Abstract
Receptive fields in early visual cortex change systematically with eccentricity and visual area. Model-based functional neuroimaging approaches provide a method to map and interrogate population receptive fields (pRF) and spatial frequency tuning (pSFT), though how closely the response properties derived from these methods correspond with one another remains unclear. Here, we compared pRF maps to pSFT estimates in early visual cortex (V1 – V3). Preferred spatial frequency and bandwidth of voxel-wise pSFT were estimated with a log Gaussian function (see Aghajari, Vinke, & Ling, 2020). We then compared pSFT to receptive field sizes that were derived from a separate fMRI pRF mapping experiment. Our results confirm an inverse relationship between peak population spatial frequency preference and receptive field size that was strongest in V1 but persists across V2 and V3 (p<0.001). In V1, we found that spatial frequency preference decreases by -0.38 cycles per degree (cpd) for every degree of receptive field size. In V2 and V3, however, spatial frequency preference decreases by -0.23 and -0.27 cpd for every degree of receptive field size, respectively. Conversely, bandwidth increases with receptive field size in V1 (0.14 octaves per degree of receptive field size). And in V2, this relationship weakens (0.07 octaves/degree), while in V3 it virtually remains unchanged (-0.009 octaves/degree). This work provides evidence that spatial frequency preference is inversely proportional to receptive field size across visual area, while bandwidth increases with receptive field size only in primary visual cortex.