|M.Sc Student||Dvorkin Roman|
|Subject||Similarity of Activity-dependent Structural Changes of|
Synapses Formed between Identical Neurons
|Department||Department of Medicine||Supervisor||Professor Noam Ziv|
|Full Thesis text|
Chemical synapses are specialized sites of cell-cell contact designed to transmit signals between nerve cells. It is widely believed that activity-dependent changes in synaptic function and structure represent major mechanisms for changing brain function, giving rise to phenomena collectively referred to as learning and memory. Previous studies, however, have shown that synapses, rather than being static structures, exhibit significant cellular and molecular dynamics and undergo structural changes even without electric activity. These studies raised significant questions concerning the ability of synapses to serve as reliable "devices" for modifying network properties and concerning the ability of networks to compensate for such apparently stochastic processes. To start and explore these questions we decided to address a simpler, yet important issue: What is the relative part of specific activity histories in the structural remodeling synapses undergo?
To address this question we developed means to identify, in networks of rat cortical neurons in primary culture, synapses formed between the same axonal and dendritic segments, reasoning that such synapses should have a common activity history, and thus should exhibit similar structural changes over long time periods. We then used confocal imaging combined with concomitant recording of network activity to examine, over many hours and days, the degree to which synaptic remodeling in such synapses co-varies, and how greater this covariance is in comparison to synapses connected to unrelated upstream neurons . We found that the covariance of synaptic remodeling in synapses innervated by the same neuron was slightly greater that the covariance observed for synapses connected to unrelated upstream neurons, but that this difference was rather modest (~12%). Interestingly these differences disappeared once network activity was silenced. Moreover, silencing activity led to a 20-30% increase in remodeling covariance between all synapses in the network regardless of the identity of their upstream partners. These finding suggest that the relative part of specific activity histories in determining synaptic remodeling is relatively minor (at least in our system), further underscoring the importance of the aforementioned questions concerning the long-term reliability of individual synapses.