Multiscale Modulation of Motor Primitives: A Complexity Mechanism Driving Emergent Exploration

Complex biological behaviors can emerge when a limited repertoire of elementary actions is dynamically modulated and composed across time. In movement ecology, this architecture poses a fundamental micro-to-macro question: how do microscopic motor primitives operating at distinct timescales organize to generate macroscopic exploration strategies with diverse geometries and functional consequences? Here, we address this problem by investigating the endogenous exploration of { Drosophila melanogaster} larvae under stimulus-minimal conditions. We develop a generative, multiscale framework that treats trajectory geometry as a hierarchical dynamical process rather than a static sequence of behavioral states. We demonstrate that larval movement is governed by three non-linearly coupled processes operating on separated timescales: fast propulsive cycles driving elementary displacements, intermediate steering dynamics shaping run curvature, and slow run-to-cast transitions redirecting the axis of exploration. By combining automated posture-based tracking with a minimal stochastic model of steering, we show that the classic spectrum of search behavior—from localized to dispersive and prospective exploration—arises naturally from distinct patterns of relative modulation among these same primitive components. Our results demonstrate that complex macro-level trajectories do not require specialized behavioral motifs, but instead emerge from the temporal orchestration of low-dimensional motor microstates, offering a general organizational principle for adaptive search across biological scales.