The “source specificity” hypothesis of NMDA receptor excitotoxicity states that calcium entering specifically through the NMDAR is toxic, due to coupling of the receptor to calcium dependent neuron-specific death effectors associated with the NMDAR. The alternative “calcium load” hypothesis contends that NMDARs promote death pathways solely due to the amount of calcium entering the cell. Analysis of NMDA receptor-dependent cell death revealed that it relies on both JNK and p38 pathways in neurons. Moreover, disruption of p38 signalling could be achieved by uncoupling the NR2B PDZ-binding domain from downstream proteins by using intracellular peptides mimicking the NR2B PDZ-binding motif. This peptide was found to exert its neuroprotective effects without affecting signalling to synaptic plasticity. However, the peptide had no effect on signalling to JNK, suggesting that NMDAR coupling to JNK-dependent death may be independent of the NMDAR signalling complex. Reconstitution of NR1/NR2B NMDARs in AtT20 cells revealed that dose dependent NMDAR-dependent cell death can be recapitulated in these cells. Unlike in neurons, this cell death was not inhibited by p38 inhibitors or indeed the NR2B PDZ ligand peptide, and was reliant solely on JNK signalling. Thus, pro-death signalling from the NMDAR comprises pathways that rely on the NMDAR signalling complex, and others that are triggered simply by calcium influx, encompassing elements of both hypotheses.