Reversal of LTP by theta frequency stimulation

Transcription

1 "Brain Research, 600 (1993) Elsevier Science Publishers B.V. All rights reserved /93/$06.00 BRES Reversal of LTP by theta frequency stimulation John Larson, Peng Xiao and Gary Lynch Center for the Neurobiology of Learning and Memory, University of California, IrL,ine, CA (USA) (Accepted 28 July 1992) Key words: Long-term potentiation; Long-term depression; Theta rhythm; Hippocampus; CA1; Norepinephrine; Adenosine Reversal of long-term potentiation (LTP) by physiological stimulation was tested in the CA1 field of hippocampal slices. In control medium, a one minute episode of 5 Hz (theta frequency) stimulation beginning 1-3 min after LTP had no effect on the degree of potentiation measured 30 rain later. However, in the presence of norepinephrine (200/zM), 5 Hz stimulation reduced LTP by about 30%. Theta frequency stimulation was only effective when administered within 10 min of LTP induction and had no lasting effects on non-potentiated synapses. Stimulation at 1 Hz did not reverse LTP and stimulation at 10 Hz was no more effective than 5 Hz stimulation. LTP could be nearly completely reversed by theta frequency stimulation when potentiation was induced by milder and more naturalistic stimulation patterns. Under these conditions, LTP reversal was blocked by an antagonist of adenosine A 1 receptors. These results suggest that the hippocampal theta rhythm promotes both the induction of LTP and its subsequent reversal with the latter process involving activation of adenosine receptors. Reversal of LTP may function to refine or sharpen recently encoded representations. INTRODUCTION Recent work suggests links between the endogenous firing patterns of hippocampal neurons in behaving rats and the induction of long-term potentiation (LTP) of synaptic transmission. During exploration and other behaviors associated with information acquisition, the hippocampal EEG is dominated by the theta rhythm, a large and regular 4-8 Hz activity pattern 33 (see ref. 6 for review). The excitability of pyramidal cells in the CA3 and CA1 fields is modulated by theta with cells discharging either single spikes or short bursts (2-7 spikes in ms) in phase with the rhythm v'27,3. LTP of the CA3 to CA1 connections (Schaffer collaterals/commissural system) was found to be optimally elicited by short high frequency bursts repeated at the theta frequency (theta burst stimulation; TBS); bursts repeated at higher or lower frequencies than theta produced less potentiation 22. The optimal inter-burst interval of about 200 ms reflects a priming effect whereby stimulation of one input to a CA1 cell enables a burst to the same or a separate input to induce LTP within this time window 1,19. Priming is associated with a suppression of the feed-forward inhibition that nor- mally accompanies excitatory synaptic responses19'28; disinhibition allows more effective summation of the excitation associated with burst stimulation, ultimately leading to activation of NMDA receptors and LTP induction 2 'z1. Recent evidence indicates that the IPSP suppression produced by priming stimulation involves depression of GABA release mediated by presynaptic GABA B receptors 26 (see also ref. 9). The relationship between theta, single spike activity, and LTP has received less attention. Single pulses at 5 Hz are relatively ineffective for LTP induction; rather they produce transient forms of homo- and heterosynaptic depression 12. However, low frequency stimulation (1-5 Hz) has been found to reverse LTP in situ in both acute 4 and chronic 32 preparations. The latter experiments are particularly important because they demonstrated that LTP does not reappear even twenty-four hours after the low frequency stimulation episode and the LTP induced at that time was comparable to that found in 'naive' synapses; thus, the effect observed is a true reversal of potentiation rather than a transient depression. Fujii et al. 13 subsequently obtained significant reversal of LTP in slices after 1 Hz stimulation but the reversal was not complete and Correspondence: J. Larson, Bonney Center, University of California, Irvine, CA, , USA. Fax: (1) (714)

2 98 maximal effects required several minutes of repetitive stimulation. Furthermore, control inputs were not monitored so it is unclear whether or not the observed depression was synapse-specific. The present experiments tested for interactions between theta pulse stimulation (continuous 5 Hz stimulation) and LTP induced with theta bursts. The results indicate that 5 Hz stimulation reverses LTP if it occurs shortly after induction of potentiation, that 5 Hz stimulation is more effective than 1 Hz stimulation in this regard, and that reversal is enhanced by norepinephrine. Thus, theta appears to play an important role both in inducing and in blocking the formation of long-lasting changes in synaptic strength. Finally, LTP reversal was blocked by an antagonist of adenosine A t receptors suggesting that theta pulse stimulation releases adenosine or that its reversing effects require interactions with normally present concentrations of the nucleoside. MATERIALS AND METHODS Hippocampal slices were prepared from adult male Sprague- Dawley rats ( g) and maintained in an interface chamber at 35_+ 1 C perfused with control medium containing (in mm): NaCI 124, KC1 3, KH2PO , MgSO 4 2.5, CaC12 3.4, o-glucose 10, NaHCO 3 26, and L-ascorbate 2. The upper surfaces of slices were exposed to humidified 95% O 2 and 5% CO 2. The experiments were conducted using a two-input paradigm ~2 in which different stimulation treatments were applied to independent sets of afferents in the same slice. Two bipolar stimulation electrodes (twisted strands of nichrome wire) were placed in the Schaffer collateral/commissural fibers in each slice; one was placed in the stratum radiatum of field CAla and the other in stratum radiatum of CAlc. One of the stimulation electrodes was arbitrarily designated as S1 and the other as $2 at the beginning of each experiment; the positions of these in the slice were balanced across experiments. Stimulus pulses were monophasic and 0.1 ms in duration. Population excitatory postsynaptic potentials (EPSPs) were recorded from the apical dendritic region of field CAlb using glass microelectrodes filled with 2 M NaC1. Three stimulation paradigms were used for LTP induction and reversal. In the first series of experiments, LTP was induced on each pathway (S1 and $2) independently using theta burst stimulation consisting of 10 high frequency bursts (100 Hz, 4 pulses each) repeated at five bursts per second. During TBS, the duration of each stimulus pulse was ms. LTP reversal was tested by applying theta pulse stimulation (TPS), a one-minute episode of 5 Hz stimulation to the $2 pathway. The frequency dependence of LTP reversal was tested by using a one-minute episode of either 1 or 10 Hz stimulation. In the second paradigm, LTP was induced on both pathways simultaneously using TBS as described above but without changing the pulse duration during the bursts. TPS as described above was applied to the $2 pathway to reverse LTP. In the third paradigm, LTP was induced by applying a pair of bursts (200 ms interval) to both S1 and $2 simultaneously; this was followed by a 30 s episode of TPS to the $2 pathway to reverse LTP. This pattern was repeated 5 times at two minute intervals. Responses were tested alternately on the two electrodes at s intervals during baseline periods and after LTP induction/reversal. The initial slope of the field EPSP served as an index of response size. In all experiments, the degree of LTP reversal was assessed as the difference in LTP measured 30 rain after its induction on the $2 pathway relative to S1. Norepinephrine (NE, 200 /xm) and 8-cyclopentyl-l,3-dipropylxanthine (CPX, nm) were bath-applied by perfusion. Data are expressed as means _+ S.E.M.s; statistical comparisons used Student's t-test. RESULTS Failure to reuerse L TP in control medium The stimulation paradigms used in the first series of experiments are illustrated in Fig. 1A. In control medium, an episode of theta pulse stimulation (TPS) often produced a transient homosynaptic depression lasting 5-10 min and a more variable and shorter-lasting heterosynaptic depression (3-5 min). In eight slices, homosynaptic responses were % of baseline 1 min after TPS and 102 _+ 1% of baseline min later. Thus 5 Hz stimulation of the Schaffer collateral/ commissural system did not produce any long-term effects on baseline synaptic transmission. In order to test for reversal of LTP, we compared the degree of A) n Si oli Si $2 S2 ~ t I I lilt I[I IIII IIII JlJl ]111 B/CON TPS ALONE TBS ALONE TBS de]ay TPS I I I 50 I I ~ I C) NE i) ~S 150 o I00 TPS ALONE 6o 50, r r ii) ~iii) ~ ii) ~ iii) r r TIME (MINUTES) Fig. 1. Theta frequency stimulation in the presence of norepinephrine partially reverses LTP. A: diagrams of stimulation patterns used. S1 and $2 represent independent sets of synapses made by Schaffer/commissural afferents on a common population of CA1 pyramidal cells. Theta pulse stimulation (TPS; 5 Hz for 1 rain) alone was applied to either pathway to assess its effects on control (nonpotentiated) synapses. To test for LTP reversal, one input (S1) was given theta burst stimulation (TBS: 10 bursts at 5 Hz) to induce LTP and the other input ($2) was given TBS followed 1-3 min later by TPS. B and C: representative experiments showing the effects of TPS alone (i), TBS alone (ii), and TBS plus TPS (iii) on the initial slope of the population EPSP in control medium (B, CON) and in the presence of 200/xM norepinephrine (C, NE). Graphs show measurements of the initial slope of the population EPSP over time expressed as a percentage of the baseline period. Each point is the average of four consecutive responses collected at 20 s intervals. Downward arrows indicate episodes of TPS; upward arrows indicate episodes of TBS. In these experiments, TBS to the S1 and $2 pathways was always separated by at least 20 min. Disconnected points between arrows in (iii) show TBS-induced potentiation prior to TPS. Inserts show superimposed records (averages of four consecutive responses each) taken during the baseline period and 30 min after patterned stimulation. Calibration bars: 1 mv and 10 ms.

3 99 potentiation observed in one pathway ($1) that received TBS alone with that exhibited by a second pathway ($2) that received TBS followed 1-3 min later by TPS. Results for one experiment are shown in Fig. lb. TBS to the S1 pathway induced a stable potentiation of about 45%; for the $2 pathway, TPS applied two minutes after TBS caused an initial depression but the potentiation present 30 min later was equivalent to that seen on $1. Thus in control medium, TPS had no effect on the degree of potentiation measured 30 rain after TBS (compare middle and right-hand panels of Fig. 1B). L TP reversal in slices perfused with norepinephrine The failure to reverse LTP in the slice as compared to the intact hippocampus 4'32 suggested the possibility that subcortical modulatory inputs to the hippocampus might be required for the effect. Of these, the most likely candidate is the noradrenergic system since it provides a substantial input to hippocampus 24 and has been previously linked to various forms of plasticity 5,16'17,23'29. Therefore, we repeated the above-described experiments in the presence of norepinephrine. Application of norepinephrine produced a slight depression of EPSPs evoked by single-pulse stimulation (7_+ 1% decrease in EPSP slope 30 min after onset of peffusion). TPS produced a more prolonged depression of homosynaptic responses in the presence of norepinephrine than in control medium but this depression still recovered to baseline levels within min. In eight slices, homosynaptic responses were 93 +_ 3% of control one minute after TPS and % of control min later. As in control medium, heterosynaptic responses recovered from depression within 5 min. The long-term potentiation elicited by TBS in the presence of norepinephrine did not detectably differ from that found in control slices (compare middle panels of Fig. 1B and 1C). However, theta pulse stimulation applied 1-3 min after the high frequency bursts caused an immediate depression of the potentiated responses; this showed a slight recovery over 20 rain with no further changes thereafter. As shown in Fig. 1C (middle and right panels), the amount of LTP found 30 min after TBS was noticeably less in the pathway given TPS (25% in this case) than in its matched control (50%). Several slices were followed for up to 2 h and no further changes were seen after 20 min. Fig. 2 summarizes the results for the two groups of experiments carried out using the paradigms illustrated in Fig. 1. As shown, TPS administered in the absence of norepinephrine had no effect on LTP (Fig. 2A) but produced a 29 _+ 7% reduction when delivered in its presence (Fig. 2B). The difference at thirty min- A) 10o ~ ]S: 75: B) lm 30m lm 30m i 5 I0 CONTROL NOREPINEPHRINE FREQUENCY (Hz) Fig. 2. Frequency-dependent reversal of LTP in the presence of norepinephrine. A and B: grouped data showing the percent increase in EPSP slope 1 minute and 30 min after TBS for the S1 and $2 pathways in the presence (n = 16) and absence (n = 12) of norepinephrine. In both cases there was no difference between the potentiation exhibited by $1 and $2 one minute after TBS (control: paired t H = 0.46, P > 0.05; NE: paired t15 = 1.03, P > 0.05). In control medium, TPS had no effect on LTP (30 min: paired tll = 0.48, P > 0.05). In norepinephrine (200 IxM), the potentiation expressed 30 min after TBS was significantly reduced on $2 compared to S1 (*, P < 0.001). The degree of potentiation after TBS alone was not affected by norepinephrine (1 min: t26 = 0.75, P > 0.05; 30 min: t26 = 1.30, P > 0.05). C) Frequency dependence of LTP reversal. Histograms show the % decrease in LTP 30 min after TBS for the $2 pathway relative to $1 for slices in which $2 was given a 1-min episode of 1 (n = 7), 5 (n = 16), or 10 (n = 8) Hz stimulation beginning 1 minute after TBS. Stimulation at 1 Hz did not significantly reduce LTP at 30 min (paired t 6 = 0.01, P > 0.05); 5 Hz and 10 Hz stimulation did (5 Hz: paired t15 = 4.86, P < 0.001; 10 Hz: paired t 7 = 4.24, P < 0.01). utes post-tbs in the degree of potentiation between the pathways that received TBS alone (open bars) versus TBS followed by TPS (hatched bars) in slices perfused with norepinephrine was highly significant (paired tls = 4.86, P < 0.001). It is important to note that the degree of potentiation observed for both pathways one minute after TBS (i.e., prior to TPS) did not differ. Frequency-dependence of L TP reversal In order to examine the frequency dependence of the LTP reversal effect, we compared one minute episodes of 1 Hz, 5 Hz, or 10 Hz stimulation given 1-3 min after TBS, all in the presence of norepinephrine. Again, the within slice control pathway given TBS alone was used as the standard for LTP. As shown in Fig. 2C, 1 Hz stimulation produced essentially no effect on either potentiated or control synapses. Stimulation at 10 Hz did reduce LTP, but the effect was no greater than that obtained with 5 Hz stimulation, despite the fact that twice as many stimulation pulses were given. Temporal constraints on L TP reversal Recent studies indicate that LTP induced in control medium is reversed by anoxia or activation of adenosine receptors within five minutes after induction 1' ~0 c)

4 100 Therefore we tested whether a similar time course of vulnerability was present for reversal by theta pulse stimulation in the presence of norepinephrine. TPS given 10 min or more after LTP induction was much less effective in reversing LTP than the same stimulation 1-3 min post-tbs (n = 7, t 6 = 1.02, P > 0.05, comparing S1 and $2), suggesting that vulnerability of LTP to physiological stimulation also lasts only for a short period after induction. Synapse-specificity of the reuersal effect The restricted time period during which partial reversal can be effected raises the possibility that the selectivity of the effect (i.e. S1 versus $2) is due to temporal pairing of TBS-TPS rather than to a true synapse specificity. To test this point, S1 and $2 were given TBS at the same time followed by TPS to $2 alone (i.e. paradigm 2 in Materials and Methods). Under these conditions, S1 exhibited a normal degree of LTP ( %) 30 rain later while that present in $2 was reduced by 27_+ 9% (n = 6, P < 0.05). Thus, the effects of theta pulse stimulation on LTP are specific to the synapses receiving the stimulation. TPS-induced LTP reversal is enhanced when LTP is induced under milder stimulation conditions Prior work indicates that a single primed burst induces significant LTP and that repetitions of the bursts simply increase the final magnitude of LTP induced l 'm. Cells in hippocampus have been observed to fire pairs of high frequency bursts with a theta period separating them27; therefore an additional series of experiments was conducted using this more realistic stimulation paradigm (paradigm 3). S1 and $2 were simultaneously given a pair of bursts (200 ms inter-burst interval) followed by a 30-s episode of TPS to $2. This pattern was repeated five times at two-minute intervals. Under these conditions, TPS resulted in a nearly complete reversal of LTP (86 _+ 15%; Fig. 3). As with the first paradigm, no further changes between 30 min and 2 h were seen. Thus LTP induced under milder stimulation conditions appears to be more susceptible to reversal by physiological stimulation. These results also confirm the selectivity of the effect since, again, the S1 synapses exhibited normal LTP. Since the LTP reversal effect was more robust using paradigm 3, we also tested whether the presence of norepinephrine was required. Slices were perfused with control medium and the experiments illustrated in Fig. 3 were repeated. LTP at S1 synapses was 29.78% ( _+ 4.35) and at $2 synapses was ( ), a significant difference (paired t 8 = 3.92, P < 0.01) that corresponds to a reversal of $2 LTP by 49.56% (_ ). 0-,,? th (2. 0 us W D_ 0 O io io0 50 A) s 1 B) TTTTT I I I I I I I II11 ~--,~-"-~..-~.o~ ~ ~ ~ s2 TTTTT 't~5 3b 4'5 6'o 7'5 9b TIME (MINUTES) so C) t i sl iii u) ~$2 z ~wr 40 H 20 5 lo TIME AFTER LAST BURST (MINUTES) Fig. 3. Near complete reversal of LTP in the presence of norepinephrine. S1 (A) and $2 (B) pathways were stimulated with simultaneous burst pairs (upward arrows) at two-minute intervals. $2 was stimulated with a 30-s TPS (downward arrows) immediately after each burst pair. After the first burst pair, a single response was collected for both inputs prior to the $2 TPS in order to compare initial potentiation in the two pathways. These were nearly identical. However, $2 responses showed very little potentiation thereafter, compared to robust LTP in the S1 input. Pulse duration was not increased during burst stimulation. C: grouped data from eight experiments, showing the degree of potentiation at times after the final burst pair. TPS significantly reduced the $2 LTP compared to S1 (t 7 = 7.55, P < 0.001). Thus LTP reversal was enhanced by but did not require norepinephrine using this paradigm. L TP reversal blocked by an adenosine A l receptor antagonist As noted above, stimulation of adenosine receptors shortly after LTP induction causes a reversal of LTP. Therefore we tested whether reversal of LTP by theta pulse stimulation was mediated by adenosine. Slices were tested with the paradigm used in Fig. 3 in the presence of norepinephrine alone or norepinephrine and nm 8-cyclopentyl-l,3-dipropylxanthine (CPX), a selective A 1 adenosine receptor antagonist.

5 101 TABLE I Prevention of L TP reversal by an adenosine A 1 receptor antagonist Initial potentiation represents the percent increase after the first burst pair but before TPS, post-tps gives the potentiation within 10 s after the last TPS, and potentiation at 30 min is the percent increase 30 min after the last TPS. Number of experiments is in parentheses. Percent reversal was measured 30 min after the last TPS. (* P < 0.01). NE alone (7) NE + CPX (10) S1 $2 S1 $2 Initial potentiation _ _ Potentiation post-tps _ * _ _ Potentiation at 30 min _ * _ Percent reversal _ * _ The data are summarized in Table I. As reported previously ~, CPX enhanced the size of the field EPSP (slope increase of 72.4% +_ 12.3, n = 11) but did not greatly affect the magnitude of LTP. Initial potentiation was measured by collecting a single response for both pathways 5-15 s after the first burst pair but before the TPS episode to $2. There were no differences between SI and $2 at this time for either group (NE: t 6 = 2.41, P > 0.05; NE + CPX: t 9 = 1.51, P > 0.05). Thirty minutes after the last episode of TPS, the $2 pathways in the NE alone group showed 63.6% (_+13.0) reversal relative to S1 (t 6 = 5.47, P < 0.001), replicating the results described in Fig. 3. However, the $2 pathways of the NE + CPX group did not show significant reversal (paired comparison of $1 vs $2: t 9 = 1.91, P > 0.05) and the mean percent reversal was significantly different from the NE alone group (t15 = 3.19, P < 0.001). Thus blockade of adenosine A 1 receptors prevents LTP reversal by theta pulse stimulation. DISCUSSION The present findings add a new dimension to the relationship between theta and plasticity in hippocampus. It now appears that different levels of cell activation (bursts and single spikes) modulated by an endogenous rhythm (theta) can either induce LTP or reverse it. Since bursts and spikes as well as lengthy periods of theta activity are common events in the hippocampus of behaving animals, it is not unreasonable to suppose that induction and reversal occur during behavior. The mechanisms responsible for the reversal effect are not clear. Gustafsson et al. 14 showed that LTP becomes expressed within s of inducing stimulation. TPS applied shortly after this period, but not several minutes later, substantially reduces the amount of LTP. These observations suggest that LTP, once expressed, passes through a type of consolidation or stabilization period during which it is vulnerable to physiological stimulation. The reversal is optimally elicited by stimulation in the theta frequency range, facilitated by norepinephrine, and prevented by an antagonist of adenosine A 1 receptors. Theta burst stimulation is known to gain its efficacy in inducing LTP by transiently suppressing inhibitory responses and thereby facilitating voltage-dependent NMDA receptor mediated currents m-2ms. It is possible that similar mechanisms contribute to the reversal of LTP. This hypothesis would require intense activation of NMDA receptors to set in motion the events leading to LTP expression and repeated, more modest activations to disrupt its stabilization. Perhaps related to these ideas is the evidence that different degrees of NMDA receptor activation can produce potentiation or depression in neocortex 3. The prevention of LTP reversal by CPX suggests the involvement of adenosine in the effect. One possibility is that theta burst stimulation in the presence of CPX induces a form of potentiation that is resistant to disruption by theta pulse stimulation. This would imply that adenosine is normally released during burst stimulation and has an inhibitory effect on LTP or its stabilization. That LTP was somewhat larger in the CPX experiments provides some support for this interpretation. A perhaps more attractive possibility is that theta pulse stimulation results in the release of adenosine which then has the effect of disrupting LTP stabilization. This interpretation is supported by evidence that exogenous adenosine acting on A 1 receptors can reverse LTP if applied shortly after induction 1. The cellular mechanisms responsible for this are unknown but may involve actions on calcium or potassium channels, camp levels, or G-proteins (see ref. 11 for re- view). While norepinephrine facilitated reversal, it was not necessary for the occurrence of the effect. The concentrations used were well in excess of those reported to produce physiological and biochemical changes in various type of tissues. Pilot studies using lower doses (50 /~M) did not obtain a significant improvement of the

6 102 reversal effect. Pharmacological and biochemical stud- ies of the effects of high levels of norepinephrine could provide clues as to the nature of the reversal mecha- nism. Work of this kind is also needed to evaluate the likelihood that ascending noradrenergic projections contribute to LTP reversal in situ. The reversal effect described here does not closely correspond to any of the forms of long-term depression (LTD) typically used in neural network modelsls,1s,31: it is specific to potentiated synapses, frequency depen- dent, and occurs only when the stimulation is tempo- rally coupled to LTP induction. The effect could serve to refine or sharpen recently acquired representations by removing potentiation in subpopulations of synapses, somewhat in the manner proposed by Crick and Mitchison 8 (see also ref. 15). The experiments de- scribed in Fig. 3 suggest one way in which this might work. Theta burst activity was delivered to a popula- tion of afferents (S1 + $2) sufficiently large to satisfy the cooperativity requirement for LTP induction 25 while subsequent activity in a subpopulation ($2) elimi- nated many of the potentiated connections. Over re- peated 'trials' (i.e. pairs of theta bursts) this resulted in a robust LTP in a much smaller group of synapses than otherwise would have been the case. Computer model- ing studies are needed to develop more explicit hy- potheses about the computational significance of these arrangements. Acknowledgments. This research was supported by the US Office of Naval Research. REFERENCES 1 Arai, A., Kessler, M. and Lynch, G., The effects of adenosine on the development of long-term potentiation, Neurosci. Lett., 119 (1990) Arai, A., Larson, J. and Lynch, G., Anoxia reveals a vulnerable period in the development of long-term potentiation, Brain Res., 511 (1990) Artola, A., Br6cher, S. and Singer, W., Different voltage-dependent thresholds for inducing long-term depression and long-term potentiation in slices of rat visual cortex, Nature, 347 (1990) Barrionuevo, G., Schottler, F. and Lynch, G., The effects of repetitive low frequency stimulation on control and 'potentiated' synaptic responses in the hippocampus, Life Sci., 27 (1980) Bear, M.F. and Singer, W., Modulation of visual cortical plasticity by acetylcholine and noradrenaline, Nature, 320 (1986) Bland, B.H., The physiology and pharmacology of hippocampal formation theta rhythms, Prog. Neurobiol., 26 (1986) Bland, B.H., Andersen, P., Ganes, T. and Sveen, O., Automated analysis of rhythmicity of identified hippocampal formation neurons, Exp. Brain Res., 38 (1980) Crick, F. and Mitchison, G., The function of dream sleep, Nature, 304 (1983) Davies, C.H., Starkey, S.J., Pozza, M.F. and Collingridge, G.L., GABA B autoreceptors regulate the induction of LTP, Nature, 349 (1991) Diamond, D.M., Dunwiddie, T.V. and Rose, G.M., Characteris- tics of hippocampal primed burst potentiation in uitro and in the awake rat, J. Neurosci., 8 (1988) Dunwiddie, T.V., The physiological role of adenosine in the central nervous system, Int. Reu. Neurobiol., 27 (1985) Dunwiddie, T. and Lynch, G., Long-term potentiation and depression of synaptic responses in the rat hippocampus: localization and frequency dependency, J. Physiol., 276 (1978) Fujii, S., Saito, K., Miyakawa, H., Ito, K. and Kato, H., Reversal of long-term potentiation (depotentiation) induced by tetanus stimulation of the input to CA1 neurons of guinea pig hippocampal slices, Brain Res., 555 (1991) Gustafsson, B., Asztely, F., Hanse, E. and Wigstrom, H., Onset characteristics of long-term potentiation in the guinea pig hippocampal CA1 region in vitro, Eur. J. Neurosci., 1 (1989) Hopfield, J.J., Feinstein, D.I. and Palmer, R.G., 'Unlearning' has a stabilizing effect in collective memories, Nature, 304 (1983) Hopkins, W.F. and Johnston, D., Frequency-dependent noradrenergic modulation of long-term potentiation in the hippocampus, Science, 226 (1984) Kasamatsu, T. and Pettigrew, J.D., Depletion of brain catecholamines: failure of ocular dominance shift after monocular occlusion in kittens, Science, 194 (1976) Kohonen, T., Self-Organization and Associatiue Memory, Springer-Verlag, Berlin, Larson, J. and Lynch, G., Induction of synaptic potentiation in hippocampus by patterned stimulation involves two events, Science, 232 (1986) Larson, J. and Lynch, G., Role of N-methyl-D-aspartate receptors in the induction of synaptic potentiation by burst stimulation patterned after the hippocampal theta rhythm, Brain Res., 441 (1988) Larson, J. and Lynch, G., Theta pattern stimulation and the induction of LTP: the sequence in which synapses are stimulated determines the degree to which they potentiate, Brain Res., 489 (1989) Larson J., Wong, D. and Lynch, G., Patterned stimulation at the theta frequency is optimal for the induction of hippocampal long-term potentiation, Brain Res., 368 (1986) Liang, K.C., Juler, R. and McGaugh, J.L., Modulating effects of posttraining epinephrine on memory: involvement of the amygdala noradrenergic system, Brain Res., 368 (1986) Loy, R., Koziell, D.A., Lindsey, J.D. and Moore, R.Y., Noradrenergic innervation of the adult rat hippocampal formation, J. Comp. Neurol., 189 (1980) McNaughton, B.L., Douglas, R.M. and Goddard, G.V., Synaptic enhancement in fascia dentata: co-operativity among coactive afferents, Brain Res., 157 (1978) Mott, D.D. and Lewis, D.V., Facilitation of the induction of long-term potentiation by GABA B receptors, Science, 252 (1991) Otto, T., Eichenbaum, H., Wiener, S.I. and Wible, C.G., Learning-related patterns of CA1 spike trains parallel stimulation parameters optimal for inducing hippocampal long-term potentiation, Hippocampus, 1 (1991) Pacelli, G.J., Su, W. and Kelso, S.R., Activity-induced decrease in early and late inhibitory synaptic conductances in hippocampus, Synapse, 7 (1991) Pettigrew, J.D. and Kasamatsu, T., Local perfusion of noradrenaline maintains visual cortical plasticity, Nature, 271 (1978) Ranck Jr., J.B., Studies on single neurons in dorsal hippocampal formation and septum in unrestrained rats. Part I. Behavioral correlates and firing repertoires, Exp. Neurol., 41 (1973) Rumelhart, D.E., Hinton, G.E. and Williams, R.J., Learning representations by back-propagating errors, Nature, 323 (1986) Staubli, U. and Lynch, G., Stable depression of potentiated synaptic responses in the hippocampus with 1-5 Hz stimulation, Brain Res., 513 (1990) Vanderwolf, C.H., Hippocampal electrical activity and voluntary movement in the rat, Electroencephalogr. Clin. Neurophysiol., 26 (1969)