024) – might imply that they were cognitively on par with amphibians and fishes. Finally, there’s no reason to think that non-bird dinosaurs of any sort had exceptionally high telencephalic neuron counts exceeding 2 or 3 billion… and even if they did it’s not clear that this would demonstrate behavioural complexity or unusual intelligence (Caspar et al. 2024).
Caption: projects like Prehistoric Planet – hadrosaur-themed images from seasons 1 and 2 of that series are shown here – depict non-bird dinosaurs living socially, indulging in parental care, and interacting with other members of their social group. If you think that animals need to be similar in intelligence to telluravian birds or primates to exhibit such behaviours… you’re wrong. Images: Prehistoric Planet / Apple TV.
Speaking as someone who’s been involved in several efforts to reconstruct and portray the intelligence of non-bird dinosaurs – most recently for the Apple TV series Prehistoric Planet – I want to point out that comparing dinosaurs to animals like turtles, lizards or crocodylians, or even amphibians or (shudder) fishes, must never be considered a ‘bad thing’. As far as I see it, it makes little to no difference in terms of how we depict or imagine them: such aspects of behaviour as foraging, drinking, nest building, post-hatching parental care, sexual display, combat, herding behaviour, and the pursuing, subduing, killing and dismembering of prey all play out in the same way no matter what ‘intelligence’ you imagine these animals to be equipped with. I don’t watch amphibians or fish engaging in sexual display, intraspecific combat, parental care or predation and find myself thinking how disappointing it is that they don’t exhibit telluravian- or primate-like intelligence.
Granted, ‘reptile smart’ dinosaurs are unlikely to mourn or pine, exhibit a theory of mind, teach their youngsters, or pay attention to celestial events. But I think we can agree that suggestions that extinct dinosaurs might have behaved in such ways are more to do with science fiction than what we know about the behaviour of real animals.
Caption: your reminder that crocodiles, alligators, various lizards, turtles and so on can all be highly trainable, and exhibit many traits that we associate with ‘intelligence’. This image shows a trained Saltwater crocodile Crocodylus porosus at Australia Zoo. Image: Australia Zoo, CC BY-SA 2.0 (original here).
T. rex and similar animals were not on par with avian or mammalian tool-users, they were not big-brained, and they didn’t possess billions of neurons within the telencephalon of the brain. Does this make them any less interesting, or any less sophisticated or complex, than we’ve conventionally imagined them? Definitely not.
For previous Tetrapod Zoology articles on dinosaur brains, biology and connected issues, see…
Could We Domesticate (Non-Bird) Dinosaurs?, August 2018
Postcranial Palaeoneurology and the Lifestyles of Pterosaurs, August 2018
Alternative Timeline Dinosaurs, the View From 2019 (Part 3): the Dinosauroid and its Chums, December 2019
Humanoid Dinosaurs Revisited Again: Russell and Séguin’s Dinosauroid at (Nearly) 40 Years Old, August 2021
A brain for Baryonyx: using CT-scanning to examine British spinosaurid brains, February 2023
Refs - -
Garland, T., & Ives, A. R. 2000. Using the past to predict the present: confidence intervals for regression equations in phylogenetic comparative methods. The American Naturalist 155, 346-364.
Herculano-Houzel, S. 2011. Brains matter, bodies maybe not: The case for examining neuron numbers irrespective of body size. Annals of the New York Academy of Sciences 1225, 191-199.
Herculano-Houzel, S. 2017. Numbers of neurons as biological correlates of cognitive capability. Current Opinion in Behavioral Sciences 16, 1-7.
Herculano-Houzel, S. 2019. Longevity and sexual maturity vary across species with number of cortical neurons, and humans are no exception. Journal of Comparative Neurology 527, 1689-1705.
Herculano-Houzel, S. 2022. Theropod dinosaurs had primate-like numbers of telencephalic neurons. Journal of Comparative Neurology 531 962-974.
Hurlburt, G. R., Ridgely, R. C. & Witmer, L. M. 2013. Relative size of brain and cerebrum in tyrannosaurid dinosaurs: an analysis using brain-endocast quantitative relationships in extant alligators. In Parrish, J. M., Molnar, R. E., Currie, P. J. & Koppelhus, E. B. (eds), Tyrannosaurid Paleobiology. Indiana University Press, Bloomington, pp. 134-154.
Jerison, H. J. 1973. Evolution of the Brain and Intelligence. Academic Press, New York.
Kverková, K., Marhounová, L., Polonyiová, A., Kocourek, M., Zhang, Y., Olkowicz, S., Straková, B., Pavelková, Z., Vodička, R., Frynta, D. & Němec, P. 2022. The evolution of brain neuron numbers in amniotes. Proceedings of the National Academy of Sciences 119, e2121624119.
Morhardt, A. C. 2016. Gross Anatomical Brain Region Approximation (GABRA): Assessing Brain Size, Structure, and Evolution in Extinct Archosaurs. PhD thesis, Ohio University.
Paulina-Carabajal, A., Bronzati, M. & Cruzado-Caballero, P. 2023. Paleoneurology of non-avian dinosaurs: an overview. In Dozo, M. T., Paulina-Carabajal, A., Macrini, T. E. & Walsh, S. (eds) Paleoneurology of Amniotes: New Directions in the Study of Fossil Endocasts. Springer International Publishing, Cham, pp. 267-332.
Reiner, A. 2023. Could theropod dinosaurs have evolved to a human level of intelligence? The Journal of Comparative Neurology doi 10.1002/cne.25458.
