Until the mid 1990s, psycholinguistic models of sentence processing relied almost exclusively on reading data and did not take into account the intonation and phrasing of spoken language. Today we know that prosody is immediately encoded and often guides the listeners' interpretation of sentences. In his 1999 paper in Nature Neuroscience, Dr. Steinhauer showed that ERPs provide an ideal tool to study the exciting interactions between prosody and syntax and, moreover, that the human brain elicits a specific neural response when processing prosodic boundaries: the closure positive shift (CPS). In fact, the CPS response is also elicited during silent reading (e.g., at comma positions), supporting the notion that written text is covertly recoded into speech-like representations. What we don't know yet, is whether these important findings hold universally across languages and across populations. For example, languages with multiple tone levels (e.g., Chinese) and/or different writing systems may show distinct patterns.
Another unresolved issue is which acoustic parameters are relevant to elicit the CPS: Speakers use pitch changes (boundary tones), vowel lengthening and pause insertion to mark boundaries, but most previous ERP studies used combinations of all of these cues instead of teasing them apart.
A final issue concerns the language-specificity of the CPS. Previous research suggests that a CPS-like ERP component is also elicited at musical phrase boundaries, however only in trained musicians (Knoesche et al., 2005). This contrasts strongly to the CPS in speech prosody, which has been found even in pre-linguistic infants (e.g., Pannekamp et al, 2006). An import question with critical implications for theories of the interaction of cognitive systems is, therefore, if the 'music CPS' relies on the same mechanisms as the 'language CPS'. In the next five years, my prosody program will continue to address these issues.
Two PhD students in this program (Ms E. Pauker and Ms I. Itzhak) are currently preparing their ERP thesis projects on prosody/syntax interactions in young and elderly healthy subjects as well as patients with focal brain lesions. In addition, they will systematically manipulate boundary strength by varying acoustic cues (pitch, vowel duration) in the speech signal. This research will provide important new insights into the interaction of language mechanisms and their neural underpinnings.
Starting in Winter 2009, a postdoctoral fellow from Korea, Ms Hyekyung Hwang, will work in my lab and conduct the first ERP study on overt and implicit prosodic processing in Korean. By manipulating the length of noun phrases in structurally ambiguous sentences, she will change the preferred prosodic phrasing, which alters the syntactic interpretation of the sentence. This way, the expected CPS components should predict processing difficulties later in the sentence, both in speech and during silent reading. This innovative study, co-supervised by Drs White and Goad in Linguistics, will contribute unique data on the cross-linguistic replicability of the CPS response and the validity of linguistic models. The project is also a good example of how interdisciplinary work affects the development of traditional disciplines at McGill, as outlined in the University's Strategic Research Plan.
Music. Together with my colleague and co-PI, Tier 1 CRC Dr Palmer, I will continue to conduct ERP research on musical phrasing and the online processing of different types of musical accents (employing the 'timbre-change detection paradigm' we developed in my first CRC term). A particularly promising project involves the direct comparison of the 'language CPS' and the 'music CPS' within subjects, contrasting musicians and non-musicians. This work connects the neuroscience communities in language and music at McGill.