Abstracts

RATIONALE: Epilepsy is often described as a condition of excessive neuronal spiking. A more accurate description is that of inappropriate [italic]patterns[/italic] of spiking. Relatively little is known, however, about how patterns of synaptic input give rise to patterns of output, or indeed, what the native patterns of synaptic input are in health and during seizures. To explore these issues, we studied the rates, timing precision and patterning of spikes in excitatory granule (GC) and inhibitory basket cells (BC) in slices of rat dentate gyrus, using simple stimuli and complex patterns of simulated synaptic conductance changes.
METHODS: We used an analog dynamic current-clamp circuit (J.Neurophysiol,69:992; J.Neurosci.Meth.49:157) in whole-cell patch-clamp recordings (standard solutions, room temp.) to inject currents or conductances, and to mimic the amplitudes and shapes of excitatory and inhibitory synaptic conductances. Three stimulus conditions were compared: square current steps (1 sec., 50-100 pA), square conductance steps (1 sec, 2-5 nS), and naturalistic patterns of randomly timed and sized synaptic conductance waveforms (mean rate = 100 Hz, mean amplitude = 5 nS; excitatory and inhibitory decay time constants = 5 ms and 25 ms). Spike timing precision was measured in response to 100 identical stimulus trials, and was defined as the standard deviation of a Gaussian fit to the central peak of the crosscorrelogram over all trials. Statistical significance was assessed using ANOVA with post hoc t-tests.
RESULTS: Current steps (1 sec., 50-100 pA), conductance steps (1 sec, 2-5 nS) and naturalistic conductance patterns caused similar spike rates in GCs and BCs (pooled mean [plusminus] SEM, 7 [plusminus] 1 Hz, n = 15 GC & 13 BC). Crosscorrelagrams for current steps had large side peaks, indicative of periodic spiking. Less periodicity was observed for conductance steps. Naturalistic stimuli produced no periodicity at all, but showed irregular interspike intervals with remarkable precision in spike patterns from trial to trial. Current steps yielded a precision of 3.6 [plusminus] 0.8 ms (n=8), whereas spiking was more than twice as precise with conductance steps (1.6 [plusminus] 0.3 ms, n=8), and seven times more precise with naturalistic stimuli (0.5 ms [plusminus] 0.01 ms, n=7). The spike patterns evoked with the same naturalistic stimulus were similar between cells of the same type, and the stimulus features responsible could be reconstructed from the spiketrain.
CONCLUSIONS: We conclude that GCs and BCs respond with specific patterns of spikes produced with submillisecond precision when driven by naturalistic synaptic conductance patterns. These results suggest that harmful patterns of seizure activity may arise from specific patterns of synaptic drive.
Support: This project was supported by an Epilepsy Foundation Junior Investigator Research Grant to MVJ.