An Analysis of Neuronal Receptive Fields at Three Stages of Visual Processing
Author | : Joshua Paul Van Kleef |
Publisher | : |
Total Pages | : 0 |
Release | : 2009 |
ISBN-10 | : OCLC:1443544191 |
ISBN-13 | : |
Rating | : 4/5 (91 Downloads) |
Book excerpt: Visual neurons must efficiently encode visual information that is relevant to the animals' behaviour in an ever-changing visual environment. One of the ways animals are able to adapt to these changes is to alter their behaviour. Another, the topic of this thesis is to alter the processing properties of their visual neurons. I investigate the effect of changing the stimulus on the response properties of single neurons at three stages of visual processing. A reverse-correlation technique is used to measure spatiotemporal receptive fields (STRFs) which measure how the neuron spatially and temporally integrates the linear and nonlinear contrast stimuli presented to it in space and time.This analysis is performed for stimuli with different contrast, colour and sparsity and the changes in the STRFs are analysed in terms of their ability to improve neural performance. I study the processing of ultraviolet (UV) and green light in photoreceptors of an insect simple-lens eye, the median (middle) ocellus of a dragonfly. For the first time I present physiological evidence which shows these eyes are capable of a moderate level of spatial resolution and reveal differences in the way UV and green light sequences are processed. At the next level of processing, large second-order ocellar neurons (L-neurons) were studied. By presenting monochromatic (UV or green) moving bars and gratings I demonstrate, for the first time, that ocellar L-neurons are directionally selective in UV light but not in green light. Using a novel random stimulus I show that for most L-neurons the linear STRFs alter in structure significantly with changes in the stimulus density. With decreases in the stimulus density these cells become lower latency, have better spatial resolution and are spatiotemporally tuned to faster velocities. These new STRFs account significantly better for the response of L-neurons to fast moving bars.In the last section of my thesis I investigate whether the visual stimulus can also influence the receptive fields of neurons further along the visual pathway. To achieve this goal I map the changes in spatiotemporal tuning with changes in contrast of a subset of neurons within the cat primary visual cortex. I provide further evidence that at low contrasts, particular types of neurons, known as complex cells, have response properties similar to the other major type of cortical neuron, simple-cells. These results are combined with theoretical modelling to investigate the theory that complex STRFs are composed of simple-cell components.In summary, this thesis demonstrates at three levels of visual processing how neural STRFs are altered depending on the properties of the stimulus that is driving their response.