Synaptic short-term plasticity differentially shapes signal transduction capabilities in the output stage of the basal ganglia
CBN (Computational Biology and Neurocomputing) seminars
Friday 08 June 2012
to 15:00 at
Mikael Lindahl (CB/CSC/KTH)
The basal ganglia are a group of subcortical nuclei hypothesized to control selection of actions through disinhibition of motor centers in the thalamus and in the brain stem. Recent work has shown that synaptic connections within the basal ganglia display frequency dependent responses such as facilitation or depression. Such dynamic synapses have been shown to play important roles in other brain structures, however, their functional role in the basal ganglia has not been investigated. Using a computational model we investigate the consequences of dynamical synapses onto the substantia nigra pars reticulata (SNr), one of the basal ganglia output stages. SNr receives convergent inhibitory input from the medium spiny neurons (MSN) in the striatum (Str) and from projection neurons in the globus pallidus externa (GPe) through synapses exhibiting, respectively, facilitating and depressing short-term plasticity. Using emulated MSN and GPe synaptic spike patterns, we predict that facilitating synapses in the striato-nigral pathway enhance high frequency inputs such that bursts in only 2% of the MSN pool projecting to a specific SNr neuron is sufficant to interrupt its spiking. At the same time increased non-specific inhibition due to a moderate global increase in the whole striatal pool does not occur. Simulations further show that the pallido-nigral pathway contributes with a transient action signal following the onset of GPe activation while a sustained increase in mean rate activity in GPe does not cause non-specific inhibition of SNr. Finally it is shown how coincident activation of MSNs and GPe neurons projecting to the same SNr neuron improves the temporal control of SNr inhibition. In contrast, pauses in GPe bursting, as often seen in experiments during dopamine depleted conditions can prevent effective an action signaling.