| Popis: |
Immense strides in furthering our understanding of a variety of underlying mechanisms involved in visual perception are continuously being made, however numerous questions in the field remain unanswered. Detailed here are two studies aimed at elucidating the functional processes in which the human visual system integrates visual stimuli into cohesive representations of the world around us. First, using simulations and fMRI experiments we examined how retinotopic mapping paradigms can be optimized to most efficaciously and accurately measure the underlying functional and organizational properties of early to mid-level visual cortex in both healthy and clinical populations. We focused on quantitatively comparing two common retinotopic mapping designs in terms of measurable and achievable signal-to-noise ratios under various conditions such as exogenous noise and scanning time. While multifocal retinotopy can yield a higher signal-to-noise ratio in primary visual cortex, we find that the phase-encoded technique provides qualitatively better retinotopic maps despite having a less robust point-to-point correspondence of visual field location to visual cortex. Other implications of employing these methods are also discussed.Next, fMRI experiments aimed at uncovering the spatial frequency tuning profiles of higher-level, category selective areas of cortex such as the parahippocampal place area (PPA), retrospenial cortex (RSC), occipital place area (OPA), lateral occipital cortex (LOC) and the fusiform face area (FFA) were performed. These experiments utilized both frequency filtered scenes belonging to four different categories along with unstructured stimuli of various spatial frequencies. Results from univariate, multivariate and searchlight analyses are subsequently provided and elaborated on, with the general finding that the category selective areas mentioned above are sensitive to high spatial frequencies, especially when semantic information is present. |
