The role of tuning curve variables and threshold measures in the estimation of sensory cell loss.

Auditory-evoked potential tuning curves were collected at six frequencies before and 30 days after various noise exposures in 363 chinchillas using a simultaneous masking paradigm. Traditional bivariate and multiple linear regression/correlation analyses were performed in an effort to determine the extent to which sensory cell damage could be estimated from a knowledge of audiometric and tuning curve variables. The results showed strong correlations between percent outer hair cell (%OHC) loss and permanent threshold shift (PTS) and between %OHC loss and the tuning curve variables Q10 dB and high- and low-frequency slopes (SHF, SLF). The correlations were strongest between PTS and %OHC loss. However, the proportion of variability (r2) in %OHC loss attributable to variability in the predictor variable(s) (i.e., PTS) could be increased significantly by adding the Q10 dB of the tuning curve whose probe frequency was centered in the octave band length of the cochlea corresponding to the frequency at which the PTS occurred. The r2 values could be further increased by including audiometric and tuning curve variables from frequencies adjacent to the octave band being evaluated.