Dendritic hold and read: A novel form of information storage and retrieval

Abstract

The discovery of the dual gating of the NMDA receptor was a major turning point in the field of learning and memory. Much of the excitement was directed to the ability of the NMDA receptor to act as a coincidence detector, as well as the unusually high permeability to calcium. These unique features of the NMDA receptor provided the means at the molecular level to implement Hebb's hypothesis for synaptic plasticity. In the late 1980's it was proposed that the NMDA receptor itself could hold information in one of two ways. One school of thought believed that the slow unbinding of glutamate allowed the NMDA receptors to maintain a long lasting conductance that enables a reverberating feedback circuit to sustain spiking behavior after the input into the network has ended. Bekkers and Stevens (1990) proposed an alternate hypothesis; that information could be stored by the bound-but-blocked (non-conducting) state of the NMDA receptors. Neither group experimentally tested their theoretical consideration. Suspecting that Bekkers and Stevens failed to consider a concept that was not available to them, electrical compartmentalization of distal dendrites and regenerative NMDA spikes, I reexamined their predictions with inclusion of factors required for dendritic spikes during signal read-out. Here I show that the dual gating of the NMDA receptor can hold information in the form of bound glutamate on a large population of receptors for 100's of ms, which can be read-out by a local depolarization. Key to this model is the electrical compartmentalization of a thin terminal dendrite, which allows for a regenerative all-or-none spike. The experimental observation called, "dendritic hold and read" or DHR, is a mechanism that can support a novel form of short-term memory.