Pyramidal Cell Responses to Temporally Structured Stimuli: Experiments and Computer Simulations

Citation

Protopapas, Alexander D. (1998) Pyramidal Cell Responses to Temporally Structured Stimuli: Experiments and Computer Simulations. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/zv26-tp58. https://resolver.caltech.edu/CaltechTHESIS:07252025-202233992

Abstract

Oscillations in the field potential recorded from piriform cortex can be broadly categorized into slow and fast frequency ranges. The slow wave is correlated with respiration and sniffing. During respiration it is typically in the 1-4 Hz range but during sniffing it increases in frequency and is often referred to as the theta rhythm (4-12 Hz). The faster oscillations (30-50 Hz, also called gamma) appear in response to odor stimuli and are always modulated by the slower rhythm. Oscillations in the field potential are believed to reflect synchronized synaptic input to the dendrites of pyramidal neurons in the piriform cortex. In this thesis I use a combination of experimental and computer simulation techniques to study the consequences of pyramidal cell input meant to approximate the temporal characteristics of cortical oscillations.

Because the precise spatial and temporal control of synaptic inputs is not possible in an experimental preparation, I constructed a detailed biophysical simulation of a layer II pyramidal cell from piriform cortex where such control would be possible. The passive and active properties of this model were tuned to experimental measurements that I made from pyramidal cells in vitro. The model was able to match a wide range of physiological behavior including subthreshold oscillations and responses to multiple levels of current injection. Spatio-temporal patterns of synaptic input that have been suggested to underlie gamma oscillations in piriform cortex were then used as input to the model. The effects of a single such pattern of input were longer lasting than the duration of a single gamma oscillation suggesting that a pyramidal cell integrates input over multiple gamma oscillations during the course of bursts of gamma oscillations modulated by the respiratory / theta rhythm. When bursts of gamma activity in the model were separated by 650 msec or more, the first burst had no effect on the second, implying that these neurons might be able to isolate the effects of sufficiently spaced sniffs or bouts of sniffing.

To determine how well current injections with the temporal characteristics of cortical oscillations might be represented in the spike trains of pyramidal cells, I used a reconstruction algorithm to estimate the structure of the stimulus from spike train data. By comparing the estimate to the actual stimulus I was able to quantify the amount of stimulus information contained in the spike train. I found that stimuli filtered at frequencies of 0-10 Hz and 4-12 Hz were much better represented in the pyramidal cell spike trains than 0-40 Hz stimuli designed to include the entire frequency range of cortical oscillations. The effects of norepinephrine (a neuromodulator released during arousal) on spike coding were also studied. I found that while norepinephrine increased the amount of stimulus information in the spike train, a change in decoding strategy to extract this information from the spike train was not required.

Thesis Files