Functional inertia reveals history-dependent organization of large-scale brain dynamics.

The brain does not process the present in isolation. Its responses are shaped by the accumulating weight of prior states, yet existing frameworks treat large-scale brain dynamics as sequences of transient configurations, leaving the constraining influence of history formally unspecified. Here we show that the brain possesses an intrinsic, representation-agnostic organizing constraint, the degree to which accumulated prior states resist ongoing reorganization, which we term functional inertia. Using an inertial state-space model replicated across two independent observational frameworks, we demonstrate that functional inertia operates coherently across three levels of brain organization: it structures activity into dynamical regimes, links these regimes to clinical and cognitive expression through a system-level inertial magnitude, and distributes across circuits in patterns that reconcile the longstanding coexistence of associative rigidity and sensory volatility in schizophrenia. Critically, removing the cumulative integration collapses this multilevel organization. Strikingly, the system-level inertial magnitude carries opposite cognitive signatures across diagnostic contexts, predicting better performance in healthy individuals but greater symptom severity in schizophrenia, reflecting context-dependent expression of a single constraint. These findings position functional inertia as a unifying, multilevel constraint on brain reorganization, recasting stability and volatility not as opposing properties but as context-dependent expressions of a single constraint.