Research Methodology
Combining a number of well-established speech production and perception tasks with the latest scanning techniques, we examined the acoustic and perceptual correlates of the relevant categorical linguistic and gradient paralinguistic information in the speech signal, and we identified the brain systems that are involved in processing the different types of intonation.
I. Production studies (cf. Post and Nolan 2011)
- Computer-based analysis of a range of acoustic parameters in tightly controlled laboratory experiments in which position (metrical, discourse, morpho-syntactic, etc) and/or function (e.g. linguistic vs. paralinguistic) are systematically varied while controlling for confounding factors like segmental structure (Post 2011a, 2011b, Post et al. in preparation a., Zellers et al. 2009, Zellers and Post 2010, 2011)
- Auditory analysis of speech corpus data testing a model derived from quantitative data analysis (as in 1) in semi-spontaneous and unscripted speech (Delais-Roussarie et al. 2011)
- Functional Data Analysis (FDA) is used to model systematic variation in acoustic parameters reflecting different intonational functions (here: sentence modality, discourse function, and paralinguistic meaning,). FDA is a statistical analysis technique which uses as input entire curves represented in terms of functions by means of Principal Component Analysis (PCA). Functional PCA allows the extraction of a compact description of global variations in F0 or durational patterns in a theory-neutral way (Zellers et al. 2010, Post et al. in preparation a.)
- A multiple-methods approach combining qualitative and quantitative acoustic and auditory analysis of discourse topic structure cues (Phonetics of Talk-in-Interaction, and what we termed Empirical Phonology; Zellers and Post 2012)
- Further development of transcription systems for segmental and prosodic information which integrate project insights relating to intonational segmentation, abstraction, and categorisation (Delais-Roussarie and Post 2011, 2013, applied in e.g. Delais-Roussarie et al. 2011, under review; ongoing collaboration)
II. Perception studies
- Identification and categorisation of linguistic vs. paralinguistic function in speeded judgement task (reaction time experiment) systematically varying utterance-final F0 contour (Post et al. 2013, in preparation a, b)
- Online comprehension of discourse function (topic hold vs. topic change) in speeded judgement task systematically varying F0 fall size (Zellers and Post 2010, Zellers 2011)
III. Neurobiological studies
- Functional Magnetic Resonance Imaging experiment (fMRI) using speeded judgement categorisation task (cf. above) in event-related design (Post et al. 2009, 2012, 2013, Bohr et al. under revision), including connectivity analysis (Bohr et al. under revision; Stamatakis et al. 2011)
- Electroencephalographic experiment (EEG) using discrimination of variation in F0 contour by intonational function (linguistic vs. paralinguistic) in passive listening (Post et al. in prep. a.).
The research is innovative in its methodology, since it combines acoustic analysis, statistical modelling, and auditory analysis with fMRI, EEG, and perception tests, as well as developing or further enhancing existing analysis techniques (multiple methods PTI & EP, FDA, and intonational transcription). In this way, the project has provided the tools to develop a unified model of prosody which can integrate insights from different research disciplines (i.e. phonetics, linguistic typology, cognitive psychology, and neurobiology), and it offers a template for further research.