Quantal analysis suggests strong involvement of presynaptic mechanisms during the initial 3 h maintenance of long-term potentiation in rat hippocampal CA1 area in vitro.

Long-term potentiation (LTP) is the most prominent model to study neuronal plasticity. Previous studies using quantal analysis of an early stage of LTP in the CA1 hippocampal region (<1 h after induction) suggested increases in both the mean number of transmitter quanta released by each presynaptic pulse (m, quantal content) and postsynaptic effect of a single quantum (v, quantal size). When LTP was large, it was m that increased predominantly suggesting prevailing presynaptic contribution. However, LTP consists of several temporary phases with presumably different mechanisms. Here we recorded excitatory postsynaptic potentials from CA1 hippocampal slices before and up to 3.5 h after LTP induction. A new version of the noise deconvolution revealed significant increases in m with smaller and often not statistically significant changes in v. The changes in m were similar for both early (<1 h) and later (1-3 h) post-tetanic periods and correlated with LTP magnitude. The coefficient of variation of the response amplitude and the number of failures decreased during both early and late post-tetanic periods. The results suggest that both early (<0.5 h) and later LTP components (0.5-3 h) are maintained by presynaptic changes, which include increases in release probabilities and the number of effective release sites. In addition initially silent synapses can be converted into effective ones due to either pre- or postsynaptic rearrangements. If this occurs, our data indicate that the number and the efficacy of the receptors in the new transmission sites are approximately similar to those in the previously effective sites.

Long-term synaptic changes induced by intracellular tetanization of CA3 pyramidal neurons in hippocampal slices from juvenile rats.

Minimal excitatory postsynaptic potentials were evoked in CA3 pyramidal neurons by activation of the mossy fibres in hippocampal slices from seven- to 16-day-old rats. Conditioning intracellular depolarizing pulses were delivered as 50- or 100-Hz bursts. A statistically significant depression and potentiation was induced in four and five of 13 cases, respectively. The initial state of the synapses influenced the effect: the amplitude changes correlated with the pretetanic paired-pulse facilitation ratio. Afferent (mossy fibre) tetanization produced a significant depression in four of six inputs, and no significant changes in two inputs. Quantal content decreased or increased following induction of the depression or potentiation, respectively, whereas no significant changes in quantal size were observed. Compatible with presynaptic maintenance mechanisms of both depression and potentiation, changes in the mean quantal content were associated with modifications in the paired-pulse facilitation ratios, coefficient of variation of response amplitudes and number of response failures. Cases were encountered when apparently "presynaptically silent" synapses were converted into functional synapses during potentiation or when effective synapses became "presynaptically silent" when depression was induced, suggesting respective changes in the probability of transmitter release. It is concluded that, in juvenile rats, it is possible to induce lasting potentiation at the mossy fibre-CA3 synapses by purely postsynaptic stimulation, while afferent tetanization is accompanied by long-lasting depression. The data support the existence not only of a presynaptically induced, but also a postsynaptically induced form of long-term potentiation in the mossy fibre-CA3 synapse. Despite a postsynaptic induction mechanism, maintenance of both potentiation and depression is likely to occur presynaptically.

Principal component analysis of minimal excitatory postsynaptic potentials.

'Minimal' excitatory postsynaptic potentials (EPSPs) are often recorded from central neurones, specifically for quantal analysis. However the EPSPs may emerge from activation of several fibres or transmission sites so that formal quantal analysis may give false results. Here we extended application of the principal component analysis (PCA) to minimal EPSPs. We tested a PCA algorithm and a new graphical 'alignment' procedure against both simulated data and hippocampal EPSPs. Minimal EPSPs were recorded before and up to 3.5 h following induction of long-term potentiation (LTP) in CA1 neurones. In 29 out of 45 EPSPs, two (N=22) or three (N=7) components were detected which differed in latencies, rise time (Trise) or both. The detected differences ranged from 0.6 to 7.8 ms for the latency and from 1.6-9 ms for Trise. Different components behaved differently following LTP induction. Cases were found when one component was potentiated immediately after tetanus whereas the other with a delay of 15-60 min. The immediately potentiated component could decline in 1-2 h so that the two components contributed differently into early (< 1 h) LTP1 and later (1-4 h) LTP2 phases. The noise deconvolution techniques was applied to both conventional EPSP amplitudes and scores of separate components. Cases are illustrated when quantal size (upsilon) estimated from the EPSP amplitudes increased whereas upsilon estimated from the component scores was stable during LTP1. Analysis of component scores could show apparent double-fold increases in upsilon which are interpreted as reflections of synchronized quantal releases. In general, the results demonstrate PCA applicability to separate EPSPs into different components and its usefulness for precise analysis of synaptic transmission.

Quantal analysis suggests presynaptic involvement in expression of neocortical short- and long-term depression.

Long-term depression together with long-term potentiation represent popular experimental models to study synaptic plasticity. However, analyses of the mechanisms underlying the expression of cortical long-term depression are in their infancy and have been confined to the hippocampus. Short- and long-term depression in neocortex is not well understood. Here we recorded small excitatory postsynaptic potentials intracellularly from rat visual cortex slices. The responses fluctuated between several amplitude levels suggesting a quantal nature of the synaptic transmission. Consistent changes in the quantal steps accompanied neither paired-pulse depression (50 ms interval within the pair) nor long-term depression (induced by 1 Hz, 5 min stimulation). The amplitude distributions shifted to smaller values suggesting decreases in the number of quanta released without essential changes in the postsynaptic quantal efficiency. Both the coefficient of variation of response amplitudes and the number of response failures increased; cases were encountered suggesting a very low release probability after depression. Changes in quantal content estimated from the deconvolution analysis were correlated with the magnitude of depression. The findings suggest predominantly presynaptic loci for expression of short- and long-term neocortical depressions. The likely underlying mechanism is a decrease in transmitter release probability. Long-term depression decreased the probability so strongly that some inputs became virtually silent.

Noise deconvolution based on the L1-metric and decomposition of discrete distributions of postsynaptic responses.

A statistical approach to analysis of amplitude fluctuations of postsynaptic responses is described. This includes (1) using a L1-metric in the space of distribution functions for minimisation with application of linear programming methods to decompose amplitude distributions into a convolution of Gaussian and discrete distributions; (2) deconvolution of the resulting discrete distribution with determination of the release probabilities and the quantal amplitude for cases with a small number (< 5) of discrete components. The methods were tested against simulated data over a range of sample sizes and signal-to-noise ratios which mimicked those observed in physiological experiments. In computer simulation experiments, comparisons were made with other methods of 'unconstrained' (generalized) and constrained reconstruction of discrete components from convolutions. The simulation results provided additional criteria for improving the solutions to overcome 'over-fitting phenomena' and to constrain the number of components with small probabilities. Application of the programme to recordings from hippocampal neurones demonstrated its usefulness for the analysis of amplitude distributions of postsynaptic responses.