Communicated by Mikhail Tsodyks Stable and Rapid Recurrent Processing in Realistic Autoassociative Memories

It is shown that in those autoassociative memories that learn by stor-ing multiple patterns of activity on their recurrent collateral connections, there is a fundamental conict between dynamical stability and storage capacity. It is then found that the network can nevertheless retrievemany different memory patterns, as predicted by nondynamical analyses, if its ring is regulated by inhibition that is sufciently multiplicative in nature. Simulations of a model network with integrate-and-re units con-rm that this is a realistic solution to the conict. The simulations also conrm the earlier analytical result that cued-elicited memory retrieval, which follows an exponential time course, occurs in a time linearly related to the time constant for synaptic conductance inactivation and relatively independent of neuronal time constants and ring levels.