Consensus across limbic areas predicts motivational disengagement

AbstractMotivational disengagement is widely attributed to low expected reward, yet animals and humans disengage even when reward expectations are constant. This implies a latent mechanism beyond reward. After accounting for expected reward, we found a neural signal across multiple limbic areas that predicted disengagement. As animals disengaged, neural activity across areas converged onto a shared, low-dimensional axis. Consensus across areas predicted whether disengagement persisted, suggesting why single brain area perturbations have produced inconsistent effects. We showed that engagement and disengagement operated as distinct attractor states whose strength scaled with expected reward, revealing that value modulates, but does not determine, motivational state. These findings reframe motivation as a collective computation and suggest that effective interventions for motivational deficits may require targeting multiple areas simultaneously.