Word recognition from a frame and fill perspective: Unraveling the impact of spatial frequency processing and prediction generation on reading proficiency

点击翻译按钮获取中文摘要

Given the enormous importance of good reading skills for all members of modern societies, it is necessary to understand the causes of inter-individual differences in reading ability beyond the field of learning disorders (e.g., dyslexia). Without ignoring the diversity of potentially influencing factors, this research project focuses on differences in visual perception and their influence on reading ability. Building on an established neurocognitive model for visual object recognition, the Frame and Fill model (FaF) by Moshe Bar, it will be investigated to what extent reading skill depends on two inter-related factors, namely, (1) on differences in processing high and low spatial frequencies (number of dark and bright edges per visual angle), and (2) on differences in using low-spatial frequency information for quickly generating predictions about word identities. A generalization of FaF to the word recognition domain might explain reading performance through a general model of perception and therefore generate new, explicitly verifiable hypotheses for reading research. Its specific assumptions about cortical localizations and the time-course of individual processing steps enable multi-modal testing of FaF on several levels of the word recognition domain. Research objective one comprises large-scale screening tests to identify individuals with preferences for low or high spatial frequencies using so-called Navon stimuli (i.e., multiple local, high-frequency stimuli are arranged to form global, low-frequency stimuli). Furthermore, the influence of spatial frequency preference on inter-individual differences in word recognition speed and accuracy is investigated through an innovative contour-priming task, in which low-spatial frequency components (word contours) should promote the generation of word identity predictions. Objectives two and three investigate the neuronal underpinnings of spatial frequency processing and contour-driven prediction generation in participant groups of different reading profiles. The high temporal resolution of electroencephalography will be used to investigate the exact time-course of prediction generation and subsequent processing facilitation postulated by FaF. The high spatial resolution of fMRI , in turn, will allow verifying the central FaF-hypothesis, that low-frequency-based predictions of object identity originate in the orbitofrontal cortex.