Agent-Based Modeling (ABM) as a Complementary Framework to Differential Equations for Collective Sensory Processing: An Exploratory Study

Introduction: Complex adaptive systems, particularly neural circuits that integrate sensory information, exhibit collective patterns emerging from local interactions between individual units (collective sensing). This work explores when agent-based modeling captures phenomena that population-level differential equations cannot describe, and how both frameworks articulate in a multiscale approach. Methods: We propose representing each sensory unit as a discrete agent with its own activation threshold and local interaction rules with immediate neighbors (direct coupling and imitation). The main guiding case models the propagation of a sensory disturbance through a population of sensory channels. As an illustration of the framework's extensibility, this same scheme could be transferred to economic agents to model expectation contagion or herd behavior in markets with incomplete information. Expected Results: We expect to obtain a conceptual design of the ABM and a formal approach to its continuous limit (mean-field). We aim to identify under what conditions local rules give rise to phenomena like partial or asynchronous stimulus propagation that a purely aggregate description does not capture with the same resolution. Conclusions: This framework specifies when ABM provides unique insights into collective sensory processing dynamics that a population-level differential equations approach could not show. Keywords: Agent-Based Modeling, Computational Neuroscience.