Abstract
An evolutionary account of human language as a neurobiological system must distinguish between human-unique neurocognitive processes supporting language and evolutionarily conserved, domain-general processes that can be traced back to our primate ancestors. Neuroimaging studies across species may determine whether candidate neural processes are supported by homologous, functionally conserved brain areas or by different neurobiological substrates. Here we use functional magnetic resonance imaging in Rhesus macaques and humans to examine the brain regions involved in processing the ordering relationships between auditory nonsense words in rule-based sequences. We find that key regions in the human ventral frontal and opercular cortex have functional counterparts in the monkey brain. These regions are also known to be associated with initial stages of human syntactic processing. This study raises the possibility that certain ventral frontal neural systems, which play a significant role in language function in modern humans, originally evolved to support domain-general abilities involved in sequence processing.
Introduction
A central debate in evolutionary biology and linguistics is whether the human capacity for language is a "de novo" biological miracle or an adaptation of older systems found in our primate ancestors. Specifically, researchers want to know if the brain regions we use for syntax—the rules governing word order—are unique to humans or if they perform broader, "domain-general" functions shared across species. While humans are the only species with full-blown language, other primates can learn simple "artificial grammars" (sequences of sounds following specific rules). This study addresses whether the neural substrates supporting this sequence processing are homologous across humans and macaques, potentially providing a "missing link" in the neurobiological history of language.
Methods
The researchers employed a comparative functional Magnetic Resonance Imaging (fMRI) approach involving 3 Rhesus macaques and 12 human adults.
Artificial Grammar Learning (AGL)
Both species were exposed to sequences of auditory "nonsense words" (e.g., klor, biff, jux) generated by a branching rule system. After an initial exposure phase to learn these rules implicitly, subjects were scanned while listening to "Consistent" sequences (following the rules) and "Violation" sequences (breaking the rules).
Imaging and Analysis
The team used fMRI to identify brain regions sensitive to sequence violations. They performed both whole-brain analyses and targeted Region of Interest (ROI) analyses, specifically looking at the ventral frontal and opercular cortex (vFOC) and Brodmann Areas (BA) 44 and 45. Control experiments ensured that brain responses were specific to violations of learned statistical regularities rather than just novel sounds.
Results
Behavioral Sensitivity
Both macaques and humans successfully learned the grammar, showing significantly stronger responses to violation sequences compared to consistent ones ($P < 0.001$). Macaques demonstrated this through increased looking-time toward the speaker, while humans performed a "same/different" classification task.
Neuroimaging Findings
The most striking finding was that the vFOC showed significant activation in response to sequence violations in both species. In macaques, violation sequences triggered significant clusters in the right ventral frontal cortex, specifically the opercular and dysgranular insular cortex ($Z$-scores up to 4.89). Humans showed similar, robust bilateral activation in the vFOC ($Z$-scores up to 3.55). A direct cross-species comparison confirmed that the vFOC is significantly more sensitive to these violations than the adjacent BA 44/45 across both species ($F = 234.6, P < 0.001$).
Discussion
The results suggest that the vFOC is a functionally conserved substrate for tracking local ordering relationships. Because the task used nonsense words and simple rules, it implies that the vFOC is a domain-general processor for statistical sequences that humans later co-opted for language.
In humans, complex hierarchical syntax typically engages BA 44/45. In this study’s simpler task, vFOC was the primary driver for both species, while BA 44/45 showed more variable activity. This hints at a functional hierarchy where the vFOC handles basic sequencing, while BA 44/45 may have specialized further in the human lineage for complex syntax. Furthermore, the lack of strong left-lateralization suggests that while the substrate is conserved, the specific specialization for language involves further evolutionary steps, such as increased connectivity or lateralization.
Future research could use these identified regions as "seeds" for tractography studies to see if the white-matter pathways connecting these regions have also been evolutionarily conserved or if they have uniquely expanded in humans.
Summary updated