click to enlarge

The bipolar cell ribbon synapse is part of a feedback circuit. The image to the right shows two sections from the same retina. The first section (A) shows a bipolar cell axon (bp) containing a "ribbon" (r). At the ribbon, the bipolar cell synapses on an amacrine cell (am) and a ganglion cell (gc). This synapse releases the excitatory neurotransmitter glutamate. In another section (B) about 600 nm away from the first, the amacrine cell synapses back onto the bipolar cell. This synapse releases the inhibitory neurotransmitter GABA or glycine.

Note that the entire circuit bipolar cell > amacrine cell > bipolar cell constitutes a compact negative feedback loop. The diagram to the left represents a computer simulation of the feedback loop, complete with the different receptors (mGluR, iGluR, GABA_C) and the postsynaptic current engendered by a glutamate or GABA vesicle (EPSC, IPSC). We find that the simulation causes the bipolar cell to "hiccup", releasing glutamate vesicles in brief random bursts. Electrophysiology of the ganglion cell shows that indeed the bipolar cell hiccups in just such a way. Apparently these hiccups come from the innate noisiness of vesicular release and indicate the circuit's ability to regulate release of single vesicles with millisecond accuracy  (Freed et al, 2003).

Information transmission by the bipolar cell ribbon synapse It's well accepted that chemical synapses such as the bipolar cell ribbon synapse transmit information; we have empirically estimated how much. A ganglion cell receives thousands of synapses from bipolar cells. When a naturalistic stimulus is presented, it receives from these synapses 10 to 50 bits per second (these are the same bits used to measure information stored on a computer). But the ganglion cell receives at most about .025 bits per second from each synapse and at most about 0.4 bits from each vesicle  (Freed, 2005).

The amount of information a ganglion cell receives from a vesicle is not necessarily as much as was originally transmitted. One vesicle transmits much the same kind of information as another vesicle, i.e. vesicles are redundant. To simplify, consider that each vesicle provides one copy of this redundant information, but when these copies are combined in the ganglion cell, only one copy remains. Thus a ganglion cell that receives information from few vesicles should receive more information from each vesicle; we are actively investigating this idea.