Conventional long-term potentiation (LTP) and long-term depression (LTD) are induced by different patterns of synaptic stimulation, but both forms of synaptic modification require calcium influx through NMDA receptors (NMDARs). A prevailing model (the "calcium hypothesis") suggests that high postsynaptic calcium elevation results in LTP, whereas moderate elevations give rise to LTD. Recently, additional evidence has come to suggest that differential activation of NMDAR subunits also factors in determining which type of plasticity is induced. While the growing amount of data suggest that activation of NMDARs containing specific GluN2 subunits plays an important role in the induction of plasticity, it remains less clear which subunit is tied to which form of plasticity. Additionally, it remains to be determined which properties of the subunits confer upon them the ability to differentially induce long-term plasticity. This review highlights recent studies suggesting differential roles for the subunits, as well as findings that begin to shed light on how two similar subunits may be linked to the induction of opposing forms of plasticity.