Using the N1 print tuning effect to predict poor reading in Chinese in Hong Kong children

Abstract

This study examined how we can use N1 to predict group membership of poor readers among 46 Hong Kong children (Age: M=8.26, SD=0.72 years old). An offline behavioral reading fluency measure was used to define the grouping; the children who performed in the bottom 15% (after adjusting the effect based on age) in a reading fluency task were defined as the poor readers (n=7), and the remaining children were defined as the typically developing readers (n=39). Electroencephalogram was recorded during a Chinese character lexical decision task, where the children were asked to press a button whenever they saw a real character, and press another button whenever they saw anything that was not a real character. There were four experimental conditions each with 60 trials, including real character, pseudo character, noncharacter, and stroke combination. The N1 in both the left and right occipital-temporal areas were used to predict group membership. Results indicated that the overall correct prediction was at 80.4%, with sensitivity at 71.4% and specificity at 82.1%. Among all the N1 contrasts, the N1 lexicality contrast (real character minus pseudo character) provided the highest odd ratio = 2.51 to the model. These results indicate the possibility of using the N1 to predict poor reader status.