Abstracts

RATIONALE: The balance between inhibition and excitation plays a crucial role in generation of bursting activity in neuronal circuits. In previous studies of networks of single compartment neurons, we have demonstrated that modification of this balance by changing synaptic weight can significantly alter the excitability of neural networks. The objectives of this study are to investigate the roles of the relative location of synapses on the dendritic tree and soma in the generation of bursting activity.
METHODS: We have built a reduced pyramidal model of synaptically connected neurons using the simulation software GENESIS. Three reduced pyramidal neurons and an interneuron are modeled in this study: a neuron where random input is applied to generate action potentials, two neurons connected with excitatory synapses, and an inhibitory interneuron in a negative feedback loop with one of the modeled pyramidal neurons. The synaptic weight represents the overall strength of a connection and the synaptic delay represents all delays between neurons. We have investigated several possible locations of excitatory connections between two pyramidal neurons. We have also studied the effects of different locations of inhibitory inputs relative to the soma and the excitatory connections.
RESULTS: Simulations show that a reduced pyramidal neuron model can produce repetitive burst activity. This bursting depends upon synaptic parameters (weight and delay) and the locations of synaptic inputs. To produce the same bursting pattern, the synaptic weight needs to be larger when the excitatory synaptic inputs are located on the main or branch dendrite, than when they are on the soma. Simulations including inhibition reveal that the inhibitory interneuron regulates neuronal bursting activity. The pattern of bursting behavior depends upon the synaptic weight and delay of the inhibitory conection, as well as the location of the synapse. Inhibitory action is stronger when the inhibitory synapse is close to the soma. If the inhibitory synapse is on a branch dendrite, synaptic weight has to be increased to produce the same effect.
CONCLUSIONS: A reduced multicompartmental pyramidal neuron model can replicate bursting behavior. In this model the pattern of bursting activity is dependent upon the synaptic weight and delay as well as the locations of excitatory and inhibitory synaptic connections. The inhibitory interneuron has greater influence on the pattern of bursting activity when its output is connected closer to the soma of the pyramidal neuron. These models may be useful for modeling changes in synaptic inhibitory inputs that have been observed in experimental animal models of epilepsy.
Support: NIH grant NS 38958