It is well established that cocaine induces an increase of dendritic spines density in some brain regions. However, few studies have addressed the role of this neuroplastic changes in cocaine rewarding effects and have often led to contradictory results. So, we hypothesized that using a rigorous time- and subject-matched protocol would demonstrate the role of this spine increase in cocaine reward. We designed our experiments such as the same animals (rats) were used for spine analysis and behavioral studies. Cocaine rewarding effects were assessed with the conditioned place preference paradigm. Spines densities were measured in the two subdivisions of the nucleus accumbens (NAcc), core and shell. We showed a correlation between the increase of spine density in NAcc core and shell and cocaine rewarding effects. Interestingly, when cocaine was administered in home cages, spine density was increase in NAcc core only. With anisomycin, a protein synthesis inhibitor, injected in the core we blocked spine increase in core and shell and also cocaine rewarding effects. Strikingly, whereas injection of this inhibitor in the shell immediately after conditioning had no effect on neuroplasticity or behavior, its injection 4 hours after conditioning was able to block neuroplasticity in shell only and cocaine-induced place preference. Thus, it clearly appears that the neuronal plasticity in the NAcc core is essential to induce plasticity in the shell, necessary for cocaine reward. Altogether, our data revealed a new mechanism in the NAcc functioning where a neuroplasticity transfer occurred from core to shell.