Biobehavioral Reviews, Vol. 1, pp. 143-150, 1977. Copyright © ANKHO International Inc. All rights of reproduction in any form reserved. Printed in the U.S.A. Memory Facilitation by Posttrial Hypothalamic Stimulation and Other Reinforcers: A Central Theory of Reinforcement JOSEPH P. HUSTON, CORNELIA C. MUELLER AND CESARE MONDADORI Institute of Pharmacology, University of Zftrich, Gloriastrasse 32, 8006 Zftrich, Switzerland (Received 15 June 1977) HUSTON, J. P., C. MUELLER AND C. MONDADORI. Memory facilitation by posttrial hypothalamic stimulation and other reinforcers: a central theory of reinforcement. BIOBEHAV. REV. 1(3) 143-150, 1977. - A central theory of reinforcement is presented, which is based on the assumption that reinforcers act directly on dynamic memory processes. One prediction from the model is that learning of various tasks should be improved as a result of reinforcement presented during the period of short-term memory. To test this hypothesis the reinforcers food, water or electrical stimulation of the lateral hypothalamus were presented posttrial in diverse learning situations. Posttrial food reinforcement facilitated passive avoidance learning in mice. In rats 30 sec of posttrial reinforcing brain stimulation facilitated learning of a shuttle-box avoidance, step-down avoidance, and small box avoidance. Appetitive T-maze learning was improved with posttrial reinforcing brain stimulation contingent on errors. Learning of a conditioned taste aversion was not influenced by reinforcing brain stimulation presented during the CS - UCS interval. Hypothalamic stimulation Short-term memory Reinforcement Brain stimulation THIS paper presents a central theory of reinforcement based on hypothetical memory processes, followed by a review of experimental results of relevance for the theory. A MEMORY PROCESSING THEORY OF REINFORCEMENT In the context of traditional reinforcement theory memory is often considered as something that is somehow determined by and, thus, follows reinforcement. This has usually led to a separation between theories of memory and theories of reinforcement. The following account integrates the concepts of memory and reinforcement by invoking hypothetical memory processes to account for the action of reinforcers on behavior (see also [ 16,17] ). We can start with the proposition that reinforcers act on memory processes. This simply means that reinforcing events prevent memory traces from fading. In summary, reinforcers will be considered to prevent an immediate memory trace from fading, and thus, to establish a short-term memory trace, which, if it persists, will create a long-term memory trace. In the simple operant conditioning situation the rein- forcer is said to act backwards in time on the preceding response or the stimulus-response connection, which is thus strengthened as a result. We can go further and hypothesize that the reinforcer acts on a memory of the response or of the stimulus-response contiguity. This requires the assump- tion that any response or operant (or discriminative stimulus-operant behavior contiguity) leaves an immediate memory trace, which normally fades in a short time unless it is strenghened or prolonged by some extra critical event. This critical event is the presentation of a reinforcer. The z Supported by SNSF grant No. 3, 6610.75. This manuscript is also Symposium, Magdeburg, GDR, June, 1977. reinforcer strengthens or prolongs this immediate memory to the extent that it is presented close in time to the response. Thus, the closer the reinforcer occurs in time to the response, the less likely is it that the memory has faded (the stronger the still existing memory trace), and the more likely that the trace can be maintained at a higher level, and thus, be established as a short-term memory. Figure 1 schematically shows the relationship between the delay of reinforcement and the hypothetical decay function of an immediate memory trace (arbitrarily drawn to be convex, rather than linear or concave). It shows that an immediate reinforcer (RI) will establish a more prominent short-term memory trace, than a delayed reinforcer (R~). Such a temporal gradient of strength of reinforcement has long been supposed from studies dealing with the effects of delayed reinforcement [ 32,37 ]. Hence, according to our model, a reinforcer strengthens memory simply by ensuring that it persists in time and enters a more prolonged, but labile state, which tradi- tionally is called short-term memory storage. The concept of a labile short-term memory storage has been character- ized and operationalized by a large body of evidence [25] showing that within a restricted period of time after a learning trial an electroconvulsive shock (or other treat- ments, including brain stimulation and drugs) will prevent learning (i.e. will lead to a retrograde amnesia for the task). Thus, during a limited period of time after a learning trial the memory trace is not yet permanent or fixed (or consolidated) into a long-term memory. (The period of long-term, or permanent memory storage subsequent to short-term storage can be defined, for example, by the relative failure of post-trial amnestic treatments to influ- published in the Proceedings of the 5th International Neurobiology 143