Abstract The rescue of cognitive function through environmental enrichment (EE) during aging has been extensively documented. However, the age at onset, the duration of EE, and the cerebral mechanisms required to obtain the greatest benefits still remain to be determined. We have recently shown that EE applied for 3 mo after the median lifespan, i.e., the age at which 50% of the population is still alive (from 17 to 20 mo in NMRI mice), failed to prevent cognitive deficits in senescent animals. In the present study, mice were exposed to EE prior to the median lifespan, and for a longer total duration (from 14 to 20 mo), before the assessment of memory performance and the electrophysiological properties of hippocampal neuronal networks. The EE prevented memory deficits and reduced anxiety as the animal aged. Moreover, EE attenuated the age-related impairment of basal glutamatergic neurotransmission in CA1 hippocampal slices, and reversed the decrease in isolated N-methyl-D-Aspartate receptor (NMDA-R)-dependent synaptic potentials. Surprisingly, EE did not prevent the age-related alteration of theta-burst-induced long-term potentiation (LTP). This study therefore suggests that EE needs to be initiated before the age corresponding to the median lifespan and/or required long duration (> 3 mo) to have an effect on cognitive aging. In addition, we show that EE probably acts through theta-burst-independent mechanisms of synaptic plasticity.