Spatial frequency channels implement a mental ruler in spatial vision

Perception of basic spatial properties (e.g., size, separation) varies with visual context, indicating a rescaling process in spatial vision. Previous findings have suggested that this rescaling is supported by an adjustable "mental ruler", an internal metric that can be flexibly changed by context. However, the neural implementation of this putative mental ruler remains unknown. We hypothesized that this mental ruler is represented by multiple spatial frequency (SF) channels with different tunings. In this account, the relative weighting of different SF channels sets the unit length of the mental ruler. Up-weighting of high-SF channels drives the concentration of neuronal receptive fields in early visual cortex, leading to a shorter unit length (a finer division) and perceptual inflation. Conversely, up-weighting of low-SF channels produces a longer unit length (a coarser division) and perceptual compression. Consistent with this account, we found that modulating the relative contribution of the high- and low-SF channels is coupled with a systematic distortion in perceived separation, a fundamental spatial property, and a global displacement of population receptive fields (pRFs) in primary visual cortex. Computational modeling further demonstrated that the perceptual distortion and the pRF displacements were quantitatively linked through SF channel modulation. Together, these results provide converging evidence for the neural implementation of an adjustable mental ruler and suggest a rescaling mechanism through which the visual system dynamically calibrates perceived spatial properties across different pictorial, image-based contexts.