Spatially structured food fermentation: Integrating matrix transport, microbial function, and predictive modeling

Background Food fermentation emerges from dynamic interactions between microbial metabolism and the physicochemical structure of food matrices. Although starter cultures have improved process control, microbial performance cannot be understood independently of the environment in which microorganisms grow. In structured, semi-solid, and solid foods, limitations in heat and mass transfer generate spatially heterogeneous microenvironments characterized by gradients in temperature, oxygen, pH, water activity, nutrients, and metabolites. These local conditions can influence microbial physiology, metabolic activity, interactions, and community development, while microbial activity progressively modifies the surrounding matrix. Scope and approach This review develops a matrix-aware framework for understanding and predicting food fermentation by integrating microbial ecology, matrix properties, transport phenomena, and predictive modeling. Evidence from dairy, meat, cereal, plant-based, and solid-state fermentations is examined to describe how matrix composition, structure, porosity, hydration, and process geometry regulate heat, gas, moisture, substrate, and metabolite transport and thereby determine the local physicochemical conditions experienced by microorganisms. Particular attention is given to the connection between these microenvironments and microbial responses, from metabolic regulation to population interactions and community dynamics. Mechanistic kinetic, reaction–diffusion, heat- and mass-transfer models are evaluated together with emerging approaches integrating spatial measurements, sensors, omics, machine learning, and hybrid modeling. Experimental studies are also considered to illustrate how transport limitations and matrix properties translate into measurable heterogeneity in real fermentation systems. Key findings and conclusions The available evidence indicates that bulk process variables can mask substantial spatial and temporal heterogeneity within structured foods, meaning that microorganisms located within the same fermentation system may experience markedly different physicochemical conditions. Microbial functionality therefore reflects not only intrinsic metabolic capacity or starter identity, but also the ability of the food matrix to transport heat, gases, water, nutrients, and metabolites. This relationship is bidirectional: microbial growth, substrate conversion, acidification, gas and metabolite production, heat generation, and structural changes continuously reshaping the matrix and its transport properties. A matrix-aware approach consequently provides a mechanistic bridge between food structure, local microenvironment, microbial function, community dynamics, and fermentation outcomes. Combining spatially resolved measurements with mechanistic and data-driven models could improve starter culture selection, process prediction, scale-up, and control while reducing reliance on empirical optimization. Such integration may support more consistent and resource-efficient fermentation processes, particularly for structured foods, emerging plant-based products, and the valorization of food side-streams.