Abstract
Neurons within early visual cortex are selective for basic image statistics, including the spatial frequency content of a retinal image. However, sensitivity is not uniform across all frequencies, peaking at mid-frequencies, and dropping off for both low and high frequencies. How does the window of spatial frequency sensitivity vary across the visual field and across visual areas? Although a handful of previous studies have investigated this using conventional fMRI designs and analysis methods, these measurements are time-consuming and often do not span the entire range of spatial frequencies. In this study, we introduce a model-based fMRI analysis approach that allows for fast and efficient estimation of population spatial frequency tuning (pSFT) for independent voxels. BOLD responses within early visual cortex were acquired while subjects viewed a series of full-field stimuli that swept through a large range of spatial frequency content. Each stimulus was generated by bandpass filtering white noise with a central frequency that changed periodically between a minimum of 0.5 cpd and a maximum of 12 cpd. To estimate the underlying frequency tuning of each voxel, we assumed a Gaussian pSFT and optimized the parameters of this function by fitting our model output with the measured BOLD time series. With these estimated parameters, we can investigate the relationship between spatial frequency selectivity and other factors, including retinotopic preference and receptive field size.
Meeting abstract presented at VSS 2018