Realistic Modeling of Rod Bipolar and AII Amacrine Cells: Synaptic and Intrinsic Properties of Neurons Comprising a Retinal Microcircuit

In order to elucidate how neuronal circuits transform input to output, it is important to characterize the properties of signaling between cells as well as the intrinsic features of individual cells. In this work, we seek to address both of these areas in the context of a retinal microcircuit by studying rod bipolar (RB) and AII amacrine (AII) cells. Our work presented here describes two complementary features of circuit processing: first, we elucidate the mechanisms of synaptic communication between presynaptic RBs and postsynaptic AIIs; second, we describe the membrane conductances and the electrotonic structure of the AII. In both of these areas, we employ computational modeling in close conjunction with experimental recordings.;For the first line of inquiry, we study physiological signaling between presynaptic RBs and postsynaptic AIIs. Using experimental data from the Singer Laboratory, we employ techniques from systems identification to construct a phenomenological description of the transfer function at this synapse. To generate mechanistic insight into these experimental results, we build a realistic model of vesicle cycling at the RB-AII synapse. This model captures experimentally-observed gain changes; moreover, our model demonstrates that different forms of adaptation can emerge at this synapse due to variable contributions of vesicle depletion, Ca channel inactivation, and uncorrelated release.;For the second line of inquiry studying the intrinsic properties of the AII, we aim to elucidate the mechanisms underlying the small, TTX-sensitive spikes seen at the AII soma. Using a stylized computational model, we show that firing in AIIs can be explained by an unexpected but simple electrotonic structure: spikes in the AII are initiated at a single dendritic location that is electrotonically distal to the soma. We demonstrate that the small spikes recorded at the soma represent filtered waveforms of larger dendritic events that resemble action potentials. This distal spiking is modulated by a source of slow negative feedback, likely an M-type K conductance, that is electrotonically colocalized with dendritic Na. To our knowledge, this description of distal dendritic spiking from a single location is unique among mammalian neurons.