Distributed neural computation and the evolution of the first brains
This article has been Reviewed by the following groups
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
- Evaluated articles (PREreview)
Abstract
The origin of brains in the Precambrian was a landmark in animal evolution, enabling new behavior and life histories. Brains likely evolved from diffuse nerve nets, but we do not know what the first brains looked like or how they were organized. Acoel worms, the likely sister lineage to all other animals with brains, offer a unique window into this transition. Here, we studied the acoel worm Hofstenia miamia , a marine predator that hunts planktonic invertebrates and displays other sophisticated behavior. We found that H. miamia has an unusual ‘diffuse brain’: a subepidermal network of dense neuropil exhibiting little regionalization or stereotypy in gross anatomy or distribution of neural cell types. Remarkably, we found that behavior in H. miamia is robust to large, arbitrary amputations of brain regions, suggesting that most regions can perform most computations. More brain tissue improves performance, especially on challenging tasks, but no specific brain region is required. These results lead us to propose that H. miamia ’s brain is composed of computationally pluripotent “tiles” that interact to generate coherent behavior. This architecture suggests a trajectory for nervous system evolution in which early brains may have arisen through the condensation of diffuse nerve nets into unregionalized brains, with regionalization evolving secondarily.
Article activity feed
-
This Zenodo record is a permanently preserved version of a PREreview. You can view the complete PREreview at https://prereview.org/reviews/18110759.
Across the animal kingdom, organisms must navigate a dynamic sensory world to feed, reproduce, and survive. In many species, centralized nervous systems have evolved to become both stereotyped and specialized, enabling animals to contend with and adapt to the specific challenges of their environments. While diffuse, non-specialized networks of neurons are thought to be precursors to centralized brains, it remains unknown how stereotyped regionalization and functional divisions have evolved to support different behavioral needs. The present study addresses these questions with the acoel worm H. miamia whose nervous system is uniquely positioned between diffuse nets and centralized brains. …
This Zenodo record is a permanently preserved version of a PREreview. You can view the complete PREreview at https://prereview.org/reviews/18110759.
Across the animal kingdom, organisms must navigate a dynamic sensory world to feed, reproduce, and survive. In many species, centralized nervous systems have evolved to become both stereotyped and specialized, enabling animals to contend with and adapt to the specific challenges of their environments. While diffuse, non-specialized networks of neurons are thought to be precursors to centralized brains, it remains unknown how stereotyped regionalization and functional divisions have evolved to support different behavioral needs. The present study addresses these questions with the acoel worm H. miamia whose nervous system is uniquely positioned between diffuse nets and centralized brains. Leveraging a wide array of imaging techniques, single-cell transcriptomics, and behavioral manipulations, the authors present evidence that H. miamia brains lack regionalization and stereotypy, and are instead composed of uniform, repeating units that work in concert to support ethologically relevant foraging behaviors.
This study represents a valuable contribution to the question of how neural circuits evolve for behavior. A key strength is its framing of H. miamia brains as composed of computationally equivalent tiles that can be integrated to enhance foraging—this is an intriguing proposal for how early nervous systems may have coordinated behavior before regional specializations. While anatomical and molecular data provide compelling support for a lack of regionalization, the behavioral manipulations could benefit from more direct demonstrations of the functional uniformity of the tiles and their combined effect on behavior. Some ambiguity also remains regarding how the central network interfaces with sensory and motor regions of the body.
Major Comments
1. The authors propose that the H. miamia brain consists of "computationally pluripotent" tiles that all access the same information (lines 280-283), as foraging remains largely intact after amputations. However, the anatomical data reveal that tiles often include sensory specializations (line 79) or densities associated with specific organs (e.g., mouth cilia or statocyst innervation), suggesting heterogeneity. Consistent with this, the authors find foraging deficits in mouth-amputated animals, which seems at odds with computational uniformity. Higher resolution anatomical analyses could reveal whether tiles near the mouth receive denser or distinct sensory inputs. If feasible, the EM dataset could be leveraged to trace sensory inputs to tiles. Targeted laser ablations of central tiles that preserve the sensory periphery would also disentangle sensory versus central network contributions. Any of these approaches would help refine the understanding of tiles as repeating computational units and to what extent they require intact sensory inputs to coordinate behavior.
2. The authors conclude that more tiles enhance "perceptual resolution" (lines 387-88), based on worse foraging deficits the larger the amputation size and with more head segmentations, but these manipulations likely affect other tissues besides the tiles. Larger amputations appear to remove more of the sensory structures, which likely affects food-sensing, and the data do not show whether cutting the head into more pieces disrupts specific commissures or densities near specific organs. Ablations that remove defined numbers of tiles, or correlating the number of remaining tiles post-ablation with foraging across individuals, could provide more direct support than comparing behavior to amputation size, which is an indirect measure of tile number. The observed variability in tile number across wild-type individuals could also be used to directly relate tile counts with foraging. For segmentations, higher-resolution images of the cuts could verify which tissues are sliced. Alternatively, the authors could modify the text to reflect the caveat that effects of tile number and connectivity are being inferred from the manipulations.
Minor Comments
3. A reorganization of figures could help to streamline the manuscript and improve conceptual clarity. Figure 1 followed by Figure 3 could first lay the anatomical and molecular description of tile uniformity, while subsequent behavioral analyses in Figure 2 and 4 would provide the functional level and implications of such an organization.
4. Arena exploration is used as a measure of movement defects (Figure S2d), but because it appears negatively correlated with time spent near food, the interpretation is unclear. For instance, the text states that half-head worms forage normally while mouth-ablated worms do not, with neither showing movement impairments (lines 188-9 and lines 194-96), but Figure S2d shows higher exploration in mouth-ablated worms yet lower exploration in half-head worms. Using a metric independent of food proximity, such as movement in a food-free arena, might better isolate motor effects. A series of video examples from key-point clusters would also illustrate that behavior features look similar across manipulations.
5. A schematic of how the central brain is related to the organs, sensory structures, and muscles would help provide a high-level view of how this novel system is organized. Are the cells that make up the 2-layer ring distinct from the sensory cells interspersed in the tiles? Are there differences in cell patches that connect to sensory structures versus those that do not? Where does the central network begin, and where do the sensory and motor periphery and body organs end?
6. Though right-half amputations do not affect time spent near food (Figure 2e), there may be subtle compensatory changes, given the loss of all right-sided sensory cilia (Figure S1d). Comparing individual trajectories of pose-tracked animals could reveal behavioral biases. Regardless of the results, since most bilaterians use left/right differences to localize food, it would be interesting discussing alternative strategies in the Discussion.
7. In the Introduction, an explicit description of the diffuse-to-centralized axis, and how H. miamia differs from other distributed systems, e.g., hydra, jellyfish, or planaria, would provide relevant context for its representations as an intermediate.
8. Figure S3c: Plotting transition matrices for each condition rather than bar plots may help show the similarities and differences in transitions between conditions.
9. Figure 2k, l (and lines 268-9): The posture maps for showing effects such as "lack of foraging" or "no turning bias" (line 273) are rather abstract. A matrix of example video clips, e.g., rows representing different clusters and columns representing within-cluster examples, might be more intuitive. A similar visualization would make it easier to appreciate how similar or different a behavior is after a given manipulation.
10. Brief descriptions in the text for why each imaging method was used and the choice of markers would improve the interpretability of the experiments for general readers.
Competing interests
The author declares that they have no competing interests.
Use of Artificial Intelligence (AI)
The author declares that they did not use generative AI to come up with new ideas for their review.
-
This Zenodo record is a permanently preserved version of a PREreview. You can view the complete PREreview at https://prereview.org/reviews/18110761.
Brains evolved to perform sensorimotor transformations, enabling animals to adapt to changing environments and giving rise over time to increasingly specialized and stereotyped regions. In mammals, for example, feeding depends on distinct neural circuits for smell, taste, chewing, and internal-state integration. How and why such specialized architectures first emerged, however, remains a fundamental question.
The authors address this question in the acoel worm Hofstenia miamia, a lineage positioned between animals with diffuse nerve nets and those with regionalized brains. Using imaging, single-cell RNA-seq analyses, and targeted perturbations during foraging, they propose that the H. miamia…
This Zenodo record is a permanently preserved version of a PREreview. You can view the complete PREreview at https://prereview.org/reviews/18110761.
Brains evolved to perform sensorimotor transformations, enabling animals to adapt to changing environments and giving rise over time to increasingly specialized and stereotyped regions. In mammals, for example, feeding depends on distinct neural circuits for smell, taste, chewing, and internal-state integration. How and why such specialized architectures first emerged, however, remains a fundamental question.
The authors address this question in the acoel worm Hofstenia miamia, a lineage positioned between animals with diffuse nerve nets and those with regionalized brains. Using imaging, single-cell RNA-seq analyses, and targeted perturbations during foraging, they propose that the H. miamia brain comprises repeated computational modules capable of carrying out the sensorimotor operations required for foraging. This modular organization represents a compelling potential evolutionary intermediate from which more complex, specialized regions could arise.
This evolutionary framing is a major strength. By identifying putative ancestral building blocks of sensorimotor computation, the study provides a powerful framework for thinking about how specialized brain regions may have evolved. At the same time, the manuscript would benefit from a more explicit conceptual model describing how repeated units integrate sensory cues and drive motor programs.
Major comments:
1. The authors argue that the H. miamia brain consists of independent units capable of performing sensorimotor transformations during foraging, based on preserved prey-approach behavior after removal of different brain regions (Fig. 2e–g). However, the deficits observed following mouth removal (Fig. 2e), which likely reflects disruption of key sensory inputs, complicate the claim of full computational independence. The mouth experiments, together with the anatomical data (Fig. 2e, S1), point to specialized peripheral structures, such as the frontal organ, statocyst, and pharyngeal network, whose connectivity appears essential for effective sensorimotor transformations.
Clarifying in the text that the proposed computational units depend on specialized peripheral sensory and/or motor elements would better align the interpretation with the data. Such a clarification would not undermine the concept of repeated sensorimotor units in the central brain, but rather refine the model by emphasizing that their computational competence is conditional on intact peripheral inputs. If feasible within the scope of the existing EM dataset, tracing connections between peripheral sensory and motor structures and individual brain units could further test whether all units are equally equipped to support foraging behavior, while also highlighting the exceptional richness of the presented anatomical data.
Minor comments:
2. The behavioral impairment observed only in the six-fragment condition (Fig. 4f) is difficult to interpret. It is unclear why six-fragment cuts disrupt foraging whereas two- or four-fragment cuts do not. One possibility is that essential pathways to peripheral structures, such as the frontal organ, are severed only at higher fragmentation levels. If possible, the above EM tracings could provide anatomical insights. Alternatively, explicitly discussing such testable hypotheses would help clarify the interpretation of this result.
3. The authors emphasize the lack of stereotypy in H. miamia brains, showing that cluster number increases with age but varies widely among same-age individuals (Fig. 1l). This variability is important for interpreting the foraging assays and offers an opportunity to further test the idea that an increased number of cell clusters supports foraging efficiency. Because removing the same anatomical region likely eliminates different numbers of clusters across animals, quantifying cluster counts removed in each manipulation and testing whether animals with fewer initial clusters show stronger deficits would clarify how cluster number relates to behavioral impact.
4. The existing variability also enables a non-manipulative test: correlating natural variation in cluster number with foraging efficiency, within and across age groups, could strengthen the proposed link between computational unit number and behavior. In addition, the reported dorso–ventral asymmetry, with the ventral half containing fewer edges (L. 84), suggests a testable prediction. Comparing dorsal versus ventral amputations could provide further support for the model.
5. The sex of H. miamia animals is not mentioned. Given the complex hermaphroditic biology of this species, anatomical or behavioral differences among males, females, or sequential hermaphrodites, such as the number of cell clusters, could influence the results and interpretation. If such data are available, including them would strengthen the manuscript. Alternatively, the sex or reproductive state of the experimental animals should be specified in the Methods, and the potential impact of sex differences briefly discussed.
6. Bilaterians, including H. miamia, may localize prey by comparing sensory input across left and right sides. A directed analysis of the hunting assay (Fig. 4c) could clarify this. Does removal of brain tissue cause a tracking deficit specifically on the corresponding side? Such an analysis would help refine the computational model: the absence of lateralized deficits may support the independence of units, whereas side-specific deficits may reveal interactions among units.
7. The authors employ a broad variety of methods to characterize the H. miamia nervous system. At times, it may be difficult for a general audience to understand why specific methods or markers were chosen. For example, the rationale for using voltage dye to visualize the neuropil (Fig. 1c) or for selecting Par3 or pERK as neuronal markers (Fig. 1e,f) is not fully explained. Are these markers validated in H. miamia, or inferred from related systems? The EM dataset may provide additional evidence (e.g., synaptic boutons) to support neuronal identity. Including brief explanations of marker selection and validation, either in the main text or the Methods, would improve clarity.
8. The authors describe the repetitive nature of the nervous system using varied terminology, such as "regions," "cell clusters," "edges," and "tiles." Although these terms appear to refer to distinct units, their relationships remain unclear. A schematic illustrating the anatomical units, their constituent cell types (including non-neuronal types), and their connectivity would help readers understand the organization of the system. Such a schematic could also serve as a simplified computational model and aid comparison with cnidarian diffuse nets and centralized bilaterian brains (Fig. 1a).
9. The term "pluripotency" is commonly used to describe stem cells with broad differentiation potential. To avoid confusion, an alternative term, such as "replica" or "repeated computational unit", may more accurately communicate the intended meaning.
10. In Fig. 2l, the amputation conditions used to generate postural cluster occupancy plots are not clearly described. In addition, legends overlap with plotted data in several figures (Figs. 4a,f; S2m,n); moving legends outside the plotting area would improve readability.
Competing interests
The author declares that they have no competing interests.
Use of Artificial Intelligence (AI)
The author declares that they did not use generative AI to come up with new ideas for their review.
-
