Shuttling between critical regimes to break the responsivity–persistence trade-off

Collective systems confront a fundamental trade-off: they must remain responsive to external cues while maintaining persistence against noise. While the criticality hypothesis posits that information processing is optimized in the near-critical regime, the microscopic mechanisms enabling systems to dynamically navigate this balance remain poorly understood. Here, we identify the Motion Salience Threshold, a parameter regulating the switch between averaging and selective interactions, as a tunable control parameter that organizes critical-like collective dynamics. Through the analysis of cascade dynamics via collective sensitivity and the empirical branching ratio, we demonstrate that adjusting this threshold precipitates a crossover from a sensitive supercritical regime to a robust subcritical one. Mapping these dynamical states onto a performance landscape reveals that responsivity peaks near the crossover region, whereas persistence dominates the subcritical regime. Crucially, we show that this trade-off is not insurmountable. We introduce a decentralized adaptive strategy driven by local salience variance. This mechanism allows the collective to dynamically shuttle between responsive (supercritical) and persistent (subcritical) states. Consequently, the system breaks the static trade-off curve, yielding a 20%–30% enhancement in responsivity at equivalent persistence levels. Our study establishes a mechanistic link between microscopic attention rules and macroscopic collective phenomena, offering a theoretical foundation for the design of adaptive swarm systems.