Retinal circuits transfer prodigious amounts of information using graded release of synaptic vesicles. Vesicles are expensive to manufacture, transport, and refill; also their quantal postsynaptic currents discharge ionic batteries that use about one-third of the retina's total energy. Therefore, vesicles must be used judiciously. This implies that upstream of the well-studied spike code, there must be a """"""""vesicle code"""""""". Although the mechanics of vesicle release are no longer a complete mystery, the important question of how vesicles encode information has hardly been broached. The information that one vesicle conveys to a ganglion cell is not fixed. Rather, it depends on how many other vesicles are delivering the same message. Release from multiple sites of the same presynaptic neuron and from different neurons with overlapping receptive fields carries redundant messages and thus reduces information per vesicle. Vesicle redundancy, though expensive, is essential. For example it improves signal-to-noise and helps create a receptive field center. Here I propose to apply methods from studies of spike coding to measure information rates for vesicles and thus discover some principles for a vesicle code. Recording under voltage clamp, I will measure bits/vesicle and test the following hypotheses: (1) higher rates of bipolar cell quanta in a ganglion cell should reduce net information per quantum. Evoke similar quantal rates in different ganglion cell types (brisk-sustained and sluggish) and test whether brisk-sustained cells, with less redundancy, have more bits/vesicles. (2) Total information conveyed separately by amacrine and bipolar cell vesicles should exceed that of their combined quantal currents. Isolate quantal currents from amacrine and bipolar cells and determine separate information rates. Record combined current and determine combined information rate. (3) Information is lost between the vesicle code and spike code. Compare information rates for quantal currents vs. spike train. (4) Ganglion cells select from multiple bipolar cell types with distinct pass bands. Determine if different frequency components produce distinct quantal currents. Characterize the pass bands of identified types of bipolar cell. Determine which bipolar cell types contact different ganglion cell types. The proposed studies, by measuring information flow at the level of synaptic currents, will develop explicit rules that connect low-level synaptic mechanisms to higher-level circuit architectures and thus contribute to fundamental understanding of retinal function.
Showing the most recent 10 out of 16 publications