Neuroprotectant strategies that have worked in rodent models of stroke have failed to provide protection in clinical trials. Here we show that the opposite circadian cycles in nocturnal rodents versus diurnal humans1,2 may contribute to this failure in translation. We tested three independent neuroprotective approaches-normobaric hyperoxia, the free radical scavenger α-phenyl-butyl-tert-nitrone (αPBN), and the N-methyl-D-aspartic acid (NMDA) antagonist MK801-in mouse and rat models of focal cerebral ischaemia. All three treatments reduced infarction in day-time (inactive phase) rodent models of stroke, but not in night-time (active phase) rodent models of stroke, which match the phase (active, day-time) during which most strokes occur in clinical trials. Laser-speckle imaging showed that the penumbra of cerebral ischaemia was narrower in the active-phase mouse model than in the inactive-phase model. The smaller penumbra was associated with a lower density of terminal deoxynucleotidyl transferase dUTP nick end labelling (TUNEL)-positive dying cells and reduced infarct growth from 12 to 72 h. When we induced circadian-like cycles in primary mouse neurons, deprivation of oxygen and glucose triggered a smaller release of glutamate and reactive oxygen species, as well as lower activation of apoptotic and necroptotic mediators, in 'active-phase' than in 'inactive-phase' rodent neurons. αPBN and MK801 reduced neuronal death only in 'inactive-phase' neurons. These findings suggest that the influence of circadian rhythm on neuroprotection must be considered for translational studies in stroke and central nervous system diseases.