Low-level visual processing and its relation to neurological disease

Retinal neurons extract changes in image intensity across space, time, and wavelength. Retinal signal is transmitted to the early visual cortex, where the processing of low-level visual information occurs. The fundamental nature of these early visual pathways means that they are often compromised by neurological disease. This thesis had two aims. First, it aimed to investigate changes in visual processing in response to Parkinson’s disease (PD) by using electrophysiological recordings from animal models. Second, it aimed to use functional magnetic resonance imaging (fMRI) to investigate how low-level visual processes are represented in healthy human visual cortex, focusing on two pathways often compromised in disease; the magnocellular pathway and chromatic S-cone pathway. First, we identified a pathological mechanism of excitotoxicity in the visual system of Drosophila PD models. Next, we found that we could apply machine learning classifiers to multivariate visual response profiles recorded from the eye and brain of Drosophila and rodent PD models to accurately classify these animals into their correct class. Using fMRI and psychophysics, found that measurements of temporal contrast sensitivity differ as a function of visual space, with peripherally tuned voxels in early visual areas showing increased contrast sensitivity at a high temporal frequency. Finally, we used 7T fMRI to investigate systematic differences in achromatic and S-cone population receptive field (pRF) size estimates in the visual cortex of healthy humans. Unfortunately, we could not replicate the fundamental effect of pRF size increasing with eccentricity, indicating complications with our data and stimulus.