Design and Validation of a Hybrid Stimulus fMRI Paradigm for Simultaneous Retinotopy and Mapping of Reading-Selective Visual Cortex

Abstract

Reading selectively recruits specialized cortical patches within the left ventral occipitotemporal cortex (vOTC), such as the visual word form area (VWFA). Under the framework of the interactive account of word recognition, these neural responses emerge from dynamic, bidirectional loops where top-down linguistic predictions continuously modulate and constrain bottom-up sensory visual inputs. Consequently, identifying and isolating these top-down cognitive signals fundamentally requires a precise characterization of the underlying low-level sensory baseline. This dense-sampling fMRI study (11 right-handed adults, 10 sessions each) systematically addresses the spatial and functional architecture of the reading network using population receptive field (pRF) modeling and functional localizers (fLoc). First, we conduct a rigorous group- and subject-level replication of stimulus-dependent pRF eccentricity shifts (words and false fonts versus checkers) observed in previous research. To overcome known barriers to individual parameter stability, we systematically manipulate and evaluate stimulus Frequency, bar Width, and element Size (FWS) across eight sessions per subject to isolate the precise factors driving test-retest reliability. Second, we design and validate a novel Word-Center (WC) paradigm acquired across one/two sessions per participant. This hybrid paradigm was designed to decouple moving spatial mapping carriers (checkerboard bars) from central reading processes (word or false-font streams in the central fixation location). Functional data were denoised with NORDIC, preprocessed via fMRIPrep, and projected to the cortical surface for analysis via Nilearn GLMs to model word-responsive regions and mrVista to obtain pRF estimates. Individual-level spatial maps replicated previous results in approximately 70% of subjects. The lack of replication in the remaining subjects, whether driven by methodological factors or inherent individual differences, highlights ongoing baseline mapping challenges. Optimizing this mapping is sensitive to stimulus carrier type and specific FWS configurations. Furthermore, the hybrid WC paradigm successfully demonstrates stimulus equivalence, concurrently yielding robust reading-selective functional localization and reliable retinotopic estimates. By breaking the spatial-lexical confound of traditional mapping protocols, this paradigm provides an innovative framework to separate bottom-up sensory sweeps from top-down central linguistic processing. This vision-centric framework serves as a methodological baseline toward a comprehensive, multi-signal understanding of the human reading hierarchy.