Predicting speech-in-noise ability in normal and impaired hearing based on auditory cognitive measures

Abstract

Problems with speech-in-noise (SiN) perception are extremely common in hearing loss. Clinical tests have generally been based on measurement of SiN. My group has developed an approach to SiN based on the auditory cognitive mechanisms that subserve this, that might be relevant to speakers of any language. I describe how well these predict SiN, the brain systems for them, and tests of auditory cognition based on them that might be used to characterise SiN deficits in the clinic.

Keywords: auditory cognition; behaviour; brain; cortex; speech in noise.

FIGURE 1

Figure-ground stimulus. The “ground” stimulus contains random elements distributed in frequency-time space. At a certain point in time, shown by the first vertical dashed line, a certain number of elements (four here–shown by arrows) are constrained to remain constant from one time frame to the next. If there are enough elements and they are on for long enough a “figure” emerges perceptually from the background.

FIGURE 2

Correlation between performance for the figure ground task (SFG, stochastic figure ground) and a widely used US measure of sentence-in-noise perception, AzBio. The data are for 47 cochlear-implant users (see text).