Different paths, same functions: the convergence of neural circuits
The inevitable emergence of good solutions?
If you find Complexity Thoughts interesting, follow us! Click on the Like button, leave a comment, repost on Substack or share this post. It is the only feedback I can have for this free service. The frequency and quality of this newsletter relies on social interactions. Thank you!
→ Don’t miss the podcast version of this post: click on “Spotify/Apple Podcast” above!
Imagine we’re trying to build a brain from scratch: would we end up with the six‑layered neocortex of mammals, the nuclear pallium of birds or something entirely different?
Surprisingly, evolution seems less interested in reusing the same “blueprint” than in rediscovering the same tricks. Cognition‑supporting circuits across birds, mammals and reptiles arise from distinct embryonic roads and, yet, they converge on remarkably similar computational designs, such as dense connectivity, recurrent loops, neuromodulatory control and cooperative subsystems. In other words: different brains, same functions.
In this post we are going to quickly overview the new results and then discuss them — within a broader a picture — with Prof. Ricard Solé.
Convergent evolution
When two lineages tackle similar challenges — e.g., navigating complex landscapes, solving novel problems, social coordination — they may independently evolve analogous solutions. Think about wings on bats and birds, streamlined bodies in sharks and dolphins, echolocation mechanisms in bats and whales.
Three perspectives, one question
But what happens when those “wings” are neural circuits? Do similar cognitive capacities demand the same structures or can very different architectures produce equivalent functions?
In the April 2025 issue of Science, three groups tackle this question by comparing the telencephalic pallium of chickens, geckos, turtles, mice and humans. The challenge: are mammalian neocortex and avian pallium homologous, or are they dazzling examples of functional convergence? Together, they suggest a picture of “constrained roads to complex brains” where similar functions emerge from different developmental and regulatory starting points (for further details, see this paper):
Despite vast diversity in behavior and cognition across the tree of life, there is surprising similarity in brain structures and even gene expression across amniotes, the largest group of vertebrates that includes reptiles, birds, and mammals.
Rueda-Alana et al. integrated transcriptional analysis at single-cell resolution and mathematical modeling to investigate the development of sensory circuits in chicken, gecko, and mouse. → Paper
Zaremba et al. generated a spatially resolved cell atlas in chicken pallium and compared it with mouse, two lizards, and turtle to interrogate the conservation of cell states. → Paper
Hecker et al. developed deep learning models to identify shared and divergent regulatory signatures across telencephalon cell types in chicken, human, and mouse. → Paper
Evolutionary convergence of sensory circuits in the pallium of amniotes
Do sensory‑processing neurons in birds and mammals share a common developmental timetable?
“For decades, scientists have debated the homologies between the mammalian neocortex and the pallium of other vertebrates”
The amniote pallium contains sensory circuits that are structurally and functionally equivalent, yet their evolutionary relationship remains unresolved. We used birthdating analysis, single-cell RNA and spatial transcriptomics, and mathematical modeling to compare the development and evolution of known pallial circuits across birds (chick), lizards (gecko), and mammals (mouse). We reveal that neurons within these circuits’ stations are generated at varying developmental times and brain regions across species and found an early developmental divergence in the transcriptomic progression of glutamatergic neurons. Our research highlights developmental distinctions and functional similarities in the sensory circuit between birds and mammals, suggesting the convergence of high-order sensory processing across amniote lineages.
“Our study demonstrates that high-order sensory processing circuits have evolved separately in different vertebrate taxa, converging into a functionally similar circuit […] Evolution tinkered with pallial circuit development, structure, and function. And likely, convergent evolution sculpted the formation of the components of the sensory circuits in amniote species.
Developmental origins and evolution of pallial cell types and structures in birds
How similar are the cell types of the chicken pallium to those of mammals and reptiles?
“Overall, our study resolves long-standing debates on the amniote pallium, offering valuable insights into the evolutionary trajectory and diversification of neural cell types and structures crucial for advanced behaviors”
Innovations in the pallium likely facilitated the evolution of advanced cognitive abilities in birds. We therefore scrutinized its cellular composition and evolution using cell type atlases from chicken, mouse, and nonavian reptiles. We found that the avian pallium shares most inhibitory neuron types with other amniotes. Whereas excitatory neuron types in amniote hippocampal regions show evolutionary conservation, those in other pallial regions have diverged. Neurons in the avian mesopallium display gene expression profiles akin to the mammalian claustrum and deep cortical layers, while certain nidopallial cell types resemble neurons in the piriform cortex. Lastly, we observed substantial gene expression convergence between the dorsally located hyperpallium and ventrally located nidopallium during late development, suggesting that topological location does not always dictate gene expression programs determining functional properties in the adult avian pallium.
Enhancer-driven cell type comparison reveals similarities between the mammalian and bird pallium
Do the DNA “switches” (enhancers) that control neuron‑type genes betray a shared regulatory code?
“We show that chicken enhancer sequences exhibit activity in the corresponding mammalian telencephalic cell types when assayed in mouse brains”
Combinations of transcription factors govern the identity of cell types, which is reflected by genomic enhancer codes. We used deep learning to characterize these enhancer codes and devised three metrics to compare cell types in the telencephalon across amniotes. To this end, we generated single-cell multiome and spatially resolved transcriptomics data of the chicken telencephalon. Enhancer codes of orthologous nonneuronal and γ-aminobutyric acid–mediated (GABAergic) cell types show a high degree of similarity across amniotes, whereas excitatory neurons of the mammalian neocortex and avian pallium exhibit varying degrees of similarity. Enhancer codes of avian mesopallial neurons are most similar to those of mammalian deep-layer neurons. With this study, we present generally applicable deep learning approaches to characterize and compare cell types on the basis of genomic regulatory sequences.
“[…] enhancer codes reveal both expected and unexpected correspondences between cell types in the mammalian and avian telencephalon, indicating conserved regulatory programs that likely originated in the common amniote ancestor and have been co-opted or diversified”
Good “solutions” are inevitable?
I discussed about the above papers, and more broadly about the convergent evolution to biological innovation enabling clear advantages in nature, with Ricard Solé.
As part of his research on the origins and evolution of biological complexity, Ricard has been working on cognitive systems, using theoretical models and experiments with ant colonies and synthetic biology. Some years ago, he proposed the idea of “liquid brains” as a unifying concept to deal with those cognitive networks (from social insect colonies and the immune system to swarms) that depart from the standard “solid” brains where neurons have fixed spatial locations in space. He has also advocated exploiting concepts from natural collective intelligence to design synthetic circuits using genetic engineering.
Ricard, given the evidence for convergent development of sensory circuits in birds and mammals, to what extent do you think similar computational “solutions” might emerge in non‑amniote lineages (e.g., cephalopods or even plant signaling networks), and what does this imply about the universality of certain neural architectures?
It is remarkable that, in octopi, we find a reinvention of brain structures that strongly converge with the vertebrate brain despite their completely independent origins. This supports the idea that strong constraints exist on the potential cognitive architectures that can evolve. Concerning plant signalling, I am sceptical. Their nature (highly modular and redundant) and structural architecture, along with the lack of true neural-like elements, make them an entirely different class of agents.
Could parallel studies in “liquid brains” of ant colonies or in decentralized cephalopod nervous systems uncover analogous regulatory convergences, suggesting a shared logic of intelligence beyond neuron‑based architectures?
This is an open question for liquid brains, since the intrinsic nature of the individual interactions (which is not stable over time) profoundly limits their cognitive complexity. Concerning cephalopods, one interesting point is that their behavioural repertoire (memory, learning, association) is shared with that displayed by vertebrates. Convergent designs point to similar solutions to the universal behavioural repertoire.
If highly sophisticated cognitive circuits can evolve independently from distinct embryological substrates, how should this reshape our perspective on the human neocortex’s supposed uniqueness and the anthropocentric view of intelligence as a linear progression?
Along with those constraints imposed by developmental pathways and computational constraints, we must remember the social component that seems equally prominent in humans. Our brains develop over a long period after birth, being exposed to the constant influence of other individuals, learning language and different skills, and actively seeking information. That is (along with other unique properties, such as mental time travel) a marked trait not shared by other species. I doubt that the linear progression concept is correct.
Considering the recurrent emergence of similar network motifs (such as dense recurrent loops, neuromodulatory hubs, cooperative subsystems) across phyla, do you believe there exist “inevitable innovations” in information processing that any evolving system must stumble upon, and if so, how might we identify them experimentally?
I do. Evolution has been experimenting with the space of the possible over millions of years, and it is suspicious that some key components of the connectome (across scales) are found everywhere. Part of it results from optimality, whereas some regularities might result from the conflict between segregation (division of labour) and integration (via sync). This is particularly well explained by early work by Sporns and colleagues on the evolution of in silico networks when complexity is used as the fitness function.
Which core elements (eg, network motifs, molecular rules, or developmental constraints) do you believe constitute the fundamental logic driving this functional convergence of cognitive circuits across such diverse lineages?
Yes, but this needs to be tested using the right combination of experiments (using animal models), theory and molecular phylogenetics. It is a good question that can be accurately tested.
I thank Ricard for taking time to discuss with me about this fascinating topic. I hope that someone, somewhere in the world, is thinking about experimental tests.
Ultimately, the amniote pallium reveals a mosaic of deep homologies — conserved inhibitory neuron types and regulatory codes — and parallel reinventions of excitatory circuits, reminding us that brain evolution is both inheritance and innovation, as well as that only by integrating developmental timing, circuit mapping and functional assays across taxa can we uncover the universal logic of intelligence.
After resetting the universe, would life evolution tell a different story?
If you find Complexity Thoughts interesting, follow us! Click on the Like button, leave a comment, repost on Substack or share this post. It is the only feedback I can have for this free service. The frequency and quality of this newsletter relies on social interactions. Thank you!