Social hierarchy as a heuristic prior: A cross-species conceptual framework for reward processing

Social hierarchy profoundly shapes reward processing across species, yet the principles linking them remain unclear. This review proposes a unifying conceptual framework in which an individual's own social status is treated, heuristically, as a Bayesian prior, an expectation shaped by competitive experience that is proposed to constrain reward anticipation and modulate prediction error signaling. Critically, the framework distinguishes between two forms of hierarchical representation. Representations of one's own status can serve as an internal prior, whereas representations of others' status inform social valuation. Synthesizing convergent evidence from rodents, non-human primates, and humans, the framework further draws a distinction between two levels of neural implementation. At the representational level, conjunctive coding in the amygdala enables single neurons to jointly encode social rank and reward magnitude. At the level of proposed mechanism, active prior weighting offers a heuristic account of how self-related hierarchical expectations might shape anticipatory neural activity before outcomes are observed. This distinction clarifies a key ambiguity in the field and generates specific, testable hypotheses. Convergent evidence consistent with this conceptual architecture includes grid-like entorhinal codes for social hierarchy and dopaminergic social reward prediction error signals. Conserved principles are also identified, including the amygdala's dual role in integrating rank and value, alongside species-specific divergences in prefrontal circuit architecture, from action-oriented competition in rodents to inference-based navigation in primates. By conceptualizing social hierarchy as a heuristic prior, the framework offers a conceptual bridge between social structure and reward circuitry.