Speech discrimination ability and its relation to psychophysical tuning curves (PTCs)

Abstract

PTCs were obtained from normal and hearing-impaired listeners using a simultaneous masking paradigm. All impaired listeners presented fiat moderate sensorineural hearing losses with excellent (90%) speech discrimination ability (NU-6) in quiet. Probe stimuli were presented at 10 dB SL for impaired listeners and were varied from 10–60 dB SL for normal subjects. PTCs were quantified in terms of Q10 and tip-to-tail intersect. Performance-intensity functions for monosyllabic words were obtained as a function of signal-to-noise ratio for broadband and low-pass (500 Hz) filtered noise. Normal listeners demonstrated a decrease in tip-to-tail intersect and an increase in Q10 with increasing probe level. Regardless of whether comparisons were made across constant SL or constant SPL, impaired listeners exhibited smaller tip-to-tail differences than normal subjects with less obvious changes noted in Q10. While speech discrimination performances in broadband noise was similar for both groups, performance in low-pass noise was poorer for the impaired group. The significance of changes in PTC shape in the presence of hearing loss will be discussed.