Human discrimination and modeling of high-frequency complex tones shed light on the neural codes for pitch
Fig 5
Comparison of simulated autocorrelograms at low and high frequencies for the DBL stimulus.
(A) Autocorrelograms for simulated high-spontaneous-rate auditory-nerve-fiber responses [52] for the low-frequency DBL stimulus (left column) and the high-frequency DBL stimulus (right column). Here, one masker F0 was set to 5.5 ST below the target F0 while the other masker F0 was set to 6 ST above the target F0. The figure was constructed by fixing the target F0 and then varying the simulated CF (y-axis) linearly over a fixed range. For each CF, the instantaneous firing rate of the auditory-nerve model was computed and then the autocorrelation of that response was computed. Color (from purple/blue [low] to yellow/green [high]) indicates the autocorrelation value at each lag-CF point, analogous to color indicating intensity in a spectrogram. In each column, the first four panels show autocorrelograms for varying TMRs and the bottom panel shows the summary autocorrelation function (sACF; the sum of the autocorrelograms along the CF axis). Red boxes and labels indicate the zoomed-in views plotted in the following subfigures. (B) A zoomed-in view of autocorrelograms and the sACF for 0, 5, and 10 dB TMR at low frequencies for lags of 3.5 to 4.5 times the period of the target F0. Vertical lines indicate lags corresponding to the periods of each tone in the stimulus or to multiples of those periods (marked using the shorthand “2F0” to refer to the lag twice as long as the period of the tone). Lags corresponding to multiples of the target period are indicated with “XF0”, while lags corresponding to multiples of the masker periods are indicated with “XF0L” and “XF0U” for the lower and upper maskers, respectively. (c) A zoomed-in view of autocorrelograms and the sACF for 5 and 10 dB TMR and without maskers at high frequencies for lags of 1.5 to 2.5 times the period of the target F0. Vertical lines indicate the same as in the previous subfigure. Only vertical lines corresponding to the target period are plotted in the last subpanel because this simulation was conducted without the maskers. See Materials and Methods for more details on the simulations.