Temporal And Frequency Characteristics Analysis Of Retinal Ganglion Cells Population Activities

The retina is the primary stage of visual information processing. The Photoreceptors transmit the light stimulus into electrical signals for being processed further in the retinal circuit. The retinal ganglion cells, which are the output neurons of retina, elicit spike trains containing the original visual information and transmit these spike trains through their axons eventually to brain. The neuronal synchrony makes great significance to the encoding and transmitting of visual information. The recently developed multi-electrode array (MEA) allows researchers to record multiple neurons simultaneously, which offers one way to explore the mechanism of neural encoding. Using chicken's retina as our experimental material, we first design and programme several visual stimulus protocols, then analyze the experimental data based on information discrepancy to measure the temporal characteristics of a group of synchronous retinal ganglion cells and at last calculate the correlation of coupled neurons in frequency domain by a method named coherence analysis. The main efforts made in this dissertation are listed as below:1) The miscellaneous stimulus patterns we design and realize are aimed to explore the retinal ganglion cells, including flashing dot, checkerboard, and sliding bar from different directions as well as C ring-like ones by employing C# and DirectX graphic programming language. In addition, the stimulus system we set up is modular and scalable for further extensions.2) A new method based on information discrepancy is applied and the results show that the neuron population exhibite increasing synchrony over adaptation time.3) The correlation of coupled neurons in frequency domain is also investigated by introducing coherence analysis into our experimental data. Comparing to correlation in time domain, this coherence analysis doesn't advances only in determining the correlation of the neuron couple, but also finding out at which frequency they are more like to pairing through calculating an objective confidence limit. This method gives an simple expression for fast computation. Further more, the population of coupled neurons can also be analyzed at the same time if the correlation induced by stimulus between channels of MEA can be eliminated.Our results show that the pair-wised as well as the population neurons under the stimulus exhibit strong synchronization and the synchronization increases over time; which means that they are more"tight"during the adaptation time course. The temporal precision is in millisecond. This indicates that the stimulated population of neurons evolves dynamic connection without being physically connected. The synchronized EPSPs are able to transmit the electrical signal that embeds the stimulus information in their temporal structure to post neurons by smaller latency but higher precision and efficiency. This temporal encoding mechanism possesses less energy consumption, more transmitting efficiency and metabolic significance.