Frequency-dependent properties of neurons and synapses in an oscillatory network

The oscillatory activities form neural networks are involved in many behaviors, and animals need to be able to control the frequencies of these activities to respond to the environmental challenges. Neurons in many systems have frequency-dependent properties and preferred frequencies (also known as resonance). We hypothesize that the activity frequency of an oscillatory network is determined by the preferred frequencies of the neurons and of the synapses in the network. We examined this hypothesis by investigating the frequency-dependent properties of the neurons and of the synapses in the pyloric network in the crab Cancer borealis. We also examined what factors affect the preferred frequencies and how changing in these factors influence the frequency of the network activity. We first showed that the preferred frequency of neurons could be measured with the voltage-clamp technique. Measuring the preferred frequency with the voltage-clamp technique allowed us to have a full control of the voltage range and the waveform of the oscillation. By shifting the voltage range of the oscillation, we found that the pacemaker PD neuron has a higher preferred frequency when it is oscillating at a higher voltage range, and the preferred frequency of the follower LP neuron is only affected by the upper bound of the oscillation. The PD neuron also has different preferred frequencies when oscillating with different waveforms. Specifically, one waveform parameter, the 75 - 100% rising slope, showed a negative correlation with the preferred frequency. After knowing that the voltage range and the waveform of the oscillation are correlated with the preferred frequency, we used dynamic clamp to alter the voltage range and the waveform of the PD oscillation during the ongoing activity and measured the pyloric frequency. Based on our hypothesis, we expected the voltage range and the waveform would have similar effects on the pyloric frequency as they do on the preferred frequency. Indeed, our result showed that the shifts in the pyloric frequency during the dynamic clamp experiments could be explained by the changes in the voltage range and in the waveform parameters. Finally, we examined the frequency-dependencies of the amplitude and the phase of the synaptic current. The amplitudes of the synaptic currents between the AB/PD and LP neuron showed preferred frequencies. Interestingly, the preferred frequencies of the synapses were significantly lower than those of the presynaptic neurons and also than the pyloric frequency. While the voltage range of the presynaptic PD oscillation did not affect the preferred frequency of the AB/PD to LP synapse, the preferred frequency of the LP to PD synapse was higher when the upper bound, but not the lower bound, of the LP oscillation was increased. Moreover, the strength of the synaptic resonance depended on the upper bound of the presynaptic oscillation. To produce the strongest resonance, the upper bound of the presynaptic oscillation need to be within the voltage range at which the synapse is most sensitive to the presynaptic membrane potential. In addition to the amplitude, the phase of the synapses also showed frequency-dependence. At low frequencies ( 1 Hz). Overall, in this study, we demonstrated that many properties of the neurons and synapses depend on the frequency of the oscillation and have preferred frequencies. Moreover, these preferred frequencies can be regulated by many factors, including the voltage range and the waveform of the oscillation. Because some of these frequency-dependent properties are able to influence the network frequency, the factors affecting their preferred frequencies could change the network frequency in the same way. As a result, the frequency of an oscillatory network is not determined by a single factor, but by the dynamic interactions among the frequency-dependent properties of the network components.Ph. D.Includes bibliographical referencesIncludes vitaby Hua-an Tsen