Katharina von Kriegstein



Human Communication

When different speakers say the same sentence, the speech content is the same while the acoustic signals can be very distinct. These acoustic differences are introduced by inter-individual variability in vocal tract anatomy (e.g., the difference in the vocal tract size between speakers). Acoustic variability between speakers allows humans to recognize each other by voice (speaker recognition). However, for speech recognition, acoustic variability makes sophisticated processing necessary. My research has the goal to investigate (i) the brain mechanisms that exploit speaker variability for speaker recognition, and (ii) the brain mechanisms that deal with speaker variability in order to extract the communicated message, i.e. speech recognition.

Speech and speaker recognition are inherently auditory tasks. However, typically, people talk to us face-to-face, resulting in concurrent and correlated auditory and visual input. Thus we expect a taller speaker to sound different than a smaller speaker and we expect that specific lip-movements are associated with corresponding speech-sounds. It is likely that the brain has evolved over somatic and evolutionary timescales to use these audio-visual correlations in order to optimize behaviour. Therefore, my aim is, besides investigating auditory aspects of inter-speaker variability, to investigate how visual information, over different speakers, interacts with auditory processing.

To address these aims I am using methods of systems neuroscience (psychophysics, functional and structural magnetic resonance imaging). The studies involve participants with selective perceptual deficits as well as participants without deficits.