The role of NMDA receptors in information processing

In this review, we have concentrated on the parallels between the cellular properties of the NMDA receptor and a variety of functional properties within sensory and motor systems. Of course, the NMDA channel exists within the cell in conjunction with a variety of other channels, including non-NMDA channels. Although the NMDA receptor is unique in a cellular sense - it is the only ligand-gated channel that is also voltage dependent and calcium permeable - it is not unique in a functional sense. A cell that has non-NMDA receptors and voltage-sensitive channels will also exhibit nonlinear behavior. Moreover, Buhrle and Sonnhof (1983) demonstrated some time ago that calcium flows into frog motor neurons through more than one type of calcium channel. The contribution to the inflow of calcium from NMDA channels may vary from cell to cell and could easily be a minor proportion of the total. Many authors have pointed out that the NMDA channel has a low conductance at a resting potential of -70 mV. However, many cells in the nervous system are depolarized from -70 mV by excitatory input. Thus, as pointed out above, NMDA receptors make a contribution to the tonic or spontaneous activity of cells in both visual cortex and spinal cord. In practice, many cells are probably working in a range of membrane potentials where the NMDA channels are always open to some extent. Even in the hippocampal slice where a substantial amount of afferent input is removed, NMDA receptors contribute to spontaneous activity (Sah et al 1989). Does the NMDA receptor act as a switch? Does it act as an AND gate? The suggestion that it may act as a switch comes from work on LTP in the hippocampus, which is readily produced by high-frequency stimulation and is abolished by APV. However, activation of the NMDA receptor is only the first in a sequence of reactions leading to LTP: In theory, switchlike behavior could also be produced by calcium-buffering systems within dendritic spines, or by enzymatic processes (Lisman 1985 Zador et al 1990). Fox and Daw (1992) have modeled the action of NMDA and non-NMDA receptors that are activated in parallel with each other, and shown that the occurrence of switch-like behavior depends on the relative density of NMDA versus non-NMDA receptors. Switch-like behavior is not seen in the visual cortex, but might be seen in the hippocampus if the relative density of NMDA receptors there was higher than in the visual cortex. Our review did not turn up any specific examples of situations in which switch-like behavior or AND gate behavior has been proved to be caused by NMDA receptors. In fact, quite the opposite. We found that, far from creating a sharp threshold, the voltage dependence of the NMDA receptor generally just leads to an amplification. This area is clearly one for fruitful investigation in the future. On the positive side, several cellular properties of the NMDA channels low rise time, prolonged effect, temporal summation, and amplification of responses due to the voltage dependency - can be related to function. This is particularly true when the function is frequency specific, as in motor rhythms, or develops slowly over time, as in nociception. There are also situations in which NMDA receptors contribute to a low frequency function, and non-NMDA receptors to a high frequency function-jamming avoidance response and 'chirping' in the fish electric organ; different frequencies of oscillation in the spinal cord - in systems where both types of receptor are found in parallel with each other. Whereas the role of NMDA receptors in plasticity, LTP, and cell death is now quite well established, investigation of their role in information processing could continue to be fruitful for some time to come.