Poster B47, Thursday, August 16, 3:05 – 4:50 pm, Room 2000AB

Does grapheme knowledge or phonological awareness determine detailed speech perception in preliterate children?

Anne Bauch1, Claudia K. Friedrich1, Ulrike Schild1;1University of Tuebingen, Germany

Literacy acquisition might modulate implicit aspects of speech recognition: Reading children might exploit more acoustic detail and/or might command more fine-grained word form representations than children who are not able to read yet (Schild, Röder, & Friedrich, 2011). Here we test whether letter knowledge or phonological awareness mediate more elaborated speech processing. German preschoolers (5 to 6 years old) participated in three different types of training: One group received a phonological-only training, in which we instructed participants about how selected German phonemes sound and vary. A second group with the same phonological training additionally learnt about the letters that correspond to these selected German phonemes (phonological-orthographic training). A control group received a training on early numerical skills. All children completed tests on explicit phonological awareness and letter knowledge before and after the training. After intervention, children additionally conducted an auditory word onset priming experiment. Prime-target combinations either matched in the initial phoneme (Identity condition, e.g., “Ki - Kino”), differed in the initial phoneme’s voicing feature (Variation condition, e.g., “Gi - Kino”) or were unrelated (Control condition, e.g., “Ba - Kino”). We recorded children’s lexical decision latencies and event related potentials (ERPs) to the target words. We expected training effects for both phonological groups in the explicit phonological awareness measurements, and a training effect for letter knowledge in the combined phonological-orthographic group. In general, we took facilitated responses and ERP amplitude differences (compared to the control condition) as indices of facilitated lexical access exerted by the primes. In particular, we took responses to target words in the variation condition as indicating the phonemic detail that listeners use for lexical access. If phonological awareness modulates implicit aspects of spoken word recognition, mismatching phonemic detail should restrict lexical access in both phonological training groups. If letter knowledge modulates implicit aspects of spoken word recognition, mismatching phonemic detail should restrict lexical access especially in the phonological-orthographic group. While all groups improved over time in explicit phonological awareness and letter knowledge, both phonological training groups gained more scores in a standardized phonological awareness test compared to the control group. Moreover, the phonological-orthographic training group gained more knowledge in trained letters that the other two groups. Analyses of the response latencies revealed faster reaction times in the Identity condition compared to the Variation and Control condition for all groups. ERPs showed facilitated processing of the Identity condition compared to the Variation condition for both phonological groups but not for the control group. Both phonological groups differed in ERP morphology: Children receiving phonological-orthographic training showed left-anterior amplitude differences. Taken together the results indicate that phonological awareness rather than grapheme knowledge might modulate the accuracy of speech processing. However, grapheme knowledge might contribute to the development of adult-like left-lateralized implicit speech perception. Schild, U., Röder, B., & Friedrich, C. K. (2011). Learning to read shapes the activation of neural lexical representations in the speech recognition pathway. Developmental Cognitive Neuroscience, 1, 163‐174.

Topic Area: Language Development