Growth of loudness in sensorineural impairment: Experimental results and modeling implications

Several modifications of Stevens' power law have been proposed to account for the low-level curvature in the loudness function in quiet and in noise. This chapter extends these modeling applications to cochlear impairment. Loudness functions are generated in cochlear-impaired hearing by a battery of psychophysical procedures. The results are compared to loudness values determined from three proposed power-law modifications: L = k(X - X-t)(n) (equation 1), L = k(X-n - X-t(n)) (equation 2), and L = k[(X-2 + cX(t)(2))(n) - (cX(t)(2))(n)] (equation 3), where X is the stimulus intensity in equations 1 and 2, and X-2 is proportional to the stimulus intensity in equation 3. Unlike the two more widely employed modifications expressed by equations 1 and 2, both the bandwidth of the presumed intrinsic or extrinsic masking noise and the critical bandwidth at the stimulus frequency are incorporated into the structure of equation 3. Model predictions obtained for individuals and groups show that equation 3 more accurately describes the overall shape and slope of the loudness function in cochlear impairment than either equations 1 or 2. These findings provide additional evidence for the general applicability of equation 3.