Abstracts

Rationale: Inhibition in the brain is mediated by a variety of inhibitory neurons, each with distinctive short-term synaptic dynamics in excitatory inputs and inhibitory outputs. Excitation from regular spiking (RS) neurons onto low-threshold spiking (LTS) neurons facilitates, and inhibition from LTS onto RS neurons depresses. It has been suggested that strong facilitation of RS-to-LTS synapses could engage LTS-mediated inhibition to protect cortex from overexcitation. Empirically-based computational models of RS-LTS networks have revealed, however, that LTS-to-RS depression limits the inhibitory efficacy of LTS cells. As characterization of short-term synaptic dynamics is crucial for understanding the protective role of LTS cells, we examine them experimentally in detail. Methods: Whole cell patch-clamp recordings were made from synaptically connected LTS-RS cell pairs in layer 2/3 of neocortex from juvenile mice. A subset of LTS cells was visualized using the GFP-expressing inhibitory neuron (GIN; Jackson Labs) strain. Trains (12, 50 or 100 spikes) were induced in the presynaptic cell at frequencies from 3 to 60 Hz. Recovery kinetics were determined from responses to single spikes at post-train delays from 200 ms to 5 s. The amplitude of the postsynaptic potential (PSP) was measured as a function of train frequency and recovery time for both excitatory and inhibitory synapses. Results: EPSPs (RS-to-GIN/LTS) in response to short trains strongly facilitated above 10 Hz but were weak and unreliable below 10 Hz. For longer trains (100 stimuli), however, lower frequencies showed delayed yet persistent facilitation and higher frequencies showed rapid facilitation followed by strong depression. By contrast, we found that IPSPs (GIN/LTS-to-RS) were depressing, with larger depression at higher frequencies, although depression at high frequencies was less than predicted from computational models with a constant recovery time. Surprisingly, recovery from depression also showed a brief period of augmentation for about 1.5-2 s before returning to baseline. Lastly, we found that for longer trains (50-100 stimuli) IPSP depression recovered during the train at lower frequencies while augmentation lasted longer (up to 5s) at high frequencies. Conclusions: Excitatory RS cell-to-GIN cell synapses showed strong facilitation at high frequencies of activation, as described previously. For longer stimulus trains, however, we found that lower frequencies were still effective in recruiting the GIN network because facilitation was delayed but long-lasting. Additionally, depression of inhibition was not as large as expected: 1) high frequencies showed less depression than models predict, 2) augmentation of inhibition occurred after high frequency stimulation, and 3) recovery from depression could begin prior to the end of the PSP train at low frequencies. These factors may allow the GIN/LTS cell network to be recruited during high frequency or persistent excitation and provide significant inhibitory tone to the network in response to enhanced activity that precedes seizures.