Enhancement of Information Transmission Efficiency via Feedback in Synapses of the Fly Early Visual System

Metabolic efficiency and the rate of information transmission are both important factors in biological design. Producing synaptic vesicles, for example, comes at a metabolic cost, and while fidelity of synaptic transmission can be made indefinitely high by increasing the rate of vesicle release, such an adaptation will also use up an indefinite amount of energy. A driving principle in the design of the synapse is thus the optimization of the bit rate per vesicle in operational conditions. A study of the contrast power transfer spectrum and the noise power spectral density of large monopolar cells (LMCs) of the blowfly C. vicina predicts a sustained minimum rate of >10⁵ synaptic vesicles per second per LMC. The calculation of this rate assumes that the exocytosis of vesicles occurs as an unstructured (Poisson) process. If instead, we assume that vesicle release is more structured, noise may be suppressed at the low end of the spectrum, substantially bringing down the release rate required for observed experimental results. Weckström and Laughlin (2010) show that significant electric potentials are generated in the photoreceptor-LMC extracellular space as a response to visual activity. We propose that regularization of vesicles occurs through the control of the voltage sensors responsible for vesicle release by these feedback potentials. We demonstrate signatures of synaptic control experimentally, and through modeling, the consequences of the measured extracellular responses on the bit rate per vesicle.