Objective Most rhyme awareness assessments do not encompass measures of the global similarity effect (i.e., children who are able to perform simple rhyme judgments get confused when presented with globally similar non-rhyming pairs). The present study examines the neural nature of this effect by studying the N450 rhyme effect. Methods Behavioral and electrophysiological responses of Dutch pre-literate kindergartners and literate second graders were recorded while they made rhyme judgments of word pairs in three conditions; phonologically rhyming (e.g., wijn-pijn), overlapping non-rhyming (e.g., pen-pijn) and unrelated non-rhyming pairs (e.g., boom-pijn). Results Behaviorally, both groups had difficulty judging overlapping but not rhyming and unrelated pairs. The neural data of second graders showed overlapping pairs were processed in a similar fashion as unrelated pairs; both showed a more negative deflection of the N450 component than rhyming items. Kindergartners did not show a typical N450 rhyme effect. However, some other interesting ERP differences were observed, indicating preliterates are sensitive to rhyme at a certain level. Significance Rhyme judgments of globally similar items rely on the same process as rhyme judgments of rhyming and unrelated items. Therefore, incorporating a globally similar condition in rhyme assessments may lead to a more in-depth measure of early phonological awareness skills. Highlights Behavioral and electrophysiological responses were recorded while (pre)literate children made rhyme judgments of rhyming, overlapping and unrelated words. Behaviorally both groups had difficulty judging overlapping pairs as non-rhyming while overlapping and unrelated neural patterns were similar in literates. Preliterates show a different pattern indicating a developing phonological system.