Bringing the Natural Auditory Environment into the Laboratory
Development of Speech and Sound Source Modelling Techniques for Cognitive Brain Research
Principal Investigator: PAAVO ALKU
Researcher: Santeri Yrttiaho
Laboratory of Acoustics and Audio Signal Processing, Helsinki University of Technology
This project will focus on the development of new digital signal processing (DSP) methodologies to model the production of speech in voice communication situations in natural environments. This will, firstly, make possible high-quality, yet controllable synthesis of complex acoustical stimuli to be implemented in non-invasive brain measurements in humans. Secondly, the methodologies to be developed will allow us to delineate what the most essential acoustical cues in different natural auditory environments are and how these cues can be quantified. The methodological work to be conducted approaches the challenging research task mainly from the perspective of computational modelling of voice production. In addition, we will study physical modelling of speech production and will combine modelling of three-dimensional acoustical space in speech messages.
Contact: paavo.alku(at)hut.fi, tel. +358 9 451 5680
Computational Modelling of Cortical Processes
Principal Investigator: PATRICK J.C. MAY
Researcher: Kevin I. Hynnä
Department of Psychology, University of Helsinki
Although speech processing is one of the central abilities of the human brain, very little is known about the underlying neural mechanisms. The project aims to reveal the computational principles underlying the neural processing of speech. Speech perception will be explored by combining computational modelling of neuronal dynamics with non-invasive brain measurements (EEG, MEG, fMRI). Firstly, we will describe how areas of the brain analyzing sound features react to speech sounds such as vowels. Secondly, we will describe how different areas of the brain interact with one another during the processing of continuous speech, such as spoken sentences. Special emphasis will be laid on delineating how the comprehensibility and semantic content of speech signals is reflected in the short-term functional re-organization of the human brain. In using computational modelling, the project aims to go considerably beyond descriptions of "where" and "when" brain areas are activated and to deliver a biologically realistic explanation of how humans process speech.
Contact: patrick.may(at)helsinki.fi, tel. +358 9 1912 9483
Brain Dynamics and Boundary Conditions of Speech Perception
Principal Investigator: HANNU TIITINEN
Researcher: Sakari Leino
Department of Psychology, University of Helsinki
The project will utilize non-invasive mass-action brain measures (EEG, MEG, fMRI) in studying how compression of speech modifies human brain dynamics and behavioural performance. Digital signal compression techniques will be used to transform speech from completely natural to unnatural, thus allowing us to monitor how the brain deals with the consequent changes in the comprehensibility and semantic content of speech. Both short- and long-duration stimuli (vowel sounds and sentences) are to be used in the experimentation. The project is expected to reveal hemispheric asymmetries in the spatiotemporal orchestration of brain areas extracting information from stimuli of varying compression levels, the boundary conditions of human speech perception, and their relation to behavioural performance. The project will provide directly relevant data for the construction of a computational model of human auditory cognition. The results of the project might also provide solutions for applied research and technological innovations in, for example, telecommunications.
Contact: hannu.tiitinen(at)helsinki.fi, tel. +358 9 1912 9482