Modulation Of Adaptation Currents On The Response Characteristics Of Retinal Ganglion Cells And Electrically Coupled Neuronal Networks

Adaptation is an extraordinary phenomenon widely observed in neura l systems, and its typical manifestation is that neurons dynamic ally adjust their firing rate(primarily the gradual decreas e of firing rate) during sustained stimuli. Previous reports have pointed out that adaptation exhibits conspicuous roles in helping neurons encode external stimulus information, such as : modulating the selectivity of neurons to various stimuli, lowering the redundancy of neural information encoding, and so on. However, some detailed functions that adaptation performs in neural information processing are not completely clear yet, such as : possible contributions to the generation of diverse firing patterns in neurons, and spec ific modulations on the population coding manners that neural networks adopt. Mechanis ms underlying the generation of adaptation can be c lassified into two parts :(1) intrins ic mec hanis ms(ion currents) and extrins ic mechanis ms(synaptic inputs). In this thes is, starting from the intrins ic mec hanis ms, the roles of adaptation currents in modulating the intrins ic firings of s ingle retinal ganglion c ells(RGCs), and in controlling the synchronous activities of electric ally coupled neuron populations were investigated respectively by computational models combined with several reported experimental results.The main work consisted of two parts. The first part primarily focused on single RGC. First, by proposing a modified ionic model, the functional roles of adaptation currents in generating several intrins ic firings of RGCs were analyzed; Second, based on the modified models combining with a previous ly reported RGC model, the different firing dynamics that these intrins ically distinct RGCs exhibit in encoding external inputs were compared. The second part primarily used the methods derived from graph theory, and analyzed how the recorded cells electric ally coupled with each other at the population level. Then, by constructing several typical electric ally coupled neuronal networks, how the population activities of networks formed by non-adapting neurons differ from the networks that formed by adapting neurons were investigated.The main res ults are:(1) computational model confirmed that two ionic currents(s lowly inactivated sodium current and delayed-rectifier potassium current) intrins ic ally exist on the membranes both partic ipate in modulating the adapting activity of RGCs. Moreover, the two adaptation currents activate collectively to contribute to the generation of two typical firing behaviors, i.e., tonic and phas ic firings, observed in RGCs;(2) three intrins ically distinct RGCs can exhibit the Hodgkin’s three c lasses of firing dynamics. In addition, in encoding periodic stimuli, different RGCs behave differentially but complementarily;(3) electrically coupled networks derived from the recorded neuron populations can be approximately described by small-world networks. Furthermore, adaptation observably decreases the correlation strength between electric ally coupled neuron pairs. However, for population neuron activities, adaptation behaves in a non-monotonous ly way: weak adaptation reduces the population synchronization, while strong adaptation stabilizes the population sync hronization. Model results further indic ated that the presence of adaptation would induc e the activity of neurons switch from spiking to bursting.The results demonstrated in this work revealed that s lowly inactivated sodium current and delayed-rec tifier potass ium current prominently influence the generation of tonic and phas ic activities of RGCs. In addition, the three intrins ically distinct RGCs can exhibit three different exc itabilities, respectively, suggesting that intrins ic mechanis ms partic ipate in the different coding manners performed by RGCs. At the level of population neuron activities, networks composed of neurons with different degrees of adapting activity produc e rather different population patterns, indic ating that neuron ens embles can make information effic iently processed us ing various modes of population code. In summary, the above res ults suggested that adaptation currents are imperative for promoting the effic ient coding of neural information.