An investigation of behavioral and electrophysiological effects of orthographic similarity on lexical processing

Five experiments investigated the effect of lexical orthographic similarity on behavioral and electrophysiological measures during the processing of individual words and pronounceable nonwords. The goal was to generate a new perspective for an area of research that has received considerable attention in recent years, but only moderate consensus (Andrews, 1997). Previous research has shown that similarity among lexical items can improve or hinder performance depending on the operational definition (e.g., neighborhood size vs. neighborhood frequency) and the task used (e.g., lexical decision, perceptual identification, or semantic categorization). Experiments 1 and 2 used a lexical decision task and a semantic categorization task, respectively, to examine the effect of lexical similarity defined in terms of neighborhood size. These two studies established a reliable event-related potential (ERP) lexical similarity effect while replicating existing behavioral findings. Negativity in the region of the N400 ERP component was greater for stimuli with more orthographic lexical neighbors than stimuli with few orthographic lexical neighbors. Experiments 3, 4, and 5 also used lexical decision and semantic categorization tasks. The goal of these studies was to extend the findings from Experiments 1 and 2 to another corpus of materials while simultaneously considering the additional factors of neighborhood frequency and word frequency. The only consistent result that emerged from these studies was an extension of the ERP neighborhood size effect established in Experiments 1 and 2. Word frequency and neighborhood frequency had less consistent results. The implications of these findings are then weighed against the functional significance of the N400 and frameworks of visual word recognition.