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

Measurement and analysis of postsynaptic potentials using a novel voltage-deconvolution method

Accurate measurement of postsynaptic potential amplitudes is a central requirement for the quantification of synaptic strength, dynamics of short-term and long-term plasticity, and vesicle-release statistics. However, the intracellular voltage is a filtered version of the underlying synaptic signal and so a method of accounting for the distortion caused by overlapping postsynaptic potentials must be used. Here a voltage-deconvolution technique is demonstrated that defilters the entire voltage trace to reveal an underlying signal of well-separated synaptic events. These isolated events can be cropped out and reconvolved to yield a set of isolated postsynaptic potentials from which voltage amplitudes may be measured directly-greatly simplifying this common task. The method also has the significant advantage of providing a higher temporal resolution of the dynamics of the underlying synaptic signal. The versatility of the method is demonstrated by a variety of experimental examples, including excitatory and inhibitory connections to neurons with passive membranes and those with activated voltage-gated currents. The deconvolved current-clamp voltage has many features in common with voltage-clamp current measurements. These similarities are analyzed using cable theory and a multicompartment cell reconstruction, as well as direct comparison to voltage-clamp experiments.

Timing and contributions of pre-synaptic and post-synaptic parameter changes during unitary plasticity events at CA3-CA1 synapses

At individual synapses, post-synaptic responses include a mixture of "successes" and "failures" in which transmitter is released or not released, respectively. Previously we measured synaptic strength at CA3-CA1 synapses averaged over all trials, including both successes and failures, using an induction protocol that allowed us to observe potentiation and depression events as step-like changes. Here we report quantal properties of 15 of the earlier experiments, including 14 potentiation events and eight depression events. In five experiments both potentiation events and depression events were evoked at the same synapse. During potentiation, success rate increased from 0.56 +/- 0.14 (mean +/- SD) to 0.69 +/- 0.12, and during depression, success rate decreased from 0.70 +/- 0.09 to 0.51 +/- 0.10. During potentiation potency increased from 10 +/- 5 to 19 +/- 9 pA, and during depression, potency decreased from 18 +/- 12 to 12 +/- 7 pA. On average, changes in potency accounted for 76% of the change in response size in potentiation events and 60% of the change in depression events. A reduced-assumption spectral analysis method showed evidence for multiple quantal peaks in distributions of post-synaptic current amplitudes. Consistent with the observed changes in potency, estimated quantal size (Q) increased with potentiation and decreased with depression. A change in potency, which is thought to reflect post-synaptic expression mechanisms, was followed within seconds to minutes by a change in success rate, which is thought to reflect pre-synaptic expression mechanisms. Synaptic plasticity events may therefore consist of changes that occur on both sides of a synapse in a temporally coordinated fashion.

Joint application of independent component analysis and non-stationary fluctuation analysis for studying the mechanisms of the early phase of long-term potentiation in the rat hippocampus

Interest in studies of synaptic plasticity have led to the development of specific methods for analyzing evoked postsynaptic currents: quantum analysis, component analysis, non-stationary fluctuation analysis, etc. However, the use and interpretation of these methods are not always consistent, which leads to the acquisition of contradictory results from similar experiments. In the present study, simulations were used to analyze the influences of the heterogeneity of the shapes of postsynaptic currents on the results obtained by non-stationary fluctuation analysis. Assessment of postsynaptic channel conductivity was found to depend on the number of synapses involved in generating the response. It is hypothesized that the increase in the assessment of AMPA channel conductivity reported in the literature in conditions of long-term potentiation may be related to changes in the synaptic composition of postsynaptic currents. The hypothesis was tested using a new method for independent component analysis. Studies using the simulation showed that the joint application of this method and non-stationary fluctuation analysis avoids errors of this type. The procedures developed here were applied to data obtained from physiological experiments; the results of this exercise provided general support for the hypothesis.

Rapid change of quantal size in PC-12 cells detected by neural networks

The basic building block of synaptic transmission-the number of molecules released per vesicle (quantal size (QS)) often changes with stimulation, but there is no agreement about what factors regulate it. To throw more light on this problem spontaneous quantal release was recorded amperometrically in PC-12 cells. Amperometric current spikes, representing single vesicle release, were detected by thresholding and were separated from spurious events on the basis of their amplitude and time course using a pattern recognition system based on the principal component neural network methods. The frequency of current spikes, their amplitude, quantal size, rise time and decay time were typically non-stationary, even in the absence of stimulation. Their running values changed much more than those of memoryless stationary random data with the same probability density distribution. Irrespective of how much the quantal size, rise and decay times varied, their amplitude dependence remained constant, or changed with a very different time course. In conclusion, the quantal size is highly labile in PC-12 cells. The lability does not appear to result from the changes of fusion pore dynamics or the mechanism of release of vesicular content, but because of the preferential release of large vesicles.