Abstracts

Rationale: Though < 20% of neurons, INs shape normal and pathological circuit dynamics. Parvalbumin (PV) and somatostatin (SOM)-INs are two key types. PV-INs innervate somas, are divergent to 1000 or more excitatory cells, are fast-spiking, and are strongly driven by excitatory inputs for feedforward inhibition. They shape normal circuit oscillations over multiple frequencies and are implicated as early participants with seizures. SOM-INs act on distal dendrites and fire sparsely for more local processing important with feedback inhibition. To probe the role of stimulation patterns of DBS on INs, we evaluate the properties of PV- and SOM-INs to various patterns of electrical stimulation in vitro using calcium imaging of IN subtypes. In addition, computational models of GABA-AR phasic, spillover, and tonic currents predict post- and extra-synaptic inhibitory output responses from INs driven with various frequency/burst stimulation patterns.

Methods: INs (PV- or SOM-Cre) were injected with virus (AAV1-Syn-FLEX-GCaMP6f) 4 wks before hippocampal slice preparation with imaging (50 Hz) of responses to evoked stimulation (10 – 0.1mA) by paired-pulses (1280 – 40 ms) or brief stimulus trains (1 – 50 Hz). Miniature, spontaneous, and evoked/paired-pulse IPSCs as well as GABA-AR mediated tonic currents were recorded from dentate gyrus granule cells in hippocampal slices. Computational models of synaptic, tonic, and evoked currents using a 7-state receptor kinetic model for GABA-ARs were optimized to fit experimental IPSCs/tonic currents.

Results: SOM-IN calcium fluorescence effectively tracks individual pulses while PV-IN activation is favored by high-frequency trains of stimulation. After hi-freq hi-intensity activation, the PV-IN fluorescence lasts > 2-3x longer than the SOM-IN responses, often exceeding10s. SOM-INs have lower thresholds for activation compared to PV-INs, with robust induced fluorescence over 0.1 to 10 mA intensities. Spontaneous hypersynchronous activity in SOM-INs also is more easily induced with 0 Mg++ aCSF. For GABA-AR models, above 40 Hz the extrasynaptic δ subunit-containing GABA-ARs cannot track inputs and integrates individual evoked responses. The DC contribution from δ relative to g2 subunit-containing GABA-ARs becomes more pronounced with increase to 150 Hz, with significant desensitization impacting the latter. Between 3-6 Hz, responses are oscillatory and track stimulation while also maintaining synaptic desensitization / circuit hyperactivity. Adding ‘bursting’ to low frequency driving increases the peak inhibition 60%, but accentuates desensitization 250%.

Conclusions: Specific patterns of stimulation selectively engage subpopulations of INs as well as the post- and extrasynaptic GABA-ARs that mediate their output control of excitatory cells. The effects of even brief patterns of stimulation can long outlast the stimulus by 10s of sec or longer. Optimization of complex and/or intermittent stimulation should facilitate specific targeting of cell subtypes, as well as their downstream outputs, for greater control of circuit dynamics.

Funding: Please list any funding that was received in support of this abstract.: None.