| || |
|Title:||BRAIN MECHANISMS OF SPEECH PROCESSING IN NEWBORNS: TWO OPTICAL IMAGING STUDIES|
| || |
|Authors/Affiliations:||1 Judit Gervain*; |
1 Department of Psychology, UBC;
| || |
|Content:||Objectives: Behavioral results (Marcus et al. 1999 etc.) show that young infants can extract basic structures from the language input. Another line of research, behavioral and neurophysiological (Ramus et al. 1999; Pena et al. 2003), suggests that neonates exhibit efficient perceptual learning abilities. As a synthesis, we carried out two near-infrared spectroscopy (NIRS) experiments to investigate how early the precursors of syntax acquisition appear.|
Materials and Methods: Neonates’ brain activity was measured in two experiments (N1=20, N2=22) using NIRS. Both experiments compared responses to two artificial languages, a test language with a predefined structure and a control language with no structure, but otherwise similar to the test language. In Experiment 1, the control language generated trisyllabic sequences with no internal structure (e.g., ABC: “mubage”), whereas the test language consisted of trisyllabic sequences with an immediate repetition (ABB: “mubaba”). In Experiment 2, the control was identical, while the test language had an ABA structure, i.e. a non-adjacent repetition. The two languages were played to newborns in 18-second-long blocks. Fourteen blocks were created for both languages, and were presented in an intermixed, pseudo-random order.
The optical sensors were placed on the frontal and temporal areas of both hemispheres (12 channels/hemisphere), measuring changes in oxygenated hemoglobin concentrations. To analyze the data, we conducted an analysis of variance with factors Grammar (test/control), Hemisphere (LH/RH) and ROI (temporal/frontal).
Results: In Experiment 1, we obtained a significant main effect of Grammar (F(1,19)=5.516, p=0.030), due to a larger activation for the ABB grammar overall. The interactions Grammar X ROI (F(1,19)=6.321, p=0.021) and Hemisphere X ROI (F(1,19)=6.603, p=0.019) were also significant. The interaction Grammar X Hemisphere showed a trend to significance (F(1,19)=3.094, p=0.095). We also analyzed the temporal evolution of responses during the course of the experiment. In an ANOVA with factors Grammar (test/control) X Time (beginning/end), we obtained a significant main effect of Grammar (F(1,21)=7.174, p=0.015). There was a significant interaction between Grammar and Time (F(1,21)=6.136, p=0.023). This suggests that the newborn brain detects immediate repetitions in speech, as indicated by the differential activation to ABB. The increase of this response over time is interpreted as a memory trace for the abstract pattern.
In Experiment 2, we obtained a significant effect of ROI (F(1,21)=9.506, p=0.006) due to greater activation in the temporal than in the frontal areas. No other main effects or interactions were significant. These results suggest that non-adjacent repetitions are not detected as perceptual Gestalts by the newborn brain.
Conclusion: Our findings imply that the perceptual system may play an important role in learning the structural regularities of language. For neurodevelopment, the results suggest that the newborn brain shows some functional specializations characteristic of the mature brain.
| || |