From Brainstem Arousal to Cortical Prediction: A Circuit-Based Framework for Active Flavor Inference

Flavor perception exhibits substantial variability across physiological and contextual states, yet how peripheral gustatory signals are transformed into these subjective experiences remains incompletely understood. Traditional models of taste processing have largely focused on bottom-up sensory relay from peripheral receptors to cortical regions, and have limited capacity to explain this variability across physiological and contextual conditions. To bridge this gap, this study proposes Active Flavor Inference (AFI) as a conceptual framework that reinterprets gustatory processing in light of predictive coding and interoceptive regulation. AFI conceptualizes flavor perception not as an isolated sensory outcome, but as an active inferential process arising from functional interactions among brainstem arousal circuits, thalamic gating mechanisms, and cortical networks, through which sensory evidence, physiological state, and higher-order predictions are integrated. Drawing on converging evidence from animal models and human neuroimaging, we propose that ascending sensory signals and descending regulatory influences converge through distributed architectures rather than single linear pathways. Ultimately, AFI provides a structured framework to interpret state-dependent flavor distortions in clinical contexts and generates experimentally testable predictions to guide future high-resolution gustatory research. • Flavor perception is reframed as an active inferential process rather than a fixed sensory output. • Active Flavor Inference (AFI) integrates sensory evidence, physiological state, and prediction. • Human gustatory processing is interpreted through convergent evidence across sensory domains. • Brainstem and neuromodulatory systems are emphasized as indirect, state-dependent regulators. • AFI provides a conservative framework for generating testable questions in human taste research.