Induction of sharp wave–ripple complexes in vitro and reorganization of hippocampal networks Christoph J Behrens 1 , Leander P van den Boom 1 , Livia de Hoz 1 , Alon Friedman 1,2 & Uwe Heinemann 1 Hippocampal sharp wave–ripple complexes (SPW-Rs) occur during slow-wave sleep and behavioral immobility and are thought to represent stored information that is transferred to the neocortex during memory consolidation. Here we show that stimuli that induce long-term potentiation (LTP), a neurophysiological correlate of learning and memory, can lead to the generation of SPW-Rs in rat hippocampal slices. The induced SPW-Rs have properties that are identical to spontaneously generated SPW-Rs: they originate in CA3, propagate to CA1 and subiculum and require AMPA/kainate receptors. Their induction is dependent on NMDA receptors and involves changes in interactions between clusters of neurons in the CA3 network. Their expression is blocked by low-frequency stimulation but not by NMDA receptor antagonists. These data indicate that induction of LTP in the recurrent CA3 network may facilitate the generation of SPW-Rs. After initial encoding, memory traces require a process of consolidation to become long lasting 1 . This process is believed to involve interactions between the hippocampus and neocortical cell assemblies and to result in the stabilization of memory traces across the cortical mantle 2 . It has been hypothesized that consolidation is dependent on hippocampal network oscillations, such as SPW-Rs, which are observed in vivo during slow-wave sleep 3,4 and behavioral immobility 4,5 . SPW-Rs are fast (140–200 Hz) oscillations in field potential recordings that are superimposed on a slow field potential transient, where the ripples result from synchronized discharges of hippocampal CA3 cells. The mechanisms underlying the induction of SPW-Rs remain to be elucidated. It is generally accepted that during learning, synaptic plasticity results in changes in the strength of the connections between neurons within the recruited population. LTP 6 is a long-lasting change in the responses of single neuronal synapses that can be induced by high- frequency stimulation (HFS) and theta burst stimulation (TBS) pat- terns both in vivo and in vitro 6,7 . LTP is considered to be a good cellular model of learning and memory 8,9 . In the present study, we explored whether physiological processes that result in synaptic plasticity in the hippocampus can also lead to the generation of SPW-Rs. We specifically tested the hypothesis that protocols that induce LTP in the CA3 region can result in the subsequent generation of SPW-Rs in this area such that strengthening of synaptic coupling becomes associated with a tendency of these neurons to fire synchronously as it occurs during SPW-R. We used an in vitro preparation to activate CA3 from different stimulation sites with recurrent HFS or TBS. Recordings were made in areas CA3, CA1 and the subiculum. The physiological and pharmacological characteristics of the induced SPW-Rs were then compared with those of SPW-Rs that occurred spontaneously in ventral hippocampal slices. In a subset of experiments, the activity of individual CA3 pyramidal cells was recorded at the same time as the induction of the SPW-Rs. Thus we determined not only the contribution of a single cell to the synchronized network activity during SPW-Rs but also whether this contribution changed during SPW-R induction. The main findings of this series of in vitro experiments are that SPW-Rs can be induced with stimulation protocols that are known to induce LTP in vivo and in vitro 10–12 and that the induction is accompanied by synaptic reorganization within the recurrent network of area CA3. RESULTS Repeated HFS induces SPW-Rs Our main finding is that repeated application of HFS trains (three tetani of 40 pulses at 100 Hz; 40-s interval; repeated up to 15 times every 5 min) to the stratum radiatum of area CA1 (Fig. 1a), a protocol that successfully induced LTP (Fig. 1b), led to the generation of SPW-Rs in areas CA3 and CA1 (Fig. 1c). Such stimuli reliably induced SPW-Rs when repeated three to five times (Fig. 1d). The induced SPW-Rs persisted for at least 2 h. Similar persistence was also observed when SPW-Rs were induced with 7 instead of 15 stimulation trains (n ¼ 6; data not shown). The HFS-induced SPW-Rs, when recorded in the stratum pyramidale or stratum oriens, consisted of 30- to 80-ms positive field potential transients (sharp waves, SPWs), superimposed by a series of small population spikes (ripples) with a frequency of about 180 Hz, as shown by power spectrum analysis and the interval between the first and second peak in auto-correlation analysis, which indicated a mean interspike interval of 5.8 ms (Fig. 2a–c). Although the frequency of ripples, once induced, did not Received 11 July; accepted 21 September; published online 16 October 2005; doi:10.1038/nn1571 1 Institute for Neurophysiology, Charite ´-Universita ¨tsmedizin Berlin, Tucholskystrasse 2, 10117 Berlin, Germany. 2 Department of Neurosurgery, Zlotowski Centre for Neuroscience, Ben-Gurion University of the Negev and Soroka Medical Center, Beer-Sheva, 84105 Israel. Correspondence should be addressed to U.H. (uwe.heinemann@charite.de). 1560 VOLUME 8 [ NUMBER 11 [ NOVEMBER 2005 NATURE NEUROSCIENCE ARTICLES © 2005 Nature Publishing Group http://www.nature.com/natureneuroscience