Visual influences on speech sound discrimination: A parametric fMRI study

Vera Blau, Dept. of Cogntive Neuroscience, Maastricht University, the Netherlands; Maastricht Brain Imaging Center (M-BIC), the Netherlands

Abstract
The superior temporal cortex is activated during integration of letters and speech sounds [1] and has recently been implicated in the visual enhancement of speech intelligibility in the presence of acoustic masking noise [2]. The present fMRI study [N=9] used degraded auditory and visual input to unravel the relative contributions of visual letters on speech sound discrimination.

Speech sounds were presented in isolation or with matching or non-matching letters (congruency factor). Speech sounds were degraded using pink noise enabling non-optimal identification; letters were degraded using a line mask corresponding to good, medium and low perceptibility of the visual letter (visual noise factor). Subject discriminated between speech sounds /a/ and /e/.

Simultaneous presentation of a congruent visual letter facilitated subjects’ speech sound discrimination performance compared to unimodal auditory conditions, whereas incongruent letters increased reaction time. Only well-perceivable visual letters in congruent presentations led to increased activation in superior temporal areas. In lateral occipito-temporal regions the difference between congruent and incongruent combinations of letters and speech sounds scaled parametrically with the levels of visual noise.

The present study reveals a systematic influence of audiovisual congruency on the auditory discrimination of speech sounds in superior temporal and in occipito-temporal areas.

References:
[1] Van Atteveldt et al. (2004), Neuron, 43, 271-282.
[2] Callan et al. (2003), NeuroReport, 14, 2213-2218.

Not available

Back to Abstract