1. The excitability of CA3 pyramidal neurons was assessed with intracellular recordings in hippocampal slices from behaviorally naive rabbits. CA3 pyramidal neurons had large (-13.1 ± 0.3 mV; mean ± SE) postburst afterhyperpolarizations (AHPs) and exhibited robust spike-frequency adaptation (accommodation/to prolonged (800-ms) depolarizing current injection at resting potentials of -68 mV. AHP and accommodation measures differed in scale but not in kind from those obtained in stable recordings from CA1 pyramidal neurons in the same slices or from the same rabbits, with CA3 neurons having larger longer AHPs but fewer spikes during accommodation. 2. Groups of rabbits were trained in a simple, associative-learn ing task, trace eye-blink conditioning, which requires an intact hippocampus for successful acquisition. Memory consolidation in this task also involves the hippocampus, whereas long-term retention of the learned response does not. The time course and magnitude of learning specific changes in excitability were assessed in 201 CA3 pyramidal neurons. 3. Learning increased the excitability of CA3 pyramidal neurons soon after acquisition (within 1-24 h). The mean postburst AHP was reduced to approximately half (-6.4 ± 0.3 mV) the basal amplitude of the AHP observed in naive controls. The area and duration of the postburst AHP similarly were reduced. Approximately half of all pyramidal neurons tested soon after learning exhibited significantly reduced AHPs, whereas none exhibited enhanced AHPs. 4. Trace conditioning also reduced accommodation of CA3 pyramidal neurons 1-24 h after learning. Neurons from successfully trained rabbits fired significantly more action potentials (5.6 ± 1.5) in response to prolonged depolarization than did neurons from naive controls (4.1 ± 0.2). The magnitude of the learning-specific change in accommodation was less than that for the AHP. Approximately 45% of neurons tested exhibited significantly reduced accommodation soon after learning. 5. Both learning-specific changes in CA3 increased neuronal excitability. Both changes were highly time dependent. AHPs were reduced maximally 1-24 h after learning, then increased, returning to basal (naive) levels within 7 days and remaining basal thereafter. The decay rate of accommodation to basal levels preceded that of the AHP by several days. 6. Other membrane properties, including action potential char acteristics, resting potential, and input resistance, were unchanged by learning. The restriction of the observed changes to two interrelated measures of excitability concurs with earlier reports that learning-specific changes in the mammalian hippocampus are linked to changes in a limited number of membrane conductances. 7. Learning, not long-term memory or performance of the learned behavior, was linked to the excitability changes. Neurons from rabbits that failed to acquire the task after considerable training exhibited no excitability changes. Neurons from pseudoconditioned rabbits were indistinguishable from neurons of behaviorally naive controls. Finally, neurons from rabbits that explicitly demonstrated long-term retention of the conditioned response were indistinguishable from those of naive controls. 8. Behavioral changes persisted for extremely long periods, but the observed changes in hippocampal excitability were transient and greatest soon after learning. Excitability was enhanced for a period of a few days, a period demonstrated in other eyeblink studies to be required for memory consolidation. Because hippocampal excitability then returned to basal levels but memory of the learned task persisted, postconsolidation memory traces (the 'engram') must be extrahippocampal.
ASJC Scopus subject areas