Hippocampal gamma-frequency oscillations: from interneurones to pyramidal cells, and back
1 University Laboratory of Physiology, Oxford University, Parks Road, Oxford OX1 3PT, UK
2 Centre for the Biology of Memory, Norwegian University of Science and Technology, N-7489 Trondheim, Norway
Abstract
GABAergic interneurones are necessary for the emergence of hippocampal gamma-frequency network oscillations, during which they play a key role in the synchronization of pyramidal cell firing. However, it remains to be resolved how distinct interneurone subtypes contribute to gamma-frequency oscillations, in what way the spatiotemporal pattern of interneuronal input affects principal cell activity, and by which mechanisms the interneurones themselves are synchronized. Here we summarize recent evidence from cholinergically induced gamma-frequency network oscillations in vitro, showing that perisomatic-targeting GABAergic interneurones provide prominent rhythmic inhibition in pyramidal cells, and that these interneurones are synchronized by recurrent excitation. We conclude by presenting a minimal integrate-and-fire network model which demonstrates that this excitatory-inhibitory feedback loop is sufficient to explain the generation of intrahippocampal gamma-frequency oscillations.
, http://www.100md.com
Spatio-temporal characteristics of hippocampal interneurone activity
GABAergic interneurones constitute a diverse population of neurones, and several criteria have been used to subdivide them into different classes. Probably the most successful of these classification schemes is based on target selectivity. Two main cellular targets have been identified anatomically and physiologically, principal neurones and interneurones, and distinct classes of interneurone show a preference for either of these neuronal types (Fig. 1). In addition, interneurones that preferentially or exclusively innervate principal neurones synapse onto specific subcellular membrane domains. Thus, three broad classes of interneurones can be distinguished: interneurone-selective interneurones, interneurones that terminate on dendritic membrane domains of principal neurones, and those that target perisomatic regions of principal neurones (Buhl et al. 1994; Halasy et al. 1996; Acsady et al. 1996; Gulyas et al. 1996; Cobb et al. 1997; for review see Freund & Buzsaki, 1996).
, 百拇医药
A–C, light microscopic reconstructions of the dendritic and axonal arborizations of synaptically coupled neurones in the hippocampal CA1. A, reciprocally connected basket (grey) and pyramidal (black) cell pair, showing that the basket cell preferentially synapses onto the pyramidal cell at perisomatic sites. Reproduced with permission from Buhl et al. (1994); Nature Publishing Group (http://www.nature.com). B, Schaffer-associated interneurone and two postsynaptic pyramidal cells, showing preferential innervation of pyramidal cell dendrites in the stratum radiatum. Reproduced from Vida et al. (1998). C, a synaptically connected basket-to-basket cell pair, reproduced from Cobb et al. (1997) with permission from Elsevier. D, schematic diagram of two principal mechanisms by which local hippocampal network oscillations could be generated: synaptic recurrent feedback loops between excitatory principal cells (E) and inhibitory interneurones (I), and entrainment by a mutually connected interneuronal network. (Arrows mark the axon initial segments; oriens, stratum oriens; pyr, stratum pyramidale; rad, stratum radiatum; lm, stratum lacunosum-moleculare.)
, http://www.100md.com
The spatial segregation of interneurone terminals on pyramidal neurones suggests functional specialization among GABAergic interneurones. Perisomatic-targeting interneurones, including basket cells and axo-axonic cells, are apposite to precisely control the timing of action potential generation in principal neurones (Fig. 1A). The inhibitory inputs from dendritic-targeting cells are more suited to shunt coaligned excitatory inputs and control dendritic electrogenesis (Miles et al. 1996) (Fig. 1B). Therefore, while excitatory interactions between principal cells are likely to mediate information processing and storage in the hippocampus, the pattern of activity across distinct interneuronal subtypes may be critical in determining the computational functions performed by the network (Paulsen & Moser, 1998).
, 百拇医药
The inputs from distinct interneuronal subtypes may be compartmentalized not only in space, but also in time. Hippocampal neurones often display periods of synchronized rhythmic activity, which are evidenced by oscillations in the extracellularly recorded field potential. These network oscillations have been categorized into frequency bands, with correlated mental and/or behavioural activities, and it has been shown that different interneurone subtypes display stereotypical temporal firing patterns in relation to theta-frequency (4–12 Hz) and ripple (100–200 Hz) oscillations (Klausberger et al. 2003, 2004). Gamma-frequency oscillations (30–100 Hz) are commonly observed superposed on theta-frequency oscillations in the hippocampus (Bragin et al. 1995; Csicsvari et al. 2003; Buzsaki et al. 2003), but apart from a report on three anatomically identified basket interneurones in CA1 (Penttonen et al. 1998), no systematic study has been performed to characterize the firing properties of GABAergic interneurones during hippocampal gamma oscillations. Such gamma-frequency network oscillations have been implicated in memory processing (Lisman & Idiart, 1995; Jensen & Lisman, 1996), and to assess such hypotheses requires knowledge of the spatiotemporal patterns of inhibition and excitation involved, i.e. what cellular events do gamma-frequency oscillations in the field potential represent These underlying events cannot simply be deduced from the firing patterns of different cell classes, with one important additional factor being the nature of transmission at different synapses. A clear example is the membrane potential and GABAA receptor reversal potential at the postsynaptic membrane, which determine whether fast GABAergic transmission is depolarizing or hyperpolarizing.
, http://www.100md.com
Mechanisms of hippocampal gamma oscillations
The pattern of activity during hippocampal gamma oscillations, and thus the putative functions this oscillation could support, is inextricably linked to the mechanisms by which such activity is synchronized. There are two principal mechanisms by which local hippocampal network oscillations could be generated: (i) entrainment by an interneuronal network, synchronized by mutual GABAergic connections and/or gap junctions (Whittington et al. 1995) (Fig. 1C and D), and (ii) synaptic recurrent feedback loops between principal cells and interneurones (Freeman, 1968) (Fig. 1A and D). In the model of the interneuronal network, the principal cell firing is binned into narrow time windows, without the principal cell activity itself contributing to synchronization. In the recurrent feedback model, activity gradually builds up in the pyramidal cells until recurrent inhibition is triggered, and thus this model might support the sequential activation and disbandment of neuronal assemblies.
, 百拇医药
The spatiotemporal patterns of activity during gamma-frequency network oscillations, and the mechanisms of synchronization, would ideally be explored in vivo (Bragin et al. 1995; Penttonen et al. 1998; Csicsvari et al. 2003; Buzsaki et al. 2003). However, in order to physiologically and pharmacologically dissect both pre- and postsynaptic effects across the network, it is expedient to use a relevant in vitro model to test hypotheses for the basic mechanisms involved. Cholinergically induced fast network oscillations in vitro share many of the features of intrahippocampal gamma oscillations in vivo, including pyramidal neurones firing at low frequencies (< 5 Hz) and phase-locked to the oscillation (Fisahn et al. 1998; Hajos et al. 2004), alternating pairs of current sinks (inward positive charge) and sources (outward positive charge) in the somata/basal dendrites and apical dendrites of pyramidal cells (Shimono et al. 2000; Hajos et al. 2004), and the oscillation being generated in CA3 and propagating to CA1 (Fisahn et al. 1998; for review see Mann & Paulsen, 2005). This review aims to present our current insights into the generation of intrahippocampal gamma oscillations in the CA3 (Fisahn et al. 1998; Csicsvari et al. 2003), which have been gleaned from recent experiments in the in vitro model of cholinergically induced fast oscillations.
, http://www.100md.com
The spatial distribution of extracellular current sinks and sources
While gamma rhythms represent the orchestrated activity of the hippocampal network, the actual oscillations recorded in the field potential predominantly reflect the currents flowing into and out of the principal cells. The hippocampus is laminated, with the somatodendritic axes of principal cells radially aligned, and thus common synaptic/electrogenic events within restricted membrane domains of these cells create relatively large macroscopic extracellular currents. As specific excitatory and inhibitory input pathways respect the laminar boundaries within the hippocampus, analysing these extracellular currents during network oscillations may provide clues as to the underlying cellular events.
, http://www.100md.com
The spatial distribution of macroscopic extracellular currents can be approximated using current source density (CSD) analysis (for review see Mitzdorf, 1985). Such analysis has revealed that both intrahippocampal gamma oscillations in vivo, and cholinergically induced fast network oscillations in vitro, are associated with alternating pairs of current sinks and sources in the somata/basal dendrites and apical dendrites of CA3 pyramidal cells (Shimono et al. 2000; Csicsvari et al. 2003; Hajos et al. 2004) (Fig. 2A). This CSD profile suggests that such fast network oscillations do not involve significant net rhythmic currents in the distal dendrites, which might have resulted from activity in the perforant path or lacunosum-moleculare-projecting interneurones. However, the CSD profile itself is ambiguous – currents flow in circuits, and CSD analysis does not distinguish between active current generators and passive return currents, and thus each of the observed alternating somatodendritic dipoles could result from an active perisomatic current, an active dendritic current, or both.
, 百拇医药
Individual CA3 neurones were recorded during network oscillations induced by 20–25 μM carbachol. A, intracellular current-clamp recordings from pyramidal neurones were obtained from slices mounted on 64-electrode planar arrays. Experiments were performed at room temperature, giving an oscillation frequency of 20 Hz. a, spike-triggered average of the membrane potential, showing an afterhyperpolarization which was sustained for > 400 ms. Action potential clipped at –15 mV. b, the spike-triggered average of the membrane potential at higher resolution reveals fast subthreshold oscillations. c, spike-triggered averages of the field potential across the somatodendritic axis of the pyramidal cell was also calculated, in order to construct a current source density (CSD) profile of the current sinks (red) and sources (blue) surrounding spike initiation. The CSD profile displays clear oscillations, and in comparison with b, it can be observed that perisomatic current sinks are followed by pyramidal cell depolarization, and perisomatic sources are followed by membrane hyperpolarization. B, schematic diagram of the anatomical connectivity between pyramidal cells (black) and selected interneuronal subtypes within the hippocampal CA3 network. C, characterization of the firing patterns of pyramidal cells and distinct interneurone subtypes during fast network oscillations was achieved using visually guided unit recordings, with subsequent intracellular labelling for anatomical identification. The field potential was recorded with a single extracellular electrode in the stratum pyramidale, with experiments performed at 30 ± 1°C, giving an oscillation frequency of 35 Hz. The spike times of neuronal firing are given relative to the negative peak of the field oscillation in the stratum pyramidale, with the spike probability for each neuronal subtype in any given 1 ms period represented by a gaussian function (data from Hajos et al. 2004). The gaussian functions are centred on the average spike time mode, and have a width equal to the average standard deviation of the spike time for each cell, thus giving a measure of spike jitter. Colour coded as in B. (pyr, stratum pyramidale; rad, stratum radiatum; AAC, axo-axonic cell; BC, basket cell; IS, interneurone-selective cell; O-LM, oriens-lacunosum-moleculare cell; RC, radiatum cell.)
, http://www.100md.com
Spike patterns in the network
A valuable step in deciphering the active cellular events generating the CSD profile is to examine the associated spike patterns of pyramidal cells and different interneurone subtypes within the network (Csicsvari et al. 2003). Such characterization has recently been performed for cholinergically induced fast network oscillations in vitro, using visually guided unit recordings from individual CA3 neurones, which were subsequently labelled intracellularly for anatomical identification (Hajos et al. 2004; for a similar study on kainate-induced gamma-frequency oscillations see Gloveli et al. 2005) (Fig. 2B and C). Pyramidal cells were found to fire at low frequencies (mean 3 Hz, range 1–7 Hz), with spikes clustered around the negative peak of the local field oscillation. Phase-coupled interneurones fired after the pyramidal cells, with the broad distinction emerging between perisomatic-targeting interneurones that showed highly reliable and strongly phase-coupled firing, and dendritic-targeting interneurones that fired at lower frequencies and with weaker phase coupling (Hajos et al. 2004). An exception was represented by the relatively high discharge rate of interneurones with dendritic arbours in the stratum oriens and axonal projections in the stratum lacunosum-moleculare (O-LM cells), which fired on a significant proportion of the oscillatory cycles shortly after the pyramidal cells. This is intriguing in relation to the CSD profile, since current sinks/sources were not detected in their termination zone on the distal dendritic tufts of pyramidal neurones (Hajos et al. 2004). It is possible that the slicing procedure could substantially crop the arborization of these long axonal projections (although the existence of this projection is a prerequisite of anatomical identification), but it is also possible that the postsynaptic currents associated with activation of those synapses are relatively small (Maccaferri et al. 2000). Thus, elucidating the relative spike timing of distinct neuronal types cannot alone completely disambiguate the cellular currents involved. However, what emerges strongly from the above study is the important role of perisomatic inhibition. The robust firing of perisomatic-targeting interneurones precedes the development of the perisomatic source, which is associated with an inhibition of pyramidal cell firing. Indeed, intracellular recordings from pyramidal neurones in vivo confirm that the firing of perisomatic-targeting interneurones is followed by rhythmic hyperpolarizations of the cell body (Penttonen et al. 1998) (Fig. 2A), and thus it appears that such fast network oscillations at least involve active perisomatic sources. However, the precise roles of dendritic-targeting interneurones and recurrent excitation in controlling pyramidal cell activity, and in generating the sink/source profile, have yet to be resolved.
, 百拇医药
Mechanisms of interneuronal synchronization at gamma frequencies
Given that perisomatic-targeting interneurones appear to play a prominent role in synchronizing the firing of CA3 pyramidal neurones within fast intrahippocampal network oscillations, the next question is how these interneurones themselves are synchronized. During both in vivo and cholinergically induced in vitro gamma-frequency oscillations, interneurones fire with a brief delay after pyramidal cells, which is consistent with mono-synaptic excitation (Fisahn et al. 1998; Csicsvari et al. 2003; Hajos et al. 2004). This suggests that gamma rhythm generation is supported by recurrent feedback inhibition, rather than an independent interneuronal network. Indeed, cholinergically induced fast network oscillations in vitro are blocked by AMPA receptor antagonists (Fisahn et al. 1998; Palhalmi et al. 2004). However, it is not immediately clear whether the sparse firing across the pyramidal cell population displays sufficient synchrony and strength to adequately explain the phase-locked and reliable firing of perisomatic interneurones (Traub et al. 2000).
, 百拇医药
While fast synaptic excitation may be necessary to drive perisomatic-targeting interneurones, the synchronization of these cells could be sharpened by interneurone–interneurone interactions. During cholinergically induced in vitro oscillations, interneurone-selective interneurones also fire with a delay after pyramidal cells consistent with mono-synaptic excitation (Hajos et al. 2004), and could thus promptly curtail the firing of perisomatic-targeting interneurones. Furthermore, hippocampal basket cells are coupled both synaptically and by gap junctions (Cobb et al. 1997; Fukuda & Kosaka, 2000; Meyer et al. 2002), and as a population could thus auto-tune their firing (Tamas et al. 2000; Traub et al. 2001; for review see Whittington & Traub, 2003). Although the contribution of mutual synaptic inhibition to the generation of gamma oscillations is not yet established, it appears that interneurone coupling via gap junctions is not necessary for oscillogenesis. Mice deficient in connexin-36, the predominant known neuronal gap junction protein in the brain (Rash et al. 2000; Venance et al. 2000), continue to display gamma oscillations both in vitro (Hormuzdi et al. 2001) and in vivo (Buhl et al. 2003), although the power of the gamma oscillations in these knock-outs is substantially reduced.
, 百拇医药
Another intriguing aspect of a feedback model of gamma generation is how, in each cycle, spiking in a small proportion of the pyramidal cell population can reliably discharge perisomatic-targeting interneurones (Traub et al. 2000). The anatomical connectivity between CA3 pyramidal cells and perisomatic-targeting interneurones has not been completely characterized, but there is clearly a high degree of both convergence (> 1000 x) and divergence (> 100 x) (Sik et al. 1993; Li et al. 1994; Gulyas et al. 1999). Furthermore, although these connections are often mediated by only a single synapse (Sik et al. 1993; Gulyas et al. 1993), such synapses show large-amplitude excitatory postsynaptic potentials and low failure rates (Buhl et al. 1997; Ali et al. 1998). Given this functional connectivity, it might not be surprising that even single pyramidal cells are capable of discharging postsynaptic interneurones (Miles, 1990; Csicsvari et al. 1998; Marshall et al. 2002), although the synchronized firing of several converging presynaptic pyramidal cells might normally be required (see Buhl et al. 1997; Ali et al. 1998).
, 百拇医药
Despite these mechanisms that might assist a feedback mechanism of synchronization, network modelling studies have suggested that the recorded firing of pyramidal neurones is too low to account for the phasic excitatory drive to interneurones (Traub et al. 2000, 2003a,b; for review see Whittington & Traub, 2003). It has been hypothesized that this problem can be solved by the random occurrence of ectopic spikes in pyramidal axons (1 Hz or more), which are coupled by axo-axonic gap junctions (Traub et al. 2000, 2003a,b; Schmitz et al. 2001; see also Kullmann, 2001). In this model, action potentials in the axon can percolate through the axonal plexus and thus amplify the effective postsynaptic drive to interneurones, while antidromic spike invasion is prevented by perisomatic inhibition. As gamma oscillations persist in connexin-36 knockout mice (Hormuzdi et al. 2001; Buhl et al. 2003), it has been proposed that such axo-axonic coupling is mediated by a novel gap junction protein (Traub et al. 2003a,b), possibly a pannexin (Bruzzone et al. 2003). Unfortunately, this hypothesis is difficult to test experimentally, as pharmacological gap junction blockers are notoriously non-specific (for review see Connors & Long, 2004), and thus may inhibit network activity independently of any action on gap junction coupling (Rouach et al. 2003; Fischer, 2004; Vessey et al. 2004). Therefore, it appears difficult at this time to unreservedly synthesize the evidence from both experimental and modelling studies, in order to resolve the mechanism of gamma-frequency synchronization.
, 百拇医药
A model sufficient for gamma rhythm generation
Given the recent experimental data that has provided further insight into the patterns of network activity during cholinergically induced fast network oscillations in vitro (see above), it seems appropriate to incorporate these findings into a network model to re-examine the minimal requirements for gamma-frequency oscillations. In order to test whether pyramidal cells firing at low frequencies could be sufficient to enable gamma-frequency synchronization via an inhibitory synaptic feedback loop, a network architecture was constructed containing 400 pyramidal cells and 40 interneurones, each modelled as a single-compartment leaky integrate-and-fire neurone. Connectivity was generated at random such that each pyramidal cell received inputs from 40 other pyramidal cells and 20 inhibitory interneurones, whilst each inhibitory interneurone received inputs from 200 pyramidal cells. The membrane potential of the pyramidal cells was initialized at random between –80 and –70 mV, and each pyramidal cell received a tonic depolarizing current that was the minimum to reach threshold (+0.175 nA; threshold, –50 mV). To model the long-lasting afterhyperpolarization observed in pyramidal cells (see Fig. 2A), which presumably contributes to their low firing rate, the membrane potential of pyramidal cells was reset to –80 ± 5 mV following each spike (full network details will be published separately, and can be obtained from the authors). Given these parameters, the network initially displayed clusters of gamma-frequency activity (Fig. 3A and B), but rapidly settled into a continuous gamma frequency oscillation, with a low firing rate in pyramidal neurones (< 5 Hz) and a high firing rate in interneurones (> 30 Hz) (Fig. 3C and D). Therefore, in this minimal model, excitation from subpopulations of pyramidal neurones appears to be sufficient to phasically drive interneurones, and thus synchronize the entire network through recurrent feedback inhibition.
, http://www.100md.com
A network model of leaky integrate-and-fire neurones was constructed, comprising 400 excitatory pyramidal cells (pyr) and 40 inhibitory interneurones (int). Synaptic conductances were modelled as an exponentially decaying function with a finite increment in conductance following each presynaptic spike. The connectivity in the network was generated at random, with 50% connectivity between pyramidal cells and interneurones and 10% connectivity between pyramidal cells. The membrane potential of each pyramidal cell was initialized at random between –80 and –70 mV, and was reset to –80 ± 5 mV following each spike. Each pyramidal cell received a tonic input current sufficient to reach threshold. A, raster plot of the spike times of pyramidal cells and interneurones, binned every 1 ms over a 20 s trial. Box marks period expanded in C. B, inter-spike interval histograms for pyramidal cells (black) and interneurones (grey). Inhibitory cells fire on almost every oscillatory cycle, whereas pyramidal cells fire at lower rates (< 5 Hz). C, the final second of the raster plot in A. D, spike time histograms for pyramidal cells (black) and interneurones (grey) binned every 1 ms over final second of the trial, showing that only a subpopulation of pyramidal cells spike on each cycle, and that these cells fire prior to interneurones.
, http://www.100md.com
This model was not designed to replicate the full behaviour of the CA3 hippocampal network, but reproduces some essential experimentally determined features of cholinergically induced fast network oscillations, such as low firing rate of pyramidal neurones, higher firing rate of GABAergic interneurones and the necessity of both AMPA receptor-mediated excitation and fast GABAA receptor-mediated inhibition. Similar behaviour of an integrate-and-fire recurrent network model, with faster kinetics of excitation than of inhibition and a high excitation/inhibition ratio, has been reported earlier (Brunel & Wang, 2003), and similar results have been obtained in conductance-based models (Kopell & LeMasson, 1994). It is likely that the inclusion of functional diversity between different interneurone subtypes, in addition to both chemical and electrical synapses between members of the same subtype, would modify the network pattern in important ways.
, 百拇医药
Conclusions
Converging evidence from in vivo recordings in anaesthetized and freely moving rats (Penttonen et al. 1998; Csicsvari et al. 2003), cholinergically induced fast network oscillations in vitro (Fisahn et al. 1998; Hajos et al. 2004) (Fig. 2), and models of leaky integrate-and-fire neurones (Brunel & Wang, 2003) (Fig. 3), suggests that synaptic recurrent inhibitory feedback loops between CA3 pyramidal cells and perisomatic-targeting interneurones are both necessary and sufficient to explain intrahippocampal gamma-frequency oscillations. This does not rule out additional functions of other synaptic and non-synaptic mechanisms in oscillatory activity. The main implication of these studies is that gamma-frequency oscillations serve to synchronize the output of pyramidal neurones, and enable their sequential activation. This leaves the potential for dendritic-targeting interneurones to perform distinct functions within the network.
, 百拇医药
Footnotes
This report was presented at The Journal of Physiology Symposium in honour of the late Eberhard H. Buhl on Structure/Function Correlates in Neurons and Networks, Leeds, UK, 10 September 2004. It was commissioned by the Editorial Board and reflects the views of the authors.
References
Acsady L, Gorcs TJ & Freund TF (1996). Different populations of vasoactive intestinal polypeptide-immunoreactive interneurons are specialized to control pyramidal cells or interneurons in the hippocampus. Neuroscience 73, 317–334.
, http://www.100md.com
Ali AB, Deuchars J, Pawelzik H & Thomson AM (1998). CA1 pyramidal to basket and bistratified cell EPSPs: dual intracellular recordings in rat hippocampal slices. J Physiol 507, 201–217.
Bragin A, Jando G, Nadasdy Z, Hetke J, Wise K & Buzsaki G (1995). Gamma (40–100 Hz) oscillation in the hippocampus of the behaving rat. J Neurosci 15, 47–60.
Brunel N & Wang XJ (2003). What determines the frequency of fast network oscillations with irregular neural discharges I. Synaptic dynamics and excitation-inhibition balance. J Neurophysiol 90, 415–430.
, http://www.100md.com
Bruzzone R, Hormuzdi SG, Barbe MT, Herb A & Monyer H (2003). Pannexins, a family of gap junction proteins expressed in brain. Proc Natl Acad Sci U S A 100, 13644–13649.
Buhl EH, Halasy K & Somogyi P (1994). Diverse sources of hippocampal unitary inhibitory postsynaptic potentials and the number of synaptic release sites. Nature 368, 823–828.
Buhl DL, Harris KD, Hormuzdi SG, Monyer H & Buzsaki G (2003). Selective impairment of hippocampal gamma oscillations in connexin-36 knock-out mouse in vivo. J Neurosci 23, 1013–1018.
, 百拇医药
Buhl EH, Tamas G, Szilagyi T, Stricker C, Paulsen O & Somogyi P (1997). Effect, number and location of synapses made by single pyramidal cells onto aspiny interneurones of cat visual cortex. J Physiol 500, 689–713.
Buzsaki G, Buhl DL, Harris KD, Csicsvari J, Czeh B & Morozov A (2003). Hippocampal network patterns of activity in the mouse. Neuroscience 116, 201–211.
Cobb SR, Halasy K, Vida I, Nyiri G, Tamas G, Buhl EH & Somogyi P (1997). Synaptic effects of identified interneurons innervating both interneurons and pyramidal cells in the rat hippocampus. Neuroscience 79, 629–648.
, http://www.100md.com
Connors BW & Long MA (2004). Electrical synapses in the mammalian brain. Annu Rev Neurosci 27, 393–418.
Csicsvari J, Hirase H, Czurko A & Buzsaki G (1998). Reliability and state dependence of pyramidal cell-interneuron synapses in the hippocampus: an ensemble approach in the behaving rat. Neuron 21, 179–189.
Csicsvari J, Jamieson B, Wise KD & Buzsaki G (2003). Mechanisms of gamma oscillations in the hippocampus of the behaving rat. Neuron 37, 311–322.
, 百拇医药
Fisahn A, Pike FG, Buhl EH & Paulsen O (1998). Cholinergic induction of network oscillations at 40 Hz in the hippocampus in vitro. Nature 394, 186–189.
Fischer Y (2004). The hippocampal intrinsic network oscillator. J Physiol 554, 156–174.
Freeman WJ (1968). Relations between unit activity and evoked potentials in prepyriform cortex of cats. J Neurophysiol 31, 337–348.
Freund TF & Buzsaki G (1996). Interneurons of the hippocampus. Hippocampus 6, 347–470.
, http://www.100md.com
Fukuda T & Kosaka T (2000). Gap junctions linking the dendritic network of GABAergic interneurons in the hippocampus. J Neurosci 20, 1519–1528.
Gloveli T, Dugladze T, Saha S, Monyer H, Heinemann U, Traub RD, Whittington MA & Buhl EH (2005). Differential involvement of oriens/pyramidale interneurons in hippocampal network oscillations in vitro. J Physiol 562, 131–147.
Gulyas AI, Hajos N & Freund TF (1996). Interneurons containing calretinin are specialized to control other interneurons in the rat hippocampus. J Neurosci 16, 3397–3411.
, 百拇医药
Gulyas AI, Megias M, Emri Z & Freund TF (1999). Total number and ratio of excitatory and inhibitory synapses converging onto single interneurons of different types in the CA1 area of the rat hippocampus. J Neurosci 19, 10082–10097.
Gulyas AI, Miles R, Sik A, Toth K, Tamamaki N & Freund TF (1993). Hippocampal pyramidal cells excite inhibitory neurons through a single release site. Nature 366, 683–687.
Hajos N, Palhalmi J, Mann EO, Nemeth B, Paulsen O & Freund TF (2004). Spike timing of distinct types of GABAergic interneuron during hippocampal gamma oscillations in vitro. J Neurosci 24, 9127–9137.
, http://www.100md.com
Halasy K, Buhl EH, Lorinczi Z, Tamas G & Somogyi P (1996). Synaptic target selectivity and input of GABAergic basket and bistratified interneurons in the CA1 area of the rat hippocampus. Hippocampus 6, 306–329.
Hormuzdi SG, Pais I, LeBeau FE, Towers SK, Rozov A, Buhl EH, Whittington MA & Monyer H (2001). Impaired electrical signaling disrupts gamma frequency oscillations in connexin 36-deficient mice. Neuron 31, 487–495.
, http://www.100md.com Jensen O & Lisman JE (1996). Theta/gamma networks with slow NMDA channels learn sequences and encode episodic memory: role of NMDA channels in recall. Learn Mem 3, 264–278.
Klausberger T, Magill PJ, Marton LF, Roberts JD, Cobden PM, Buzsaki G & Somogyi P (2003). Brain-state- and cell-type-specific firing of hippocampal interneurons in vivo. Nature 421, 844–848.
Klausberger T, Marton LF, Baude A, Roberts JD, Magill PJ & Somogyi P (2004). Spike timing of dendrite-targeting bistratified cells during hippocampal network oscillations in vivo. Nat Neurosci 7, 41–47.
, 百拇医药
Kopell N & LeMasson G (1994). Rhythmogenesis, amplitude modulation, and multiplexing in a cortical architecture. Proc Natl Acad Sci U S A 91, 10586–10590.
Kullmann DM (2001). Presynaptic kainate receptors in the hippocampus: slowly emerging from obscurity. Neuron 32, 561–564.
Li XG, Somogyi P, Ylinen A & Buzsaki G (1994). The hippocampal CA3 network: an in vivo intracellular labeling study. J Comp Neurol 339, 181–208.
, 百拇医药
Lisman JE & Idiart MA (1995). Storage of 7 +/- 2 short-term memories in oscillatory subcycles. Science 267, 1512–1515.
Maccaferri G, Roberts JD, Szucs P, Cottingham CA & Somogyi P (2000). Cell surface domain specific postsynaptic currents evoked by identified GABAergic neurones in rat hippocampus in vitro. J Physiol 524, 91–116.
Mann EO & Paulsen O (2005). Mechanisms underlying gamma (‘40 Hz’) oscillations in the hippocampus – a mini-review. Prog Biophys Mol Biol 87, 67–76.
, http://www.100md.com
Marshall L, Henze DA, Hirase H, Leinekugel X, Dragoi G & Buzsaki G (2002). Hippocampal pyramidal cell-interneuron spike transmission is frequency dependent and responsible for place modulation of interneuron discharge. J Neurosci 22, RC197.
Meyer AH, Katona I, Blatow M, Rozov A & Monyer H (2002). In vivo labeling of parvalbumin-positive interneurons and analysis of electrical coupling in identified neurons. J Neurosci 22, 7055–7064.
, http://www.100md.com
Miles R (1990). Synaptic excitation of inhibitory cells by single CA3 hippocampal pyramidal cells of the guinea-pig in vitro. J Physiol 428, 61–77.
Miles R, Toth K, Gulyas AI, Hajos N & Freund TF (1996). Differences between somatic and dendritic inhibition in the hippocampus. Neuron 16, 815–823.
Mitzdorf U (1985). Current source-density method and application in cat cerebral cortex: investigation of evoked potentials and EEG phenomena. Physiol Rev 65, 37–100.
, 百拇医药
Palhalmi J, Paulsen O, Freund TF & Hajos N (2004). Distinct properties of carbachol- and DHPG-induced network oscillations in hippocampal slices. Neuropharmacology 47, 381–389.
Paulsen O & Moser EI (1998). A model of hippocampal memory encoding and retrieval: GABAergic control of synaptic plasticity. Trends Neurosci 21, 273–278.
Penttonen M, Kamondi A, Acsady L & Buzsaki G (1998). Gamma frequency oscillation in the hippocampus of the rat: intracellular analysis in vivo. Eur J Neurosci 10, 718–728.
, 百拇医药
Rash JE, Staines WA, Yasumura T, Patel D, Furman CS, Stelmack GL & Nagy JI (2000). Immunogold evidence that neuronal gap junctions in adult rat brain and spinal cord contain connexin-36 but not connexin-32 or connexin-43. Proc Natl Acad Sci U S A 97, 7573–7578.
Rouach N, Segal M, Koulakoff A, Giaume C & Avignone E (2003). Carbenoxolone blockade of neuronal network activity in culture is not mediated by an action on gap junctions. J Physiol 553, 729–745.
, http://www.100md.com
Schmitz D, Schuchmann S, Fisahn A, Draguhn A, Buhl EH, Petrasch-Parwez E, Dermietzel R, Heinemann U & Traub RD (2001). Axo-axonal coupling. a novel mechanism for ultrafast neuronal communication. Neuron 31, 831–840.
Shimono K, Brucher F, Granger R, Lynch G & Taketani M (2000). Origins and distribution of cholinergically induced beta rhythms in hippocampal slices. J Neurosci 20, 8462–8473.
Sik A, Tamamaki N & Freund TF (1993). Complete axon arborization of a single CA3 pyramidal cell in the rat hippocampus, and its relationship with postsynaptic parvalbumin-containing interneurons. Eur J Neurosci 5, 1719–1728.
, http://www.100md.com
Tamas G, Buhl EH, Lorincz A & Somogyi P (2000). Proximally targeted GABAergic synapses and gap junctions synchronize cortical interneurons. Nat Neurosci 3, 366–371.
Traub RD, Bibbig A, Fisahn A, LeBeau FE, Whittington MA & Buhl EH (2000). A model of gamma-frequency network oscillations induced in the rat CA3 region by carbachol in vitro. Eur J Neurosci 12, 4093–4106.
Traub RD, Cunningham MO, Gloveli T, LeBeau FE, Bibbig A, Buhl EH & Whittington MA (2003a). GABA-enhanced collective behavior in neuronal axons underlies persistent gamma-frequency oscillations. Proc Natl Acad Sci U S A 100, 11047–11052.
, http://www.100md.com
Traub RD, Kopell N, Bibbig A, Buhl EH, LeBeau FE & Whittington MA (2001). Gap junctions between interneuron dendrites can enhance synchrony of gamma oscillations in distributed networks. J Neurosci 21, 9478–9486.
Traub RD, Pais I, Bibbig A, LeBeau FE, Buhl EH, Hormuzdi SG, Monyer H & Whittington MA (2003b). Contrasting roles of axonal (pyramidal cell) and dendritic (interneuron) electrical coupling in the generation of neuronal network oscillations. Proc Natl Acad Sci U S A 100, 1370–1374.
, 百拇医药
Venance L, Rozov A, Blatow M, Burnashev N, Feldmeyer D & Monyer H (2000). Connexin expression in electrically coupled postnatal rat brain neurons. Proc Natl Acad Sci U S A 97, 10260–10265.
Vessey JP, Lalonde MR, Mizan HA, Welch NC, Kelly ME & Barnes S (2004). Carbenoxolone inhibition of voltage-gated Ca channels and synaptic transmission in the retina. J Neurophysiol 92, 1252–1256.
Vida I, Halasy K, Szinyei C, Somogyi P & Buhl EH (1998). Unitary IPSPs evoked by interneurons at the stratum radiatum-stratum lacunosum-moleculare border in the CA1 area of the rat hippocampus in vitro. J Physiol 506, 755–773.
, 百拇医药
Whittington MA & Traub RD (2003). Interneuron diversity series: inhibitory interneurons and network oscillations in vitro. Trends Neurosci 26, 676–682.
Whittington MA, Traub RD & Jefferys JG (1995). Synchronized oscillations in interneuron networks driven by metabotropic glutamate receptor activation. Nature 373, 612–615., http://www.100md.com(Edward O. Mann, Catrin A.)
2 Centre for the Biology of Memory, Norwegian University of Science and Technology, N-7489 Trondheim, Norway
Abstract
GABAergic interneurones are necessary for the emergence of hippocampal gamma-frequency network oscillations, during which they play a key role in the synchronization of pyramidal cell firing. However, it remains to be resolved how distinct interneurone subtypes contribute to gamma-frequency oscillations, in what way the spatiotemporal pattern of interneuronal input affects principal cell activity, and by which mechanisms the interneurones themselves are synchronized. Here we summarize recent evidence from cholinergically induced gamma-frequency network oscillations in vitro, showing that perisomatic-targeting GABAergic interneurones provide prominent rhythmic inhibition in pyramidal cells, and that these interneurones are synchronized by recurrent excitation. We conclude by presenting a minimal integrate-and-fire network model which demonstrates that this excitatory-inhibitory feedback loop is sufficient to explain the generation of intrahippocampal gamma-frequency oscillations.
, http://www.100md.com
Spatio-temporal characteristics of hippocampal interneurone activity
GABAergic interneurones constitute a diverse population of neurones, and several criteria have been used to subdivide them into different classes. Probably the most successful of these classification schemes is based on target selectivity. Two main cellular targets have been identified anatomically and physiologically, principal neurones and interneurones, and distinct classes of interneurone show a preference for either of these neuronal types (Fig. 1). In addition, interneurones that preferentially or exclusively innervate principal neurones synapse onto specific subcellular membrane domains. Thus, three broad classes of interneurones can be distinguished: interneurone-selective interneurones, interneurones that terminate on dendritic membrane domains of principal neurones, and those that target perisomatic regions of principal neurones (Buhl et al. 1994; Halasy et al. 1996; Acsady et al. 1996; Gulyas et al. 1996; Cobb et al. 1997; for review see Freund & Buzsaki, 1996).
, 百拇医药
A–C, light microscopic reconstructions of the dendritic and axonal arborizations of synaptically coupled neurones in the hippocampal CA1. A, reciprocally connected basket (grey) and pyramidal (black) cell pair, showing that the basket cell preferentially synapses onto the pyramidal cell at perisomatic sites. Reproduced with permission from Buhl et al. (1994); Nature Publishing Group (http://www.nature.com). B, Schaffer-associated interneurone and two postsynaptic pyramidal cells, showing preferential innervation of pyramidal cell dendrites in the stratum radiatum. Reproduced from Vida et al. (1998). C, a synaptically connected basket-to-basket cell pair, reproduced from Cobb et al. (1997) with permission from Elsevier. D, schematic diagram of two principal mechanisms by which local hippocampal network oscillations could be generated: synaptic recurrent feedback loops between excitatory principal cells (E) and inhibitory interneurones (I), and entrainment by a mutually connected interneuronal network. (Arrows mark the axon initial segments; oriens, stratum oriens; pyr, stratum pyramidale; rad, stratum radiatum; lm, stratum lacunosum-moleculare.)
, http://www.100md.com
The spatial segregation of interneurone terminals on pyramidal neurones suggests functional specialization among GABAergic interneurones. Perisomatic-targeting interneurones, including basket cells and axo-axonic cells, are apposite to precisely control the timing of action potential generation in principal neurones (Fig. 1A). The inhibitory inputs from dendritic-targeting cells are more suited to shunt coaligned excitatory inputs and control dendritic electrogenesis (Miles et al. 1996) (Fig. 1B). Therefore, while excitatory interactions between principal cells are likely to mediate information processing and storage in the hippocampus, the pattern of activity across distinct interneuronal subtypes may be critical in determining the computational functions performed by the network (Paulsen & Moser, 1998).
, 百拇医药
The inputs from distinct interneuronal subtypes may be compartmentalized not only in space, but also in time. Hippocampal neurones often display periods of synchronized rhythmic activity, which are evidenced by oscillations in the extracellularly recorded field potential. These network oscillations have been categorized into frequency bands, with correlated mental and/or behavioural activities, and it has been shown that different interneurone subtypes display stereotypical temporal firing patterns in relation to theta-frequency (4–12 Hz) and ripple (100–200 Hz) oscillations (Klausberger et al. 2003, 2004). Gamma-frequency oscillations (30–100 Hz) are commonly observed superposed on theta-frequency oscillations in the hippocampus (Bragin et al. 1995; Csicsvari et al. 2003; Buzsaki et al. 2003), but apart from a report on three anatomically identified basket interneurones in CA1 (Penttonen et al. 1998), no systematic study has been performed to characterize the firing properties of GABAergic interneurones during hippocampal gamma oscillations. Such gamma-frequency network oscillations have been implicated in memory processing (Lisman & Idiart, 1995; Jensen & Lisman, 1996), and to assess such hypotheses requires knowledge of the spatiotemporal patterns of inhibition and excitation involved, i.e. what cellular events do gamma-frequency oscillations in the field potential represent These underlying events cannot simply be deduced from the firing patterns of different cell classes, with one important additional factor being the nature of transmission at different synapses. A clear example is the membrane potential and GABAA receptor reversal potential at the postsynaptic membrane, which determine whether fast GABAergic transmission is depolarizing or hyperpolarizing.
, http://www.100md.com
Mechanisms of hippocampal gamma oscillations
The pattern of activity during hippocampal gamma oscillations, and thus the putative functions this oscillation could support, is inextricably linked to the mechanisms by which such activity is synchronized. There are two principal mechanisms by which local hippocampal network oscillations could be generated: (i) entrainment by an interneuronal network, synchronized by mutual GABAergic connections and/or gap junctions (Whittington et al. 1995) (Fig. 1C and D), and (ii) synaptic recurrent feedback loops between principal cells and interneurones (Freeman, 1968) (Fig. 1A and D). In the model of the interneuronal network, the principal cell firing is binned into narrow time windows, without the principal cell activity itself contributing to synchronization. In the recurrent feedback model, activity gradually builds up in the pyramidal cells until recurrent inhibition is triggered, and thus this model might support the sequential activation and disbandment of neuronal assemblies.
, 百拇医药
The spatiotemporal patterns of activity during gamma-frequency network oscillations, and the mechanisms of synchronization, would ideally be explored in vivo (Bragin et al. 1995; Penttonen et al. 1998; Csicsvari et al. 2003; Buzsaki et al. 2003). However, in order to physiologically and pharmacologically dissect both pre- and postsynaptic effects across the network, it is expedient to use a relevant in vitro model to test hypotheses for the basic mechanisms involved. Cholinergically induced fast network oscillations in vitro share many of the features of intrahippocampal gamma oscillations in vivo, including pyramidal neurones firing at low frequencies (< 5 Hz) and phase-locked to the oscillation (Fisahn et al. 1998; Hajos et al. 2004), alternating pairs of current sinks (inward positive charge) and sources (outward positive charge) in the somata/basal dendrites and apical dendrites of pyramidal cells (Shimono et al. 2000; Hajos et al. 2004), and the oscillation being generated in CA3 and propagating to CA1 (Fisahn et al. 1998; for review see Mann & Paulsen, 2005). This review aims to present our current insights into the generation of intrahippocampal gamma oscillations in the CA3 (Fisahn et al. 1998; Csicsvari et al. 2003), which have been gleaned from recent experiments in the in vitro model of cholinergically induced fast oscillations.
, http://www.100md.com
The spatial distribution of extracellular current sinks and sources
While gamma rhythms represent the orchestrated activity of the hippocampal network, the actual oscillations recorded in the field potential predominantly reflect the currents flowing into and out of the principal cells. The hippocampus is laminated, with the somatodendritic axes of principal cells radially aligned, and thus common synaptic/electrogenic events within restricted membrane domains of these cells create relatively large macroscopic extracellular currents. As specific excitatory and inhibitory input pathways respect the laminar boundaries within the hippocampus, analysing these extracellular currents during network oscillations may provide clues as to the underlying cellular events.
, http://www.100md.com
The spatial distribution of macroscopic extracellular currents can be approximated using current source density (CSD) analysis (for review see Mitzdorf, 1985). Such analysis has revealed that both intrahippocampal gamma oscillations in vivo, and cholinergically induced fast network oscillations in vitro, are associated with alternating pairs of current sinks and sources in the somata/basal dendrites and apical dendrites of CA3 pyramidal cells (Shimono et al. 2000; Csicsvari et al. 2003; Hajos et al. 2004) (Fig. 2A). This CSD profile suggests that such fast network oscillations do not involve significant net rhythmic currents in the distal dendrites, which might have resulted from activity in the perforant path or lacunosum-moleculare-projecting interneurones. However, the CSD profile itself is ambiguous – currents flow in circuits, and CSD analysis does not distinguish between active current generators and passive return currents, and thus each of the observed alternating somatodendritic dipoles could result from an active perisomatic current, an active dendritic current, or both.
, 百拇医药
Individual CA3 neurones were recorded during network oscillations induced by 20–25 μM carbachol. A, intracellular current-clamp recordings from pyramidal neurones were obtained from slices mounted on 64-electrode planar arrays. Experiments were performed at room temperature, giving an oscillation frequency of 20 Hz. a, spike-triggered average of the membrane potential, showing an afterhyperpolarization which was sustained for > 400 ms. Action potential clipped at –15 mV. b, the spike-triggered average of the membrane potential at higher resolution reveals fast subthreshold oscillations. c, spike-triggered averages of the field potential across the somatodendritic axis of the pyramidal cell was also calculated, in order to construct a current source density (CSD) profile of the current sinks (red) and sources (blue) surrounding spike initiation. The CSD profile displays clear oscillations, and in comparison with b, it can be observed that perisomatic current sinks are followed by pyramidal cell depolarization, and perisomatic sources are followed by membrane hyperpolarization. B, schematic diagram of the anatomical connectivity between pyramidal cells (black) and selected interneuronal subtypes within the hippocampal CA3 network. C, characterization of the firing patterns of pyramidal cells and distinct interneurone subtypes during fast network oscillations was achieved using visually guided unit recordings, with subsequent intracellular labelling for anatomical identification. The field potential was recorded with a single extracellular electrode in the stratum pyramidale, with experiments performed at 30 ± 1°C, giving an oscillation frequency of 35 Hz. The spike times of neuronal firing are given relative to the negative peak of the field oscillation in the stratum pyramidale, with the spike probability for each neuronal subtype in any given 1 ms period represented by a gaussian function (data from Hajos et al. 2004). The gaussian functions are centred on the average spike time mode, and have a width equal to the average standard deviation of the spike time for each cell, thus giving a measure of spike jitter. Colour coded as in B. (pyr, stratum pyramidale; rad, stratum radiatum; AAC, axo-axonic cell; BC, basket cell; IS, interneurone-selective cell; O-LM, oriens-lacunosum-moleculare cell; RC, radiatum cell.)
, http://www.100md.com
Spike patterns in the network
A valuable step in deciphering the active cellular events generating the CSD profile is to examine the associated spike patterns of pyramidal cells and different interneurone subtypes within the network (Csicsvari et al. 2003). Such characterization has recently been performed for cholinergically induced fast network oscillations in vitro, using visually guided unit recordings from individual CA3 neurones, which were subsequently labelled intracellularly for anatomical identification (Hajos et al. 2004; for a similar study on kainate-induced gamma-frequency oscillations see Gloveli et al. 2005) (Fig. 2B and C). Pyramidal cells were found to fire at low frequencies (mean 3 Hz, range 1–7 Hz), with spikes clustered around the negative peak of the local field oscillation. Phase-coupled interneurones fired after the pyramidal cells, with the broad distinction emerging between perisomatic-targeting interneurones that showed highly reliable and strongly phase-coupled firing, and dendritic-targeting interneurones that fired at lower frequencies and with weaker phase coupling (Hajos et al. 2004). An exception was represented by the relatively high discharge rate of interneurones with dendritic arbours in the stratum oriens and axonal projections in the stratum lacunosum-moleculare (O-LM cells), which fired on a significant proportion of the oscillatory cycles shortly after the pyramidal cells. This is intriguing in relation to the CSD profile, since current sinks/sources were not detected in their termination zone on the distal dendritic tufts of pyramidal neurones (Hajos et al. 2004). It is possible that the slicing procedure could substantially crop the arborization of these long axonal projections (although the existence of this projection is a prerequisite of anatomical identification), but it is also possible that the postsynaptic currents associated with activation of those synapses are relatively small (Maccaferri et al. 2000). Thus, elucidating the relative spike timing of distinct neuronal types cannot alone completely disambiguate the cellular currents involved. However, what emerges strongly from the above study is the important role of perisomatic inhibition. The robust firing of perisomatic-targeting interneurones precedes the development of the perisomatic source, which is associated with an inhibition of pyramidal cell firing. Indeed, intracellular recordings from pyramidal neurones in vivo confirm that the firing of perisomatic-targeting interneurones is followed by rhythmic hyperpolarizations of the cell body (Penttonen et al. 1998) (Fig. 2A), and thus it appears that such fast network oscillations at least involve active perisomatic sources. However, the precise roles of dendritic-targeting interneurones and recurrent excitation in controlling pyramidal cell activity, and in generating the sink/source profile, have yet to be resolved.
, 百拇医药
Mechanisms of interneuronal synchronization at gamma frequencies
Given that perisomatic-targeting interneurones appear to play a prominent role in synchronizing the firing of CA3 pyramidal neurones within fast intrahippocampal network oscillations, the next question is how these interneurones themselves are synchronized. During both in vivo and cholinergically induced in vitro gamma-frequency oscillations, interneurones fire with a brief delay after pyramidal cells, which is consistent with mono-synaptic excitation (Fisahn et al. 1998; Csicsvari et al. 2003; Hajos et al. 2004). This suggests that gamma rhythm generation is supported by recurrent feedback inhibition, rather than an independent interneuronal network. Indeed, cholinergically induced fast network oscillations in vitro are blocked by AMPA receptor antagonists (Fisahn et al. 1998; Palhalmi et al. 2004). However, it is not immediately clear whether the sparse firing across the pyramidal cell population displays sufficient synchrony and strength to adequately explain the phase-locked and reliable firing of perisomatic interneurones (Traub et al. 2000).
, 百拇医药
While fast synaptic excitation may be necessary to drive perisomatic-targeting interneurones, the synchronization of these cells could be sharpened by interneurone–interneurone interactions. During cholinergically induced in vitro oscillations, interneurone-selective interneurones also fire with a delay after pyramidal cells consistent with mono-synaptic excitation (Hajos et al. 2004), and could thus promptly curtail the firing of perisomatic-targeting interneurones. Furthermore, hippocampal basket cells are coupled both synaptically and by gap junctions (Cobb et al. 1997; Fukuda & Kosaka, 2000; Meyer et al. 2002), and as a population could thus auto-tune their firing (Tamas et al. 2000; Traub et al. 2001; for review see Whittington & Traub, 2003). Although the contribution of mutual synaptic inhibition to the generation of gamma oscillations is not yet established, it appears that interneurone coupling via gap junctions is not necessary for oscillogenesis. Mice deficient in connexin-36, the predominant known neuronal gap junction protein in the brain (Rash et al. 2000; Venance et al. 2000), continue to display gamma oscillations both in vitro (Hormuzdi et al. 2001) and in vivo (Buhl et al. 2003), although the power of the gamma oscillations in these knock-outs is substantially reduced.
, 百拇医药
Another intriguing aspect of a feedback model of gamma generation is how, in each cycle, spiking in a small proportion of the pyramidal cell population can reliably discharge perisomatic-targeting interneurones (Traub et al. 2000). The anatomical connectivity between CA3 pyramidal cells and perisomatic-targeting interneurones has not been completely characterized, but there is clearly a high degree of both convergence (> 1000 x) and divergence (> 100 x) (Sik et al. 1993; Li et al. 1994; Gulyas et al. 1999). Furthermore, although these connections are often mediated by only a single synapse (Sik et al. 1993; Gulyas et al. 1993), such synapses show large-amplitude excitatory postsynaptic potentials and low failure rates (Buhl et al. 1997; Ali et al. 1998). Given this functional connectivity, it might not be surprising that even single pyramidal cells are capable of discharging postsynaptic interneurones (Miles, 1990; Csicsvari et al. 1998; Marshall et al. 2002), although the synchronized firing of several converging presynaptic pyramidal cells might normally be required (see Buhl et al. 1997; Ali et al. 1998).
, 百拇医药
Despite these mechanisms that might assist a feedback mechanism of synchronization, network modelling studies have suggested that the recorded firing of pyramidal neurones is too low to account for the phasic excitatory drive to interneurones (Traub et al. 2000, 2003a,b; for review see Whittington & Traub, 2003). It has been hypothesized that this problem can be solved by the random occurrence of ectopic spikes in pyramidal axons (1 Hz or more), which are coupled by axo-axonic gap junctions (Traub et al. 2000, 2003a,b; Schmitz et al. 2001; see also Kullmann, 2001). In this model, action potentials in the axon can percolate through the axonal plexus and thus amplify the effective postsynaptic drive to interneurones, while antidromic spike invasion is prevented by perisomatic inhibition. As gamma oscillations persist in connexin-36 knockout mice (Hormuzdi et al. 2001; Buhl et al. 2003), it has been proposed that such axo-axonic coupling is mediated by a novel gap junction protein (Traub et al. 2003a,b), possibly a pannexin (Bruzzone et al. 2003). Unfortunately, this hypothesis is difficult to test experimentally, as pharmacological gap junction blockers are notoriously non-specific (for review see Connors & Long, 2004), and thus may inhibit network activity independently of any action on gap junction coupling (Rouach et al. 2003; Fischer, 2004; Vessey et al. 2004). Therefore, it appears difficult at this time to unreservedly synthesize the evidence from both experimental and modelling studies, in order to resolve the mechanism of gamma-frequency synchronization.
, 百拇医药
A model sufficient for gamma rhythm generation
Given the recent experimental data that has provided further insight into the patterns of network activity during cholinergically induced fast network oscillations in vitro (see above), it seems appropriate to incorporate these findings into a network model to re-examine the minimal requirements for gamma-frequency oscillations. In order to test whether pyramidal cells firing at low frequencies could be sufficient to enable gamma-frequency synchronization via an inhibitory synaptic feedback loop, a network architecture was constructed containing 400 pyramidal cells and 40 interneurones, each modelled as a single-compartment leaky integrate-and-fire neurone. Connectivity was generated at random such that each pyramidal cell received inputs from 40 other pyramidal cells and 20 inhibitory interneurones, whilst each inhibitory interneurone received inputs from 200 pyramidal cells. The membrane potential of the pyramidal cells was initialized at random between –80 and –70 mV, and each pyramidal cell received a tonic depolarizing current that was the minimum to reach threshold (+0.175 nA; threshold, –50 mV). To model the long-lasting afterhyperpolarization observed in pyramidal cells (see Fig. 2A), which presumably contributes to their low firing rate, the membrane potential of pyramidal cells was reset to –80 ± 5 mV following each spike (full network details will be published separately, and can be obtained from the authors). Given these parameters, the network initially displayed clusters of gamma-frequency activity (Fig. 3A and B), but rapidly settled into a continuous gamma frequency oscillation, with a low firing rate in pyramidal neurones (< 5 Hz) and a high firing rate in interneurones (> 30 Hz) (Fig. 3C and D). Therefore, in this minimal model, excitation from subpopulations of pyramidal neurones appears to be sufficient to phasically drive interneurones, and thus synchronize the entire network through recurrent feedback inhibition.
, http://www.100md.com
A network model of leaky integrate-and-fire neurones was constructed, comprising 400 excitatory pyramidal cells (pyr) and 40 inhibitory interneurones (int). Synaptic conductances were modelled as an exponentially decaying function with a finite increment in conductance following each presynaptic spike. The connectivity in the network was generated at random, with 50% connectivity between pyramidal cells and interneurones and 10% connectivity between pyramidal cells. The membrane potential of each pyramidal cell was initialized at random between –80 and –70 mV, and was reset to –80 ± 5 mV following each spike. Each pyramidal cell received a tonic input current sufficient to reach threshold. A, raster plot of the spike times of pyramidal cells and interneurones, binned every 1 ms over a 20 s trial. Box marks period expanded in C. B, inter-spike interval histograms for pyramidal cells (black) and interneurones (grey). Inhibitory cells fire on almost every oscillatory cycle, whereas pyramidal cells fire at lower rates (< 5 Hz). C, the final second of the raster plot in A. D, spike time histograms for pyramidal cells (black) and interneurones (grey) binned every 1 ms over final second of the trial, showing that only a subpopulation of pyramidal cells spike on each cycle, and that these cells fire prior to interneurones.
, http://www.100md.com
This model was not designed to replicate the full behaviour of the CA3 hippocampal network, but reproduces some essential experimentally determined features of cholinergically induced fast network oscillations, such as low firing rate of pyramidal neurones, higher firing rate of GABAergic interneurones and the necessity of both AMPA receptor-mediated excitation and fast GABAA receptor-mediated inhibition. Similar behaviour of an integrate-and-fire recurrent network model, with faster kinetics of excitation than of inhibition and a high excitation/inhibition ratio, has been reported earlier (Brunel & Wang, 2003), and similar results have been obtained in conductance-based models (Kopell & LeMasson, 1994). It is likely that the inclusion of functional diversity between different interneurone subtypes, in addition to both chemical and electrical synapses between members of the same subtype, would modify the network pattern in important ways.
, 百拇医药
Conclusions
Converging evidence from in vivo recordings in anaesthetized and freely moving rats (Penttonen et al. 1998; Csicsvari et al. 2003), cholinergically induced fast network oscillations in vitro (Fisahn et al. 1998; Hajos et al. 2004) (Fig. 2), and models of leaky integrate-and-fire neurones (Brunel & Wang, 2003) (Fig. 3), suggests that synaptic recurrent inhibitory feedback loops between CA3 pyramidal cells and perisomatic-targeting interneurones are both necessary and sufficient to explain intrahippocampal gamma-frequency oscillations. This does not rule out additional functions of other synaptic and non-synaptic mechanisms in oscillatory activity. The main implication of these studies is that gamma-frequency oscillations serve to synchronize the output of pyramidal neurones, and enable their sequential activation. This leaves the potential for dendritic-targeting interneurones to perform distinct functions within the network.
, 百拇医药
Footnotes
This report was presented at The Journal of Physiology Symposium in honour of the late Eberhard H. Buhl on Structure/Function Correlates in Neurons and Networks, Leeds, UK, 10 September 2004. It was commissioned by the Editorial Board and reflects the views of the authors.
References
Acsady L, Gorcs TJ & Freund TF (1996). Different populations of vasoactive intestinal polypeptide-immunoreactive interneurons are specialized to control pyramidal cells or interneurons in the hippocampus. Neuroscience 73, 317–334.
, http://www.100md.com
Ali AB, Deuchars J, Pawelzik H & Thomson AM (1998). CA1 pyramidal to basket and bistratified cell EPSPs: dual intracellular recordings in rat hippocampal slices. J Physiol 507, 201–217.
Bragin A, Jando G, Nadasdy Z, Hetke J, Wise K & Buzsaki G (1995). Gamma (40–100 Hz) oscillation in the hippocampus of the behaving rat. J Neurosci 15, 47–60.
Brunel N & Wang XJ (2003). What determines the frequency of fast network oscillations with irregular neural discharges I. Synaptic dynamics and excitation-inhibition balance. J Neurophysiol 90, 415–430.
, http://www.100md.com
Bruzzone R, Hormuzdi SG, Barbe MT, Herb A & Monyer H (2003). Pannexins, a family of gap junction proteins expressed in brain. Proc Natl Acad Sci U S A 100, 13644–13649.
Buhl EH, Halasy K & Somogyi P (1994). Diverse sources of hippocampal unitary inhibitory postsynaptic potentials and the number of synaptic release sites. Nature 368, 823–828.
Buhl DL, Harris KD, Hormuzdi SG, Monyer H & Buzsaki G (2003). Selective impairment of hippocampal gamma oscillations in connexin-36 knock-out mouse in vivo. J Neurosci 23, 1013–1018.
, 百拇医药
Buhl EH, Tamas G, Szilagyi T, Stricker C, Paulsen O & Somogyi P (1997). Effect, number and location of synapses made by single pyramidal cells onto aspiny interneurones of cat visual cortex. J Physiol 500, 689–713.
Buzsaki G, Buhl DL, Harris KD, Csicsvari J, Czeh B & Morozov A (2003). Hippocampal network patterns of activity in the mouse. Neuroscience 116, 201–211.
Cobb SR, Halasy K, Vida I, Nyiri G, Tamas G, Buhl EH & Somogyi P (1997). Synaptic effects of identified interneurons innervating both interneurons and pyramidal cells in the rat hippocampus. Neuroscience 79, 629–648.
, http://www.100md.com
Connors BW & Long MA (2004). Electrical synapses in the mammalian brain. Annu Rev Neurosci 27, 393–418.
Csicsvari J, Hirase H, Czurko A & Buzsaki G (1998). Reliability and state dependence of pyramidal cell-interneuron synapses in the hippocampus: an ensemble approach in the behaving rat. Neuron 21, 179–189.
Csicsvari J, Jamieson B, Wise KD & Buzsaki G (2003). Mechanisms of gamma oscillations in the hippocampus of the behaving rat. Neuron 37, 311–322.
, 百拇医药
Fisahn A, Pike FG, Buhl EH & Paulsen O (1998). Cholinergic induction of network oscillations at 40 Hz in the hippocampus in vitro. Nature 394, 186–189.
Fischer Y (2004). The hippocampal intrinsic network oscillator. J Physiol 554, 156–174.
Freeman WJ (1968). Relations between unit activity and evoked potentials in prepyriform cortex of cats. J Neurophysiol 31, 337–348.
Freund TF & Buzsaki G (1996). Interneurons of the hippocampus. Hippocampus 6, 347–470.
, http://www.100md.com
Fukuda T & Kosaka T (2000). Gap junctions linking the dendritic network of GABAergic interneurons in the hippocampus. J Neurosci 20, 1519–1528.
Gloveli T, Dugladze T, Saha S, Monyer H, Heinemann U, Traub RD, Whittington MA & Buhl EH (2005). Differential involvement of oriens/pyramidale interneurons in hippocampal network oscillations in vitro. J Physiol 562, 131–147.
Gulyas AI, Hajos N & Freund TF (1996). Interneurons containing calretinin are specialized to control other interneurons in the rat hippocampus. J Neurosci 16, 3397–3411.
, 百拇医药
Gulyas AI, Megias M, Emri Z & Freund TF (1999). Total number and ratio of excitatory and inhibitory synapses converging onto single interneurons of different types in the CA1 area of the rat hippocampus. J Neurosci 19, 10082–10097.
Gulyas AI, Miles R, Sik A, Toth K, Tamamaki N & Freund TF (1993). Hippocampal pyramidal cells excite inhibitory neurons through a single release site. Nature 366, 683–687.
Hajos N, Palhalmi J, Mann EO, Nemeth B, Paulsen O & Freund TF (2004). Spike timing of distinct types of GABAergic interneuron during hippocampal gamma oscillations in vitro. J Neurosci 24, 9127–9137.
, http://www.100md.com
Halasy K, Buhl EH, Lorinczi Z, Tamas G & Somogyi P (1996). Synaptic target selectivity and input of GABAergic basket and bistratified interneurons in the CA1 area of the rat hippocampus. Hippocampus 6, 306–329.
Hormuzdi SG, Pais I, LeBeau FE, Towers SK, Rozov A, Buhl EH, Whittington MA & Monyer H (2001). Impaired electrical signaling disrupts gamma frequency oscillations in connexin 36-deficient mice. Neuron 31, 487–495.
, http://www.100md.com Jensen O & Lisman JE (1996). Theta/gamma networks with slow NMDA channels learn sequences and encode episodic memory: role of NMDA channels in recall. Learn Mem 3, 264–278.
Klausberger T, Magill PJ, Marton LF, Roberts JD, Cobden PM, Buzsaki G & Somogyi P (2003). Brain-state- and cell-type-specific firing of hippocampal interneurons in vivo. Nature 421, 844–848.
Klausberger T, Marton LF, Baude A, Roberts JD, Magill PJ & Somogyi P (2004). Spike timing of dendrite-targeting bistratified cells during hippocampal network oscillations in vivo. Nat Neurosci 7, 41–47.
, 百拇医药
Kopell N & LeMasson G (1994). Rhythmogenesis, amplitude modulation, and multiplexing in a cortical architecture. Proc Natl Acad Sci U S A 91, 10586–10590.
Kullmann DM (2001). Presynaptic kainate receptors in the hippocampus: slowly emerging from obscurity. Neuron 32, 561–564.
Li XG, Somogyi P, Ylinen A & Buzsaki G (1994). The hippocampal CA3 network: an in vivo intracellular labeling study. J Comp Neurol 339, 181–208.
, 百拇医药
Lisman JE & Idiart MA (1995). Storage of 7 +/- 2 short-term memories in oscillatory subcycles. Science 267, 1512–1515.
Maccaferri G, Roberts JD, Szucs P, Cottingham CA & Somogyi P (2000). Cell surface domain specific postsynaptic currents evoked by identified GABAergic neurones in rat hippocampus in vitro. J Physiol 524, 91–116.
Mann EO & Paulsen O (2005). Mechanisms underlying gamma (‘40 Hz’) oscillations in the hippocampus – a mini-review. Prog Biophys Mol Biol 87, 67–76.
, http://www.100md.com
Marshall L, Henze DA, Hirase H, Leinekugel X, Dragoi G & Buzsaki G (2002). Hippocampal pyramidal cell-interneuron spike transmission is frequency dependent and responsible for place modulation of interneuron discharge. J Neurosci 22, RC197.
Meyer AH, Katona I, Blatow M, Rozov A & Monyer H (2002). In vivo labeling of parvalbumin-positive interneurons and analysis of electrical coupling in identified neurons. J Neurosci 22, 7055–7064.
, http://www.100md.com
Miles R (1990). Synaptic excitation of inhibitory cells by single CA3 hippocampal pyramidal cells of the guinea-pig in vitro. J Physiol 428, 61–77.
Miles R, Toth K, Gulyas AI, Hajos N & Freund TF (1996). Differences between somatic and dendritic inhibition in the hippocampus. Neuron 16, 815–823.
Mitzdorf U (1985). Current source-density method and application in cat cerebral cortex: investigation of evoked potentials and EEG phenomena. Physiol Rev 65, 37–100.
, 百拇医药
Palhalmi J, Paulsen O, Freund TF & Hajos N (2004). Distinct properties of carbachol- and DHPG-induced network oscillations in hippocampal slices. Neuropharmacology 47, 381–389.
Paulsen O & Moser EI (1998). A model of hippocampal memory encoding and retrieval: GABAergic control of synaptic plasticity. Trends Neurosci 21, 273–278.
Penttonen M, Kamondi A, Acsady L & Buzsaki G (1998). Gamma frequency oscillation in the hippocampus of the rat: intracellular analysis in vivo. Eur J Neurosci 10, 718–728.
, 百拇医药
Rash JE, Staines WA, Yasumura T, Patel D, Furman CS, Stelmack GL & Nagy JI (2000). Immunogold evidence that neuronal gap junctions in adult rat brain and spinal cord contain connexin-36 but not connexin-32 or connexin-43. Proc Natl Acad Sci U S A 97, 7573–7578.
Rouach N, Segal M, Koulakoff A, Giaume C & Avignone E (2003). Carbenoxolone blockade of neuronal network activity in culture is not mediated by an action on gap junctions. J Physiol 553, 729–745.
, http://www.100md.com
Schmitz D, Schuchmann S, Fisahn A, Draguhn A, Buhl EH, Petrasch-Parwez E, Dermietzel R, Heinemann U & Traub RD (2001). Axo-axonal coupling. a novel mechanism for ultrafast neuronal communication. Neuron 31, 831–840.
Shimono K, Brucher F, Granger R, Lynch G & Taketani M (2000). Origins and distribution of cholinergically induced beta rhythms in hippocampal slices. J Neurosci 20, 8462–8473.
Sik A, Tamamaki N & Freund TF (1993). Complete axon arborization of a single CA3 pyramidal cell in the rat hippocampus, and its relationship with postsynaptic parvalbumin-containing interneurons. Eur J Neurosci 5, 1719–1728.
, http://www.100md.com
Tamas G, Buhl EH, Lorincz A & Somogyi P (2000). Proximally targeted GABAergic synapses and gap junctions synchronize cortical interneurons. Nat Neurosci 3, 366–371.
Traub RD, Bibbig A, Fisahn A, LeBeau FE, Whittington MA & Buhl EH (2000). A model of gamma-frequency network oscillations induced in the rat CA3 region by carbachol in vitro. Eur J Neurosci 12, 4093–4106.
Traub RD, Cunningham MO, Gloveli T, LeBeau FE, Bibbig A, Buhl EH & Whittington MA (2003a). GABA-enhanced collective behavior in neuronal axons underlies persistent gamma-frequency oscillations. Proc Natl Acad Sci U S A 100, 11047–11052.
, http://www.100md.com
Traub RD, Kopell N, Bibbig A, Buhl EH, LeBeau FE & Whittington MA (2001). Gap junctions between interneuron dendrites can enhance synchrony of gamma oscillations in distributed networks. J Neurosci 21, 9478–9486.
Traub RD, Pais I, Bibbig A, LeBeau FE, Buhl EH, Hormuzdi SG, Monyer H & Whittington MA (2003b). Contrasting roles of axonal (pyramidal cell) and dendritic (interneuron) electrical coupling in the generation of neuronal network oscillations. Proc Natl Acad Sci U S A 100, 1370–1374.
, 百拇医药
Venance L, Rozov A, Blatow M, Burnashev N, Feldmeyer D & Monyer H (2000). Connexin expression in electrically coupled postnatal rat brain neurons. Proc Natl Acad Sci U S A 97, 10260–10265.
Vessey JP, Lalonde MR, Mizan HA, Welch NC, Kelly ME & Barnes S (2004). Carbenoxolone inhibition of voltage-gated Ca channels and synaptic transmission in the retina. J Neurophysiol 92, 1252–1256.
Vida I, Halasy K, Szinyei C, Somogyi P & Buhl EH (1998). Unitary IPSPs evoked by interneurons at the stratum radiatum-stratum lacunosum-moleculare border in the CA1 area of the rat hippocampus in vitro. J Physiol 506, 755–773.
, 百拇医药
Whittington MA & Traub RD (2003). Interneuron diversity series: inhibitory interneurons and network oscillations in vitro. Trends Neurosci 26, 676–682.
Whittington MA, Traub RD & Jefferys JG (1995). Synchronized oscillations in interneuron networks driven by metabotropic glutamate receptor activation. Nature 373, 612–615., http://www.100md.com(Edward O. Mann, Catrin A.)