Behavioral frequency tuning in the budgerigar. Published behavioral studies employing multiple stimulus paradigms indicate a peak in behavioral tuning sharpness from 3.5-4 kHz. Red triangles are based on forward masking (Kuhn and Saunders, 1980); all other symbols indicate simultaneous masking results (black circles: Saunders et al., 1978; blue squares: Saunders et al., 1979; black diamonds: Kuhn and Saunders, 1980; black crosses: Saunders and Pallone, 1980; magenta diamonds: Saunders et al., 1979; black square: Dooling et al., 2000). The thick solid blue line is a LOESS fit to the simultaneous-masking results with a smoothing parameter of 0.65. TC: tuning curve; NBN: narrowband noise.

Budgerigar stimulus-frequency otoacoustic emissions (SFOAEs) of 22 ears in 14 animals. (a) SFOAE level ranged from 5-15 dB at stimulus frequencies from 0.6-6 kHz, showing multiple spectral notches within this frequency range, and descended into the noise floor (gray shaded region) at higher frequencies. Thick colored lines highlight SFOAE level in four representative ears. (b) SFOAE phase decreased monotonically for higher stimulus frequencies. (c) SFOAE delay in stimulus cycles, NSFOAE, calculated from individual phase gradients in (b). The thick red line and shaded region show the weighted mean and 95% confidence interval, respectively (Gaussian weighting function; sigma: 0.25 octaves; 1,000 bootstrap repetitions; see also Figs 3 & 5 for illustration of NSFOAE confidence intervals). Data from the same representative ear are presented in the same color across panels.

Budgerigar auditory-nerve recordings from 127 fibers in 6 animals. (a) Representative tuning curves (colored lines) show the threshold for excitation of the neural response as a function of tone frequency. Tuning curves are V-shaped and are approximately symmetrical around CF on a log-frequency axis. Black crosses show the threshold at the characteristic frequency (CF) for all recorded fibers. (b) Tuning-curve Q10 increases for higher CFs across the neural population. The black trend line and gray shaded region show the weighted mean Q10 (Gaussian weighting function; sigma: 0.5 octaves) and 95% confidence interval (1,000 bootstrap repetitions), respectively. The thick red line shows NSFOAE replotted from Fig. 2.

Estimation of the SFOAE tuning ratio, r, from published chicken studies. (a) Q10 values of chicken auditory-nerve fibers (ANFs; n=289; crosses: Saunders et al., 1996; circles: Manley et al., 1991) increase monotonically for higher CFs, similar to the rate at which chicken NSFOAE (Bergevin et al., 2008; n= 9 animals) increases with frequency. Thick lines and shaded regions show weighted means (Gaussian weighting function; sigma: 0.5 octaves) and 95% confidence intervals (1,000 bootstrap repetitions), respectively; back lines extending from 3-4.6 kHz show extrapolated Q10 (dotted: mean; solid ±1SE; see text). (b) The tuning ratio, r, of chicken (magenta) compared to that of budgerigars (black). r is the quotient of auditory-nerve QERB over NSFOAE at the same frequency. Magenta lines extending form 3-4.6 kHz show extended-frequency r of chicken based on extrapolated Q10. r is marginally higher in the chicken than budgerigar, and trends downward with increasing frequency in both species.

Comparison of SFOAE predictions, behavioral frequency tuning, and directly measured auditory-nerve QERB in the budgerigar (n=127 fibers from 6 animals). The SFOAE prediction based on r from chicken (empirical range: thick magenta line; extended-frequency range [means ±1SE]: thin magenta lines) slightly exceeds the upper 95% confidence limit for mean auditory-nerve QERB (Gaussian weighting function; sigma: 0.5 octaves; 1,000 bootstrap repetitions) and shows the same frequency dependence. In contrast, behavioral QERB (thick blue line) shows a fundamentally different profile characterized by a peak from 3.5-4 kHz. Mean NSFOAE of budgerigars (thick red line) is replotted from Fig. 2.