The irreversible lesions technique precludes the analysis of the possibly critical role played by discrete brain sites in the several distinct stages of mnemonic processing (acquisition, consolidation, retrieval) during which these may be specifically but transiently active. On the contrary, the reversible functional inactivation techniques, by means of stereotaxic local microinjection of active compounds, make it possible to suppress the neuronal function of a discrete volume of nervous tissue, for a pre-determined time, with the assurance of complete functional recovery within a known duration. This technique makes it possible to block the neural activity of a chosen neural site at a given stage of memory processing without any interference with the function of the same structure either during earlier or later stages of the same process. Thus, the reversible ablation results may provide information not only on the qualitative topographical but also on the quantitative temporal dimension of learning and memory. The technique employed to cause totally reversible neural inactivation is detailed. The employment of several agents to obtain functional inactivation is discussed. Of these, perhaps the safest and most manageable is tetrodotoxin when a fairly long functional inactivation (e.g., 1 h) is desired. The effects of a reversible inactivation can be quite easily and accurately assessed by observing the severity of the amnesic disruption, if any, of a conditioned response. In order to do this as well as possible, it is advantageous to employ a very simple behavioral paradigm. The passive avoidance response in the light-dark box apparatus fulfills this requirement. Moreover, this paradigm, being one-trial, provides the necessary condition of a single well-defined temporal beginning. The present protocol has been successfully employed in learning and memory research, to assess when the functional integrity of a given neural structure is necessary in order that a conditioned response may be acquired, consolidated or retrieved. The employment of this protocol in relation to the intrinsic functional characteristics of a given subcortical neural site is discussed.