Pamela Reinagel - US grants

The goal of these experiments is to study the reproducibility of firing patterns in LGN neurons in response to different types of visual stimulation. In my previous theoretical work, based on data collected by experimental collaborators, I showed that a randomly varying visual stimulus could be approximately reconstructed by simply linearly filtering an LGN spike train. Moreover, I showed that the distinctive bursts fired by LGN cells encoded visual information efficiently. In my own experiments during the past two years, I have been studying the reproducibility of individual LGN responses to dynamic visual stimuli. In my final year, I plan to extend this work to the problem of how a population of thalamic neurons collectively encodes natural stimuli. (1) I will study how the statistics of the visual stimulus ensemble affects the reproducibility of individual LGN responses. Response reproducibility will be measured for both the temporal and spatio-temporal natural stimuli. In control experiments, I will manipulate natural stimuli to vary the intensity distribution, the frequency spectrum, or the contrast of the stimuli. (2) I will study whether the noise in an individual LGN response is shared with, or independent from, the noise in other cells responses. To accomplish this I will compare the trial by trial response variations of one cell with those of another simultaneously recorded cell. These studies will shed light on how the responses of individual cells work together to encode natural visual scenes in the LGN.

[unreadable] DESCRIPTION (provided by applicant): The visual system must function under a broad range of illumination conditions. Without adaptation, the dynamic range of visual neurons (1-2 orders of magnitude) could not support functional vision over the dynamic range of natural stimuli (10-11 orders of magnitude). Therefore visual neurons adapt, not only to the mean light level, but also to the average contrast of recent visual stimuli. The physiological basis of contrast adaptation has been studied extensively in the mammalian visual system, but its functional consequences are not known. A direct test of functional consequences will require an information theoretic approach. The proposed study would test two distinct but not mutually exclusive hypotheses about the function of contrast adaptation in the LGN: (1) Contrast adaptation in the LGN serves to maintain sensitivity and therefore coding efficiency under different contrast conditions. (2) Contrast adaptation in the LGN serves to compute a contrast-invariant representation of visual information. To test these hypotheses, we will record neural responses to flickering visual stimuli using extracellular electrodes in the LGN in vivo. The visual information encoded by these responses will be assessed using several computational approaches, including information theory. First we will determine whether the changes in gain in the LGN fit theoretical predictions and test whether gain changes are correlated with the maintenance of coding efficiency across contrasts. Second, we will determine in what respect(s) the LGN response is the same when the same stimulus patterns are presented at different contrasts, and measure how much of the visual information in the LGN is contrast invariant. Finally we will construct a model to test our understanding of contrast adaptation in the LGN. By comparing the output of the model with and without adaptation, we will test the functional consequences in terms of both contrast invariance and efficient coding. This study will contribute an important component to our description of how sensory information is relayed to cortex through the thalamus and why. This could be valuable for the diagnosis or treatment of disorders of contrast adaptation, and for the development of neural prosthetics to restore vision loss. [unreadable] [unreadable] [unreadable]