Evidence of compensative listening for speech segregation in hearing-impaired adults

Hearing-impaired (HI) individuals vary widely in their ability to segregate the speech of a conversation partner from others speaking at the same time. The role of hearing loss can be difficult to gauge because reduced sensitivity to relevant speech cues may be compensated for by changes in the relative reliance listeners place on those cues. Here, we provide an example. Thirteen NH and 15 HI adults (PTA thresholds >25 dBHL from 0.25–8.0 kHz) heard natural recordings over headphones of two talkers speaking sentences concurrently. Talker B (distracter) was present on each trial, the other talker (target) was equally likely to be A or C differing from B and each other in fundamental frequency (F0) and azimuthal location (θ). The listeners’ task was to identify the target as A or C. Relative cue reliance, R = cF0/(cF0 + cθ), was determined from regression coefficients, c, relating the listener's response to trial-by-trial perturbations simulating natural variation in F0 and θ. Relative cue sensitivity, S = d′F0/(d′F0 + d′θ), was determined from listener performance, d′, for F0 and θ cues presented individually. The results show a significantly greater range of S and a positive correlation of S with R for HI compared to NH listeners. [Work supported by NIH R01 DC001262-31.]