| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Neuroscience Laboratory, Institute for Medical Sciences, Ajou University School of Medicine, Suwon 443-721, Korea
 |
ABSTRACT |
|---|
 TOP ABSTRACT References |
|---|
proposed that synaptic weight enhancement induced by conjoint activation of connected neurons can lead to the formation of a new functional assembly. Subsequent theoretical studies have confirmed that associative memories (i.e., functional assemblies) can be stored in a network of interconnected neurons on the basis of the Hebbian learning rule (Marr 1971; Kohonen 1977; Hinton and Anderson 1981; Hopfield 1984; McNaughton and Morris 1987). Because N-methyl-D-aspartate (NMDA) receptor-dependent long-term potentiation (LTP) obeys the Hebbian learning rule (Bliss and Lomo 1973; McNaughton et al. 1978; Collingridge et al. 1983; Nowak et al. 1984), it has been regarded as the substrate for associative memory. Although this view has garnered a growing body of supporting evidence (for review, see Martin et al. 2000; Martin and Morris 2002), there still exist a few missing links between LTP and memory. One essential piece of evidence for linking LTP and memory is the demonstration that LTP induction leads to changes in functional relationship among neurons.
We examined whether LTP induction leads to detectable changes in "functional connectivity" (FC) in the hippocampal neural network. FC refers to the tendency of two or more neurons to fire together, which can be assessed by examining cross-correlation between two spike trains (Perkel et al. 1967). In this study, FC was defined as the degree of synchronous firing between two neurons within a 50-msec time window (see below). Considering that the hippocampus is a leading model system for the study of memory and synaptic plasticity (Scoville and Milner 1957; Bliss and Lomo 1973; O’Keefe and Nadel 1978) and that LTP induction alters firing patterns of hippocampal neurons in vivo (Deadwyler et al. 1976; Kimura and Pavlides 2000; Martin and Shapiro 2000; Dragoi et al. 2003), we hypothesized that LTP induction would accompany detectable changes in FC in the hippocampal neural network. To test this, we performed experiments on hippocampal slices. In addition to being the most widely used preparations in studying synaptic plasticity, hippocampal slices contain only local circuitry so that the interpretation of experimental results is straightforward. Recordings were made in CA3 region because CA3 neurons are connected to each other by recurrent collateral/commissural fibers and because synaptic plasticity in this pathway obeys the Hebbian learning rule (Zalutsky and Nicoll 1990). Because of these features, the CA3 network has been proposed as the site of associative memory storage (for examples, see Marr 1971; Levy 1989; McNaughton and Nadel 1990; Treves and Rolls 1994).
The experimental protocol was approved by the Ethics Review Committee for Animal Experimentation of Ajou University. Hippocampal slices were prepared from adult male Sprague-Dawley rats (150–200 g) according to a published protocol (Yun et al. 2000). Only one slice was used from each animal. On the basis of the results from a pilot study where we varied the concentration of magnesium, the present study used 0.3 mM of magnesium to produce maximal spontaneous discharge rates of neurons without evoking a seizure-like activity. Two to six tetrodes were placed in the cell body layer under visual guidance to record unit signals (Fig. 1A). Other than tetrode placement in the slice, tetrode fabrication and recording procedures were identical as in previous in vivo recording studies (Jung et al. 1994; Song et al. 2005). A stimulating electrode, constructed of twisted strands of stainless steel wires (113 µm outer diameter), was placed in the stratum radiatum to stimulate associational/commissural fibers (Fig. 1A; Amaral and Witter 1995). Evoked field excitatory postsynaptic potentials (EPSPs) were recorded via one channel of a tetrode (filtered at band pass 1–3000 Hz). Stimulation pulses (0.1 msec) were delivered every 60 sec to obtain baseline responses (15–20 min) that were about 50% of the maximum evoked responses. Field potential responses were recorded as previously described (Yun et al. 2000). LTP was induced by applying theta burst stimulation (TBS) (Larson et al. 1986; Yun et al. 2000). Three episodes of TBS were applied with 10-sec intervals. Magnitude of LTP was assessed by measuring the percent increase of the initial slope of field EPSPs recorded during the 15- to 30-min time period following TBS over baseline. D-2-amino-5-phosphonopentanoic acid (D-AP5; Tocris Cookson, Buckhurst Hill, UK) was delivered to the perfusion medium 5 min before TBS application for 10 min using a 22-gauge needle syringe that was driven by a syringe driver.
|
; Fig. 1B). Only those clusters that were clearly separable from each other and from background noise throughout the recording session were included in the analysis. FC between two neurons was defined as the following:
|
| (1) |
), assuming that each spike train is a rate-modulated, independent 
process. Briefly, for each spike train during the baseline or post-TBS period, spike density function was generated by applying a Gaussian kernel (
t = 50 msec; t = 25 msec when assessing FC on the basis of synchronous firing within 25-msec time window) to each spike, and inhomogeneous processes were generated with the shape parameter (k) between 1 and 20 on the basis of the spike density function. The surrogate spike train with the most similar interspike interval histogram to the experimental interspike interval histogram (in the least-squares sense) was then selected. Spikes occurring in two neurons within 1 msec from each other were not detected with the same tetrode, so they were not included in the calculation of FC. All data are expressed as mean ± SEM. A P value < 0.05 was used as the criterion for a significant statistical difference. Variations in FC were compared across three different experimental groups. They were LTP induction (TBS application in the absence of AP5), AP5 + TBS (TBS application in the presence of 10 µM D-AP5), and no TBS (low frequency test stimulation every 60 sec) groups. Only those slices in which multiple single unit activities were recorded in a stable manner throughout the recording period (15 min baseline and 30 min following LTP induction) were included in the analysis. Field potential responses were significantly enhanced in the LTP induction group (n = 5 slices, 71.3 ± 4.9% enhancement over baseline, paired t-test, t4 = –14.615, P < 0.000; Fig. 2A), but not in the AP5 + TBS (n = 11 slices, 2.5 ± 2.1% enhancement, paired t-test, t10 = –1.16, P > 0.05; Fig. 2B) or no TBS (n = 12 slices, 2.3 ± 1.4% enhancement, paired t-test, t11 = –1.68, P > 0.05; Fig. 2C) groups.
|
rate|; i.e., variation in firing rate) did not vary significantly across the three experimental groups (LTP: 0.30 ± 0.07, AP5 + TBS: 0.26 ± 0.03, no TBS: 0.34 ± 0.06 Hz; Kruskal-Wallis non-parametric test, 
22 = 2.025, P > 0.05).
To assess the magnitude of change in FC (FC), we compared FC during the baseline period (15 min) to that during the 15- to 30-min time period after TBS. FC was enhanced in some neuron pairs, decreased in other neuron pairs, and did not change in the rest. Figure 3A shows three example cross-correlation histograms (CCHs) obtained from LTP-induced slices. They show enhanced, reduced, and unchanged FC after TBS. On average, TBS neither enhanced nor reduced FC in each experimental group. The averaged values of FC before and after TBS were similar in the LTP induction (0.037 ± 0.032 and –0.005 ± 0.032, respectively, paired t-test, t171 = 1.099, P > 0.05), AP5 + TBS (0.067 ± 0.020 and 0.022 ± 0.018, respectively, paired t-test, t168 = 1.886, P > 0.05) as well as no TBS group (0.028 ± 0.036 and 0.080 ± 0.026, respectively, paired t-test, t81 = –1.368, P > 0.05). The average values of FC were close to zero because the numbers of neuron pairs with positive versus negative FC were similar (LTP: 84 and 88; TBS + AP5: 76 and 93; no TBS: 40 and 42). The absolute values of FC (|FC|) were also similar before and after TBS application in the LTP induction (0.330 ± 0.020 and 0.317 ± 0.021, respectively, Wilcoxon signed rank test, z = –0.431, P > 0.05), AP5 + TBS (0.180 ± 0.016 and 0.162 ± 0.013, respectively, Wilcoxon signed rank test, = –0.015, P > 0.05) and no TBS group (0.229 ± 0.026 and 0.181 ± 0.019, respectively, Wilcoxon signed rank test, z = –1.694, P > 0.05), indicating that overall values of FC remained similar after TBS and over time.
|
FC (|FC|) regardless of its direction of change (enhancement or reduction), strikingly large FC changes were observed in LTP-induced slices compared with the other groups (Fig. 3B). Kruskal-Wallis nonparametric test revealed a significant variation in the values of |FC| across the three experimental groups (22 = 24.101, P < 0.000; Fig. 3C). No significant difference was found in |FC| between AP5 + TBS and no TBS groups (Mann-Whitney test, z = –0.154, P > 0.05), indicating that TBS delivery in the presence of AP5 did not induce significant changes in FC. On the other hand, |FC| was significantly greater in the LTP induction group compared with that in AP5 + TBS (Mann-Whitney test, z = –4.542, P < 0.000) or no TBS group (Mann-Whitney test, z = –3.518, P < 0.000).
Essentially the same results were obtained when the analyses were based on synchronous discharge within a 25-msec time window instead of a 50-msec window (Fig. 3C; Variation across groups: Kruskal-Wallis test, 22 = 17.939, P < 0.000; LTP induction vs. AP5 + TBS, Mann-Whitney test, z = –3.789, P < 0.000; LTP induction vs. no TBS, Mann-Whitney test, z = –3.139, P < 0.01). The time window was not reduced further because the number of neuron pairs with enough of the number of expected synchronous firing was too small at smaller time windows. The same conclusion was obtained when we analyzed only putative pyramidal cells (n = 34, 59, and 43 for the LTP induction, AP5 + TBS, and no TBS groups, respectively; Ranck 1973), excluding high firing rate (>2 Hz) neurons (n = 2, 4, and 5 for the LTP-induction, AP5 + TBS, and no TBS groups, respectively; data not shown), and when neuron pairs with FC values larger or smaller than three SD above or below the mean (i.e., outliers) were excluded (n = 1, 3, and 1 pairs for the LTP induction, AP5 + TBS, and no TBS groups, respectively; data not shown).
Finally, we tested the possibility that the difference in |FC| can be accounted for by the combination of factors other than LTP induction. We constructed a generalized linear model (cf. McCullagh and Nelder 1989) with mean firing rate, FC, and |FC| during the baseline period and |rate| as explanatory variables, and |FC| as the dependent variable. Then, a second generalized linear model was constructed with the experimental group (LTP induction, AP5 + TBS, and no TBS) added to the explanatory variables using two dummy variables ( distribution of the dependent variable and the reciprocal link function were employed in both models). A likelihood ratio test indicated a significant improvement of the second model compared with the first (22 = 7.961, P < 0.05), corroborating the conclusion that LTP-induction modified FC among hippocampal neurons.
The present study shows that LTP induction leads to detectable changes in FC among hippocampal neurons. The magnitude of |FC| was similar between AP5 + TBS and no TBS slices, indicating that TBS applied in the presence of AP5 induced little change in FC among hippocampal neurons. On the other hand, LTP-induced slices showed a significantly higher degree of |FC| compared with AP5 + TBS or no TBS slices, indicating that FC changes cannot be accounted for by the delivery of TBS or an unknown temporal factor. We generated surrogate spike trains on the basis of the spike density function of each recorded unit. This method takes into account local variations in firing rate. Hence, it is unlikely that larger |FC| in LTP-induced slices is an epiphenomenon secondary to a change in local firing rate or bursting pattern. As reported previously (Kimura and Pavlides 2000; Martin and Shapiro 2000; Dragoi et al. 2003), the averaged firing rate was similar before and after LTP induction, further indicating that the observed FC is not due to changes in firing rate. The overall FC was also similar before and after LTP induction, which is consistent with the previous report (Dragoi et al. 2003) that LTP induction modifies place-specific firing of hippocampal neurons without disrupting hippocampal network dynamics. Finally, the difference in |FC| across experimental groups persisted even after taking the baseline firing rate, FC, |FC|, and |rate| into consideration. These results show that LTP induction is at least partly responsible for the observed FC changes.
LTP induction did not enhance nor reduce the averaged FC. The absolute values of FC were not different before and after TBS application either. Consistent with these results, we found changes in FC among simultaneously recorded prefrontal cortical neurons in the course of learning, but the changes were bidirectional so that overall FC remains constant (see also Baeg et al. 2007). These results suggest that there exists a normalizing mechanism for the total sum of FC. The underlying neural mechanism for this observation is not clear. As discussed in the previous reports (Martin and Shapiro 2000; Kimura and Pavlides 2000; Dragoi et al. 2003), this could be due to balanced excitation/inhibition. Enhanced excitation among CA3 pyramidal cells would result in increased activation of local inhibitory neurons as well, leading to balanced excitation/inhibition. This will end up with similar average discharge rates and FC among CA3 pyramidal cells. In a similar vein, enhanced synaptic weights on excitatory as well as inhibitory neurons (Buzsaki and Eidelberg 1982; Kairiss et al. 1987; Ouardouz and Lacaille 1995; Grunze et al. 1996; Maccaferri and McBain 1996; Wang and Kelly 2001; Lamsa et al. 2005) may result in balanced enhancement/reduction of FC among neurons. Finally, it is also possible that synaptic weight enhancement in some synapses is accompanied by concurrent synaptic weight decrease in other synapses so that overall synaptic weights remain the same (Lynch et al. 1977; Royer and Pare 2003). The above possibilities are not mutually exclusive and they may work together to stably maintain overall FC at a fixed level. A number of neural network models use such normalization algorithms to maintain overall synaptic weights constant (Arbib 1995).
|
Acknowledgments |
|---|
|
FOOTNOTES |
|---|
3 Department of Psychiatry, University of Pittsburgh, Pittsburgh, PA 15213, USA;
4 Department of Neuroscience, University of Pittsburgh, Pittsburgh, PA 15260, USA.
E-mail min{at}ajou.ac.kr; fax +82-31-219-4401.
Article is online at http://www.learnmem.org/cgi/doi/10.1101/lm.466307
|
References |
|---|
|
TOPABSTRACTReferences |
|---|
Arbib, M.A. 1995. The handbook of brain theory and neural networks. MIT Press, Cambridge, MA.
Baeg, E.H., Kim, Y.B., Kim, J., Kim, J.-W., Kim, J.J., and Jung, M.W. 2007. Learning-induced enduring changes in functional connectivity among prefrontal cortical neurons. J. Neurosci. 27: 909–918.
Baker, S.N. and Lemon, R.N. 2000. Precise spatiotemporal repeating patterns in monkey primary and supplementary motor areas occur at chance levels. J. Neurophysiol. 84: 1770–1780.
Bliss, T.V. and Lomo, T. 1973. Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path. J. Physiol. 232: 331–356.
Buzsaki, G. and Eidelberg, E. 1982. Direct afferent excitation and long-term potentiation of hippocampal interneurons. J. Neurophysiol. 48: 597–607.
Collingridge, G.L., Kehl, S.J., and McLennan, H. 1983. Excitatory amino acids in synaptic transmission in the Schaffer collateral-commissural pathway of the rat hippocampus. J. Physiol. 334: 33–46.
Deadwyler, S.A., Gribkoff, V., Cotman, C., and Lynch, G. 1976. Long lasting changes in the spontaneous activity of hippocampal neurons following stimulation of the entorhinal cortex. Brain Res. Bull. 1: 1–7.[CrossRef][Medline]
Dragoi, G., Harris, K.D., and Buzsaki, G. 2003. Place representation within hippocampal networks is modified by long-term potentiation. Neuron 39: 843–853.[CrossRef][Medline]
Grunze, H.C., Rainnie, D.G., Hasselmo, M.E., Barkai, E., Hearn, E.F., McCarley, R.W., and Greene, R.W. 1996. NMDA-dependent modulation of CA1 local circuit inhibition. J. Neurosci. 16: 2034–2043.
Hebb, D.O. 1949. The organization of behavior. Wiley, New York.
Hinton, G.E. and Anderson, J.A. 1981. Parallel models of associative memory. Erlbaum, Hillsdale, NJ.
Hopfield, J.J. 1984. Neurons with graded response have collective computational properties like those of two-state neurons. Proc. Natl. Acad. Sci. 81: 3088–3092.
Jung, M.W., Wiener, S.I., and McNaughton, B.L. 1994. Comparison of spatial firing characteristics of units in dorsal and ventral hippocampus of the rat. J. Neurosci. 14: 7347–7356.[Abstract]
Kairiss, E.W., Abraham, W.C., Bilkey, D.K., and Goddard, G.V. 1987. Field potential evidence for long-term potentiation of feed-forward inhibition in the rat dentate gyrus. Brain Res. 401: 87–94.[CrossRef][Medline]
Kimura, A. and Pavlides, C. 2000. Long-term potentiation/depotentiation are accompanied by complex changes in spontaneous unit activity in the hippocampus. J. Neurophysiol. 84: 1894–1906.
Kohonen, T. 1977. Associative memory: A system theoretical approach. Springer, New York.
Lamsa, K., Heeroman, J.H., and Kullman, D.M. 2005. Hebbian LTP in feed-forward inhibitory interneurons and the temporal fidelity of input discrimination. Nat. Neurosci. 8: 916–924.[Medline]
Larson, J., Wong, D., and Lynch, G. 1986. Patterned stimulation at the theta frequency is optimal for induction of hippocampal long-term potentiation. Brain Res. 368: 347–350.[CrossRef][Medline]
Levy, W.B. 1989. A computational approach to hippocampal function. In Computational models of learning in simple neural systems: The psychology of learning and motivation (eds. R.D. Hawkins and G.H. Bower), pp. 243–305. Academic Press, San Diego. Vol. 23.
Lynch, G.S., Dunwiddie, T., and Gribkoff, V. 1977. Heterosynaptic depression: A postsynaptic correlate of long-term potentiation. Nature 266: 737–739.[CrossRef][Medline]
Maccaferri, G. and McBain, C.J. 1996. Long-term potentiation in distinct subtypes of hippocampal nonpyramidal neurons. J. Neurosci. 16: 5334–5343.
Marr, D. 1971. Simple memory: a theory for archicortex. Phil. Trans. Royal Soc. Lond. Ser. B 262: 23–81.
Martin, S.J. and Morris, R.G. 2002. New life in an old idea: The synaptic plasticity and memory hypothesis revisited. Hippocampus 12: 609–636.[CrossRef][Medline]
Martin, P.D. and Shapiro, M.L. 2000. Disparate effects of long-term potentiation on evoked potentials and single CA1 neurons in the hippocampus of anesthetized rats. Hippocampus 10: 207–212.[CrossRef][Medline]
Martin, S.J., Grimwood, P.D., and Morris, R.G. 2000. Synaptic plasticity and memory: An evaluation of the hypothesis. Annu. Rev. Neurosci. 23: 649–711.[CrossRef][Medline]
McCullagh, P. and Nelder, J.A. 1989. Generalized linear models. Chapman and Hall, London.
McNaughton, B.L. and Morris, R.G.M. 1987. Hippocampal synaptic enhancement and information storage within a distributed memory system. Trends Neurosci. 10: 408–415.[CrossRef]
McNaughton, B.L. and Nadel, L. 1990. Hebb-Marr networks and the neurobiological representation of action in space. In Neuroscience and connectionist theory (eds. M.A. Gluck and D.E. Rumelhart), pp. 1–64. Erlbaum, Hillsdale, NJ.
McNaughton, B.L., Douglas, R.M., and Goddard, G.V. 1978. Synaptic enhancement in fascia dentata: Cooperativity among coactive afferents. Brain Res. 157: 277–293.[CrossRef][Medline]
Nowak, L., Bregestovski, P., Ascher, P., Herbet, A., and Prochiantz, A. 1984. Magnesium gates glutamate-activated channels in mouse central neurones. Nature 307: 462–465.[CrossRef][Medline]
O’Keefe, J. and Nadel, L. 1978. The hippocampus as a cognitive map. Clarendon Press, Oxford, UK.
Ouardouz, M. and Lacaille, J.C. 1995. Mechanisms of selective long-term potentiation of excitatory synapses in stratum oriens/alveus interneurons of rat hippocampal slices. J. Neurophysiol. 73: 810–819.
Perkel, D.H., Gerstein, G.L., and Moore, G.P. 1967. Neuronal spike trains and stochastic point processes. II. Simultaneous spike trains. Biophys. J. 7: 419–440.[Medline]
Ranck Jr., J.B. 1973. Studies on single neurons in dorsal hippocampal formation and septum in unrestrained rats. I. Behavioral correlates and firing repertoires. Exp. Neurol. 41: 461–531.[CrossRef][Medline]
Royer, S. and Pare, D. 2003. Conservation of total synaptic weight through balanced synaptic depression and potentiation. Nature 422: 518–522.[CrossRef][Medline]
Scoville, W.B. and Milner, B. 1957. Loss of recent memory after bilateral hippocampal lesions. J. Neurochem. 20: 11–21.[Medline]
Song, E.Y., Kim, Y.B., Kim, Y.H., and Jung, M.W. 2005. Role of active movement in place-specific firing of hippocampal neurons. Hippocampus 15: 8–17.[CrossRef][Medline]
Treves, A. and Rolls, E.T. 1994. Computational analysis of the role of the hippocampus in memory. Hippocampus 4: 374–391.[CrossRef][Medline]
Wang, J.H. and Kelly, P. 2001. Calcium-calmodulin signalling pathway up-regulates glutamatergic synaptic function in non-pyramidal, fast spiking rat hippocampal CA1 neurons. J. Physiol. 533: 407–422.
Yun, S.H., Cheong, M.Y., Mook-Jung, I., Huh, K., Lee, C.-J., and Jung, M.W. 2000. Cholinergic modulation of synaptic transmission and plasticity in entorhinal cortex and hippocampus of the rat. Neuroscience 97: 671–676.[CrossRef][Medline]
Zalutsky, R.A. and Nicoll, R.A. 1990. Comparison of two forms of long-term potentiation in single hippocampal neurons. Science 248: 1619–1624.
CiteULike 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
