Postsynaptic hyperpolarization increases the strength of AMPA-mediated synaptic transmission at large synapses between mossy fibers and CA3 pyramidal cells

Using ephaptic coupling to estimate the synaptic cleft resistivity of the calyx of Held synapse

At synapses, the pre- and postsynaptic cells get so close that currents entering the cleft do not flow exclusively along its conductance, gcl. A prominent example is found in the calyx of Held synapse in the medial nucleus of the trapezoid body (MNTB), where the presynaptic action potential can be recorded in the postsynaptic cell in the form of a prespike. Here, we developed a theoretical framework for ephaptic coupling via the synaptic cleft, and we tested its predictions using the MNTB prespike recorded in voltage-clamp. The shape of the prespike is predicted to resemble either the first or the second derivative of the inverted presynaptic action potential if cleft currents dissipate either mostly capacitively or resistively, respectively. We found that the resistive dissipation scenario provided a better description of the prespike shape. Its size is predicted to scale with the fourth power of the radius of the synapse, explaining why intracellularly recorded prespikes are uncommon in the central nervous system. We show that presynaptic calcium currents also contribute to the prespike shape. This calcium prespike resembled the first derivative of the inverted calcium current, again as predicted by the resistive dissipation scenario. Using this calcium prespike, we obtained an estimate for gcl of ~1 μS. We demonstrate that, for a circular synapse geometry, such as in conventional boutons or the immature calyx of Held, gcl is scale-invariant and only defined by extracellular resistivity, which was ~75 Ωcm, and by cleft height. During development the calyx of Held develops fenestrations. We show that these fenestrations effectively minimize the cleft potentials generated by the adult action potential, which might otherwise interfere with calcium channel opening. We thus provide a quantitative account of the dissipation of currents by the synaptic cleft, which can be readily extrapolated to conventional, bouton-like synapses.

Spontaneous Infraslow Fluctuations Modulate Hippocampal EPSP-PS Coupling

Extensive trial-to-trial variability is a hallmark of most behavioral, cognitive, and physiological processes. Spontaneous brain activity (SBA), a ubiquitous phenomenon that coordinates levels and patterns of neuronal activity throughout the brain, may contribute to this variability by dynamically altering neuronal excitability. In freely-behaving male rats, we observed extensive variability of the hippocampal evoked response across 28-min recording periods despite maintaining constant stimulation parameters of the medial perforant path. This variability was related to antecedent SBA: increases in low-frequency (0.5–9 Hz) and high-frequency (40.25–100 Hz) band-limited power (BLP) in the 4-s preceding stimulation were associated with decreased slope of the field EPSP (fEPSP) and increased population spike (PS) amplitude. These fluctuations in SBA and evoked response magnitude did not appear stochastic but rather exhibited coordinated activity across infraslow timescales (0.005–0.02 Hz). Specifically, infraslow fluctuations in high- and low-frequency BLP were antiphase with changes in fEPSP slope and in phase with changes in PS amplitude. With these divergent effects on the fEPSP and PS, infraslow SBA ultimately modulates EPSP-PS coupling and thereby enables hippocampal circuitry to generate heterogeneous outputs from identical inputs. Consequently, infraslow SBA appears well suited to dynamically alter sensory selection and information processing and highlights the fundamental role of endogenous neuronal activity for shaping the brain’s response to incoming stimuli.

Ephaptic coupling to endogenous electric field activity: why bother?

There has been a revived interest in the impact of electric fields on neurons and networks. Here, we discuss recent advances in our understanding of how endogenous and externally imposed electric fields impact brain function at different spatial (from synapses to single neurons and neural networks) and temporal scales (from milliseconds to seconds). How such ephaptic effects are mediated and manifested in the brain remains a mystery. We argue that it is both possible (based on available technologies) and worthwhile to vigorously pursue such research as it has significant implications on our understanding of brain processing and for translational neuroscience.

Ephaptic communication in the vertebrate retina

In the vertebrate retina, cones project to the horizontal cells (HCs) and bipolar cells (BCs). The communication between cones and HCs uses both chemical and ephaptic mechanisms. Cones release glutamate in a Ca²⁺-dependent manner, while HCs feed back to cones via an ephaptic mechanism. Hyperpolarization of HCs leads to an increased current through connexin hemichannels located on the tips of HC dendrites invaginating the cone synaptic terminals. Due to the high resistance of the extracellular synaptic space, this current makes the synaptic cleft slightly negative. The result is that the Ca²⁺-channels in the cone presynaptic membrane experience a slightly depolarized membrane potential and therefore more glutamate is released. This ephaptic mechanism forms a very fast and noise free negative feedback pathway. These characteristics are crucial, since the retina has to perform well in demanding conditions such as low light levels. In this mini-review we will discuss the critical components of such an ephaptic mechanism. Furthermore, we will address the question whether such communication appears in other systems as well and indicate some fundamental features to look for when attempting to identify an ephaptic mechanism.

Intracellular tetanization with hyperpolarizing currents potentiates synapses formed by mossy fibers on pyramidal cells in hippocampal field CA3 in rats

Studies on living rat hippocampal slices using point recording in the whole cell configuration addressed the efficiency of the synaptic responses of pyramidal neurons in field CA3 in conditions of minimal stimulation of mossy fibers. Paired-pulse responses were recorded before and after intracellular tetanizing hyperpolarization of pyramidal neurons. In these conditions, potentiation of excitatory synaptic transmission lasting at least 20 min was seen. This phenomenon, termed hyperpolarizing tetanization-induced long-term potentiation, could arise without simultaneous mossy fiber stimulation and showed signs of having a presynaptic origin. Administration of a Ca2+ chelator into pyramidal neurons completely suppressed this potentiation. The results obtained from these experiments suggest that induction of long-term potentiation evoked by hyperpolarizing tetanization was postsynaptic, while its expression appeared to be presynaptic. These results provide evidence of the importance of gamma-rhythm hyperpolarizing oscillations in altering the efficiency of synaptic inputs and the role of its network organization in the mechanisms of cellular plasticity.

Bilateral processing in chemical synapses with electrical 'ephaptic' feedback: a theoretical model

I have developed a detailed biophysical model of the chemical synapse which hosts voltage-dependent presynaptic ion channels and takes into account the capacitance of synaptic membranes. I find that at synapses with a relatively large cleft resistance (e.g., mossy fiber or giant calyx synapse) the rising postsynaptic current could activate, within the synaptic cleft, electrochemical phenomena that induce rapid widening of the presynaptic action potential (AP). This mechanism could boost fast Ca(2+) entry into the terminal thus increasing the probability of subsequent synaptic releases. The predicted difference in the AP waveforms generated inside and outside the synapse can explain the previously unexplained fast capacitance transient recorded in the postsynaptic cell at the giant calyx synapse. I propose therefore the mechanism of positive ephaptic feedback that acts between the postsynaptic and presynaptic cell contributing to the basal synaptic transmission at large central synapses. This mechanism could also explain the supralinear voltage dependence of EPSCs recorded at hyperpolarizing membrane potentials in low extracellular calcium concentration.

Ephaptic feedback in identified synapses in mollusk neurons

The possible existence of intrasynaptic ephaptic feedback in the invertebrate CNS was studied. Intracellular recordings were made of excitatory postsynaptic potentials and currents arising on activation of the recently described monosynaptic connection between identified neurons in the snail CNS. In the presence of ephaptic feedback, tetanization of the postsynaptic neuron with hyperpolarizing impulses should activate presynaptic calcium channels, thus increasing the amplitude of excitatory postsynaptic potential, while sufficiently strong postsynaptic hyperpolarization applied during generation of the excitatory postsynaptic current should induce "supralinear" increases in its amplitude, as has been observed previously in rat hippocampal neurons. The first series of experiments involved delivery of 10 trains of hyperpolarizing postsynaptic impulses (40-50 mV, duration 0.5 sec, frequency 1 Hz, train duration 45 sec); significant changes in the amplitude of excitatory postsynaptic were not seen. In the second series of experiments, changes in the amplitude of the excitatory postsynaptic current were studied during hyperpolarization of the postsynaptic neuron. At a potential of -100 mV, the amplitude of the excitatory postsynaptic current increased significantly more than predicted by its "classical" linear relationship with membrane potential. This "supralinear" increase in the amplitude of the excitatory postsynaptic potential can be explained by the operation of ephaptic feedback and is the first evidence for this phenomenon in CNS synapses of invertebrates.

The involvement of glutamate-gated channels in negative feedback from horizontal cells to cones

Photoreceptors are the light sensitive cells in the retina. They project to horizontal cells and bipolar cells via a glutamatergic feed forward pathway. Horizontal cells are strongly electrically coupled and integrate in that way the input from the photoreceptors. Horizontal cells feedback to cones negatively. The combined signal from the photoreceptors and the horizontal cells is sent to the bipolar cells. The feedback pathway from horizontal cells to cones is thought to form the basis for the center/surround organization of bipolar cells. The nature of the feedback pathway is an issue of intense debate. It was thought for a long time that this feedback pathway was GABAergic, because cones have GABA-receptors and horizontal cells release GABA via a GABA-transporter working in the reversed direction. However, recently we showed in goldfish that horizontal cells feed back to cones via an alternative mechanism. In goldfish, negative feedback from horizontal cells to cones shifts the calcium current of the cone to more negative potentials. This feedback pathway is independent of GABA, since feedback cannot be blocked by either saturating concentrations of PTX, the GABA-transporter blocker SKF89976A, or application of GABA. The mechanism of negative feedback from horizontal cells to cones involves hemichannels located at the tips of the invaginating horizontal cells, just opposite to the calcium channels of the cones. Current flowing through these hemichannels changes the extracellular potential deep in the synaptic cleft and in that way modulates the calcium current of the cones. Such a modulation of the extracellular potential is called ephaptic. If negative feedback from horizontal cells to cones is indeed ephaptic, other channels present in the synapse should also be able to act as a current source, i.e., should also be able to change the output of the cone. We showed that glutamate-gated channels present at the tips of the horizontal cell dendrites can also mediate feedback responses. Surprisingly, although the glutamate-gated conductance of the horizontal cells is eight times the hemichannel conductance, glutamate-gated channels are not the major current source in negative feedback from horizontal cells to cones. In this chapter we present evidence that this is due to the more focal localization of the hemichannels, compared to a diffuse and extrasynaptic localization of the glutamate-gated channels.

Ephaptic interactions within a chemical synapse: hemichannel-mediated ephaptic inhibition in the retina

The two best-known types of cell-cell communication are chemical synapses and electrical synapses, which are formed by gap junctions. A third, less well known, form of communication is ephaptic transmission, in which electric fields generated by a specific neuron alter the excitability of neighboring neurons as a result of their anatomical and electrical proximity. Ephaptic communication can be present in a variety of forms, each with their specific features and functional implications. One of these is ephaptic modulation within a chemical synapse. This type of communication has recently been proposed for the cone-horizontal cell synapse in the vertebrate retina. Evidence indicates that the extracellular potential in the synaptic terminal of photoreceptors is modulated by current flowing through connexin hemichannels at the tips of the horizontal cell dendrites, mediating negative feedback from horizontal cells to cones. This example can be added to the growing list of cases of ephaptic communication in the central nervous system.

Postsynaptic depolarisation enhances transmitter release and causes the appearance of responses at “silent” synapses in rat hippocampus

Recent data indicate that most "silent" synapses in the hippocampus are "presynaptically silent" due to low transmitter release rather than "postsynaptically silent" due to "latent" receptors of alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid type (AMPARs). That synapses bearing only N-methyl-d-aspartate (NMDAR) receptors do exist is suggested by the decreased number of transmission failures during postsynaptic depolarisation and by the presence of NMDA-mediated excitatory postsynaptic currents (EPSCs) in synapses silent at rest. We tested whether these effects could be due to potentiated transmitter release at depolarised postsynaptic potentials rather than removal of Mg(2+) block from NMDARs. Using whole-cell recordings of minimal EPSCs from CA1 and CA3 neurones of hippocampal slices we confirmed decreased incidence of failures at +40 mV as compared with -60 mV. This effect was associated with a gradual increase of EPSC amplitude after switching to +40 mV and with a decrease of paired-pulse facilitation. In initially silent synapses, potentiation of pharmacologically isolated AMPAR-mediated EPSCs was still observed at +40 mV and this persisted after stepping back to -60 mV. All above effects were blocked when the cell was dialysed with the Ca(2+) chelator BAPTA (20 mM). These observations are difficult to reconcile with the "latent AMPAR" hypothesis and suggest an alternative explanation, namely that the reduction in failure rates at positive potentials is due to potentiation of transmitter release following Ca(2+) influx through NMDARs. Our results suggest that silent synapses can be mainly "presynaptically" rather than "postsynaptically silent" and thus increased transmitter release rather than insertion of AMPARs is a major mechanism of early long-term potentiation maintenance.

Combining principal component and spectral analyses with the method of moments in studies of quantal transmission

This chapter considers methods for measurements of postsynaptic responses and simple approaches to the estimation of parameters of quantal release in synapses of the central nervous system of vertebrates. The use of these methods is illustrated by the analysis of single-fibre and "minimal" monosynaptic postsynaptic potentials (PSPs) or currents (PSCs) recorded from neurons of the frog spinal cord and rat hippocampus. First, we briefly discuss traditional methods of the response measurements using peak amplitudes or areas, further focusing on a novel method based on multivariate statistical techniques of the principal component analysis (PCA). This approach provides typically better signal-to-noise ratios and is able to separate two or more response components, which can arise due to activation of more than one presynaptic fibre, axon collaterals, receptor subtypes or spatially separated transmission sites. Second, spectral analysis is introduced as the method of choice to verify whether the amplitude fluctuations of the postsynaptic responses have a quantal nature and to obtain estimations of the "basic" quantal parameters, i.e. the quantal size (Q) and mean quantal content (m), without introducing assumptions on release statistics. Third, we show how the method of moments could be applied in the framework of the Poisson and binomial models to estimate the basic quantal parameters and parameters p and n, which reflect the release probability and maximum number of quanta released (or the number of effective release sites), respectively. Fourth, we show that the analysis of the moments can also be instrumental to reveal non-uniformity of release probabilities and compare how several competing models of neurotransmitter release fit to multiple experimental data sets.

A model synapse that incorporates the properties of short- and long-term synaptic plasticity

We propose a general computer model of a synapse, which incorporates mechanisms responsible for the realization of both short- and long-term synaptic plasticity-the two forms of experimentally observed plasticity that seem to be very significant for the performance of neuronal networks. The model consists of a presynaptic part based on the earlier 'double barrier synapse' model, and a postsynaptic compartment which is connected to the presynaptic terminal via a feedback, the sign and magnitude of which depend on postsynaptic Ca(2+) concentration. The feedback increases or decreases the amount of neurotransmitter which is in a ready for release state. The model adequately reproduced the phenomena of short- and long-term plasticity observed experimentally in hippocampal slices for CA3-CA1 synapses. The proposed model may be used in the investigation of certain real synapses to estimate their physiological parameters, and in the construction of realistic neuronal networks.

Associative mossy fibre LTP induced by pairing presynaptic stimulation with postsynaptic hyperpolarization of CA3 neurons in rat hippocampal slice

Whole cell recordings of excitatory postsynaptic potentials/currents (EPSPs/EPSCs) evoked by minimal stimulation of commissural-associative (CF) and mossy fibre (MF) inputs were performed in CA3 pyramidal neurons. Paired responses (at 50 ms intervals) were recorded before, during and after hyperpolarization of the postsynaptic membrane (20-30 mV for 15-35 min). Membrane hyperpolarization produced a supralinear increase of EPSPs/EPSCs amplitude in MF-inputs. Synaptic responses remained potentiated for the rest of the recording period (up to 40 min) after resetting the membrane potential to control level (221 +/- 60%, n = 15 and 219 +/- 61%, n = 11 for MF-EPSP and MF-EPSC, respectively). We shall refer to this effect as hyperpolarization-induced LTP (HI-LTP). In the absence of afferent stimulation, membrane hyperpolarization was unable to produce HI-LTP. In contrast to MF-EPSPs, the mean amplitude of CF-EPSPs did not increase significantly after hyperpolarization relative to controls (138 +/- 29%, n = 22). HI-LTP was associated with modifications of classical indices of presynaptic release: paired-pulse facilitation, failures rate, coefficient of variation of EPSP amplitudes and quantal content. The induction of HI-LTP was NMDA independent but was dependent on metabotropic glutamate receptors (mGluRs) activation and calcium release from inositol 1,4,5-triphosphate (IP3)-sensitive intracellular stores: it was prevented by mGluR antagonist, intracellular heparin and BAPTA. We conclude that while the induction of HI-LTP was postsynaptic, its expression was presynaptic.

Mossy fibre synaptic NMDA receptors trigger non-Hebbian long-term potentiation at entorhino-CA3 synapses in the rat

Hippocampal CA3 pyramidal cells receive two independent afferents from the enthorinal cortex, i.e. a direct input via the temporoammonic pathway (TA, perforant path) and an indirect input via the mossy fibres (MF) of dentate granule cells. In spite of past suggestions that the TA is assigned an important role in exciting the pyramidal cells, little is known about their physiological properties. By surgically making an incision through the sulcus hippocampi and a small part of the dentate molecular layer, we succeeded in isolating TA-mediated monosynaptic responses in CA3 stratum lacunosum-moleculare. The TA-CA3 synaptic transmission was completely blocked by a combination of D,L-2-amino-5-phosphonopentanoic acid (AP5) and 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), NMDA and non-NMDA receptor antagonists, respectively, and displayed paired-pulse facilitation and NMDA receptor-dependent long-term potentiation, which are all typical of glutamatergic synapses. We next addressed the heterosynaptic interaction between TA-CA3 and MF-CA3 synapses. The TA-CA3 transmission was partially attenuated by single-pulse MF pre-stimulation at inter-pulse intervals of up to 70 ms. However, surprisingly, burst stimulation of the MF alone induced long-lasting facilitation of TA-CA3 synaptic efficacy. This non-Hebbian form of synaptic plasticity was efficiently prevented by local application of AP5 into the MF synapse-rich area. Therefore, MF-activated NMDA receptors are responsible for the heterosynaptic modification of TA-CA3 transmission, and thereby, the history of MF activity may be etched into TA-CA3 synaptic strength. Our findings predict a novel form of spatiotemporal information processing in the hippocampus, i.e. a use-dependent intersynaptic memory transfer.

Extracellular Ca2+ depletion contributes to fast activity-dependent modulation of synaptic transmission in the brain

Synaptic activation is associated with rapid changes in intracellular Ca(2+), while the extracellular Ca(2+) level is generally assumed to be constant. Here, using a novel optical method to measure changes in extracellular Ca(2+) at high spatial and temporal resolution, we find that brief trains of synaptic transmission in hippocampal area CA1 induce transient depletion of extracellular Ca(2+). We show that this depletion, which depends on postsynaptic NMDA receptor activation, decreases the Ca(2+) available to enter individual presynaptic boutons of CA3 pyramidal cells. This in turn reduces the probability of consecutive synaptic releases at CA3-CA1 synapses and therefore contributes to short-term paired-pulse depression of minimal responses. This activity-dependent depletion of extracellular Ca(2+) represents a novel form of fast retrograde synaptic signaling that can modulate glutamatergic information transfer in the brain.

Frequency-dependent shift from paired-pulse facilitation to paired-pulse depression at unitary CA3-CA3 synapses in the rat hippocampus

Paired recordings between CA3 interconnected pyramidal neurons were used to study the properties of short-term depression occurring in these synapses under different frequencies of presynaptic firing (n = 22). In stationary conditions (0.05-0.067 Hz) pairs of presynaptic action potentials (50 ms apart) evoked EPSCs whose amplitude fluctuated from trial to trial with occasional response failures. In 15/20 cells, paired-pulse ratio (PPR) was characterized by facilitation (PPF) while in the remaining five by depression (PPD). Increasing stimulation frequency from 0.05-0.067 Hz to 0.1-1 Hz induced low frequency depression (LFD) of EPSC amplitude with a gradual increase in the failure rate. Overall, 9/12 cells at 1 Hz became almost "silent". In six cells in which the firing rate was sequentially shifted from 0.05 to 0.1 and 1 Hz, changes in synaptic efficacy were so strong that PPR shifted from PPF to PPD. The time course of depression of EPSC1 could be fitted with single exponentials with time constants of 98 and 36 s at 0.1 and 1 Hz, respectively. In line with the inversion of PPR at 1 Hz, the time course of depression of EPSC2 was faster than EPSC1 (7 s). Recovery from depression could be obtained by lowering the frequency of stimulation to 0.025 Hz. These results could be explained by a model that takes into account two distinct release processes, one dependent on the residual calcium and the other on the size of the readily releasable pool of vesicles.

Long-term potentiation of the AMPA and NMDA components of minimal postsynaptic currents in rat hippocampal field Ca1

The aim of the present work was to study the potentiation of the AMPA and NMDA components of minimal excitatory postsynaptic currents (EPSC) evoked by activation of restricted numbers of synapses. EPSC of neurons in field CA1 in hippocampal slices were recorded in whole-call patch-clamp conditions selected such that both (AMPA and NMDA) components were present, and these were measured in parallel using computational methods in combination with pharmacological receptor blockade. There was a quite strong correlation between the amplitudes of the AMPA and NMDA components and this was regarded as evidence that they were generated by the same synapses. In cases producing this correlation, both components showed essentially equal long-term potentiation lasting from 5 min to 2 h after afferent tetanization. The data did not support the postsynaptic hypothesis and were in better agreement with the concept that the major mechanism for the persistence of the initial phase of long-term potentiation (up to 1-2 h) is based on increases in the quantity of transmitter released presynaptically.

Presynaptic R-type calcium channels contribute to fast excitatory synaptic transmission in the rat hippocampus

The possibility that R-type calcium channels contribute to fast glutamatergic transmission in the hippocampus has been assessed using low concentrations of NiCl(2) and the peptide toxin SNX 482, a selective antagonist of the pore-forming alpha(1E) subunit of R-type calcium channel. EPSPs or EPSCs were recorded in the whole-cell configuration of the patch-clamp technique mainly from CA3 hippocampal neurons. Effects of both NiCl(2) and SNX 482 were tested on large (composite) EPSCs evoked by mossy and associative-commissural fiber stimulation. NiCl(2) effects were also tested on minimal EPSPs-EPSCs. Both substances reduced the amplitude of EPSPs-EPSCs. This effect was associated with an increase in the number of response failures of minimal EPSPs-EPSCs, an enhancement of the paired-pulse facilitation ratios of both minimal and composite EPSCs, and a reduction of the inverse squared coefficient of variation (CV(-2)). The reduction of CV(-2) was positively correlated with the decrease in EPSC amplitude. The inhibitory effect of NiCl(2) was occluded by SNX 482 but not by omega-conotoxin-MVIIC, a broad-spectrum antagonist thought to interact with N- and P/Q-type calcium channels, supporting a specific action of low concentrations of NiCl(2) on R-type calcium channels. Together, these observations indicate that both NiCl(2) and SNX 482 act at presynaptic sites and block R-type calcium channels with pharmacological properties similar to those encoded by the alpha(1E) gene. These channels are involved in fast glutamatergic transmission at hippocampal synapses.

Differences in amplitude-voltage relations between minimal and composite mossy fibre responses of rat CA3 hippocampal neurons support the existence of intrasynaptic ephaptic feedback in large synapses

Computer simulations and electrophysiological experiments have been performed to test the hypothesis on the existence of an ephaptic interaction in purely chemical synapses. According to this hypothesis, the excitatory postsynaptic current would depolarize the presynaptic release site and further increase transmitter release, thus creating an intrasynaptic positive feedback. For synapses with the ephaptic feedback, computer simulations predicted non-linear amplitude-voltage relations and voltage dependence of paired-pulse facilitation. The deviation from linearity depended on the strength of the feedback determined by the value of the synaptic cleft resistance. The simulations showed that, in the presence of the intrasynaptic feedback, recruitment of imperfectly clamped synapses and synapses with linear amplitude-voltage relations tended to reduce the non-linearity and voltage dependence of paired-pulse facilitation. Therefore, the simulations predicted that the intrasynaptic feedback would particularly affect small excitatory postsynaptic currents induced by activation of electrotonically close synapses with long synaptic clefts. In electrophysiological experiments performed on hippocampal slices, the whole-cell configuration of the patch-clamp technique was used to record excitatory postsynaptic currents evoked in CA3 pyramidal cells by activation of large mossy fibre synapses. In accordance with the simulation results, minimal excitatory postsynaptic currents exhibited "supralinear" amplitude-voltage relations at hyperpolarized membrane potentials, decreases in the failure rate and voltage-dependent paired-pulse facilitation. Composite excitatory postsynaptic currents evoked by activation of a large amount of presynaptic fibres typically bear linear amplitude-voltage relationships and voltage-independent paired-pulse facilitation. These data are consistent with the hypothesis on a strong ephaptic feedback in large mossy fibre synapses. The feedback would provide a mechanism whereby signals from large synapses would be amplified. The ephaptic feedback would be more effective on synapses activated in isolation or together with electrotonically remote inputs. During synchronous activation of a large number of neighbouring inputs, suppression of the positive intrasynaptic feedback would prevent abnormal boosting of potent signals.