Poster B46, Thursday, August 16, 3:05 – 4:50 pm, Room 2000AB
Developmental differences in the neural mechanisms supporting natural sentence comprehension
Julie Schneider1, Mandy Maguire1;1Callier Center for Communication Disorders, University of Texas at Dallas
Introduction: Language comprehension requires millisecond level processing of semantic and syntactic information, yet children seem to integrate and comprehend all of this information with relative ease. Although this is done effortlessly, developmental differences exist in the speed by which children process speech (Bates, Dale & Thal, 1995; Nippold, 1988). By understanding how variation in the developmental time-course of semantics and syntax may contribute to individual differences in language comprehension, we may lay a foundation to better understand how language develops in atypical populations (Thomas & Karmiloff-Smith, 2005). This study uses electroencephalography (EEG) to investigate how early school-age children, late school-age children, and adults process semantics and syntax in naturally paced sentences. Methods: Children ages 8-9 years 9 (Mage=9.1, SDage=.60), 12-13 years 13 (Mage=12.8, SDage=.58), and adults (Mage=22.2, SDage=3.9) listened to 160 semantically and syntactically correct and incorrect naturally-paced sentences and were asked to complete an acceptability judgment task. All sentences included an inanimate noun paired with a modal verb and action verb. Grammatically incorrect sentences included either the intrusion or omission of a present participle (-ing) form of the verb (i.e., will baking, will be bake). The semantic violation introduced an unsuitable pairing of actions with agents (i.e., hose-bake). Analysis: EEG data was epoched from 500 msec before to 1500 msec after the target verb onset. Only trials in which participants responded to correctly were included in the analysis. ERPs. The mean amplitude of the pre-stimulus interval (-100 to 0 msec) was subtracted from each time point in the post-stimulus interval. Single trials were averaged together to obtain a stable waveform ERP for each condition and each electrode for every subject. Time-Frequency. The mean ERSP was computed for all data channels (3-30 Hz) and a morlet wavelet was applied to each epoch. The mean baseline power at each electrode and frequency was subtracted. Statistical significance was determined using a monte-carlo permutation analysis similar to that used by Maris and Oostenveld (2007). Results & Conclusions: When processing a semantic error, there were no developmental differences in the N400; however, increases in theta, related to semantic processing, were greater for 8-9 year olds than 12-13 year olds and adults. Therefore, the N400 may be too gross a measure to identify more subtle aspects of semantic development that are ongoing in early school-aged children. Consistent with previous research, syntactic errors resulted in a larger P600 and greater beta decrease than correct sentences (Bastiaansen et al., 2010; Davidson & Indefrey, 2007; Hagoort, Brown & Groothusen, 1993), but the location of the P600 and the amplitude of beta decreases differed as a function of age. These findings suggest specialization of syntactic skills is ongoing through adolescence and children may recruit an alternative neural process to identify a syntactic error, which beta is not sensitive to. Discussion: Our findings suggest the neural substrates underlying semantic processing reach adult-like levels at a younger age, while syntactic skills develop over a protracted time course and may require recruitment of additional neural mechanisms to support comprehension of natural language.
Topic Area: Language Development