Brain–Behavior Differences in Premodern and Modern Lineages of Domestic Dogs

Barton et al. (2025) present compelling evidence for structural brain differences between “premodern” and “modern” domestic dog lineages, with modern breeds showing cortical expansion associated with trainability, and premodern dogs exhibiting relatively larger subcortical regions, notably the amygdala, linked to fear-related behavior. This study offers important insights into the neurobiological correlates of breed-associated behavior.

However, we urge caution in interpreting these findings through the lens of evolutionary chronology. The core classification of dogs into “modern” and “premodern” groups introduces conceptual ambiguity. The so-called premodern group includes a heterogeneous mix of ancient breeds, village dogs, and New Guinea Singing Dogs (NGSDs), which differ markedly in ecological exposure, socialization, and selection history. While the authors cite population genetic studies, ancient breeds such as Shiba Inus and Samoyeds have nonetheless undergone modern selective breeding and do not represent an evolutionary baseline.

Conversely, the modern sample is heavily weighted toward working-line breeds selected for high trainability and cooperative behavior. Prior work by the same group (e.g., Hecht et al., 2019) has demonstrated that regional cortical volume correlates with trainability scores. Thus, the observed anatomical differences may reflect functional behavioral selection rather than broad evolutionary divergence. Notably, the study excludes low-trainability modern breeds (e.g., Bulldogs, Pugs), limiting the generalizability of cortical expansion as a feature of modernity per se.

Moreover, although the authors acknowledge that experiential plasticity cannot fully account for their findings, the functional specialization of the modern sample introduces a confound. Without broader sampling across behavioral spectra, it remains difficult to disentangle the contributions of selection history, environmental experience, and heritable morphology.

The evolutionary framing is also problematic. Terms such as “brain evolution,” “continued selection,” and the “premodern to modern transition” imply diachronic processes that are not directly assessed. All subjects in the study are contemporary dogs, and no fossil, archaeological, or longitudinal data are presented. While NGSDs may be genetically distinct, they are not temporally ancestral in any direct sense.

We propose a revised framework that categorizes dogs along two continuous axes: (1) behavioral selection history (e.g., working, companion, free-breeding) and (2) quantitative trainability measures (e.g., C-BARQ scores). Such an approach would allow multivariate modeling of brain-behavior relationships that better distinguishes lineage from functional specialization.

In sum, Barton et al.’s study makes a valuable contribution to canine neuroscience. Yet to avoid overextension of evolutionary claims, future work should employ terminology and sampling strategies that more precisely reflect the diversity of selection histories and behavioral phenotypes across dog populations.