Roberts Building 508 (<map>)
A Stochastic Adaptive Sampling Scheme in Fly Photoreceptors
In neuroscience, it is a traditional challenge to understand the input-output relationship of a sensory neuron. In the context of vision, how do a photoreceptor constantly adjust this relationship, so that it could effectively utilise the limited response range to represent the dramatic large input range from starlight to direct sunlight?
We recently uncovered such mysteries by generating realistic biophysical computational models of fly photoreceptors. The models explain how this adaptation process could be understood by a stochastic adaptive sampling principle out of only 4 parameters.
A Drosophila photoreceptor integrates light information by stochastic adaptive sampling rule. Its photo-sensitive waveguide (rhabdomere) consists of ~30,000 microvilli, each of which is capable of generating single photon responses (quantum bumps). So a fly photoreceptor is essentially a photon counter from a sampling point of view, with each photon counted as a bump.
In this talk, I will explain how the light adaptation dynamics emerge from the sampling of a huge population of refractory units, how the refractoriness in sampling help to boost contrast changes, and how the neuronal responses are characterised by 4 key sampling parameters. I would also explain how stochasticity in the sampling process contributes to visual information encoding, how a gain control mechanism results in adaptive sampling that approximates contrast constancy.
At last, I will show that such a stochastic adaptive sampling principle accurately predicts information processing across a range of fly species with different visual ecologies, supporting its general role in encoding sensory information. It will be interesting to examine whether this encoding principle is also applicable to other sensory neurons or to a population of synapses/neurons.
Page last modified on 27 oct 13 22:18