The fundamental frequencies of our own voice

Abstract

Own actions send a corollary discharge (CD) signal, that is a copy of the planned motor programme, to sensory-specific brain areas to suppress the anticipated sensory response, providing a neural basis for the sense of self. When we speak, the sensory consequences of the fundamental frequency ( $f_{\text{0}}$ ) of our own voice, generated by vocal fold vibrations, are suppressed. However, due to bone/air conduction filtering effects, the $f_{\text{0}}$ we self-generate is measurably different from the $f_{\text{0}}$ we subjectively perceive as defining our own voice. Using an auditory change deafness paradigm, we parametrically tested the sensitivity to auditory change in the frequency neighbourhoods of objective and subjective own voice pitches and found that participants experience change deafness for both to a similar extent, relative to a control pitch condition. We conclude that when we listen attentively, we are likely to filter out small pitch changes in the vicinity of our own objective and subjective voice $f_{\text{0}}$ , possibly as a long-term consequence of speaking-induced suppression mechanisms integrated with individual, perceptual bodily priors.

Keywords: attentive filters; auditory attention; change deafness; corollary discharge; self-generation; subjective perception.

Figure 1.

Participants sorted ascendingly by the objective pitch estimate. (left) The objective and subjective pitch estimates per participant (n = 27). (middle) The same plot for the filtered sample (n = 22). (right) The difference between both estimates per participant scales up, albeit only moderately, with the measured pitch estimates. The dichotomy (high and low clusters) codes for participants' reported sex. The estimates where measured in Hz and converted to the Mel scale to simplify comparisons in this figure.

Figure 2.

Change detection hit rate for each deviation step per pitch condition, averaged across participants. Shaded area is the s.e.m. (left, middle) Average hit rate broken down by participants’ reported sex. (right) Average hit rate for the whole sample.

Figure 3.

Change detection hit rate for each deviation step per pitch condition, averaged across participants. Shaded area is the s.e.m. (left, middle) Average hit rate broken down by the relative position of the subjective pitch estimate to the objective pitch estimate; Subjective is higher vs Objective is higher. (right) Average hit rate for the whole sample.

Figure 4.

Fundamental pitch estimates in Hz for each condition (reading, speaking and vowel) from both PDAs per participant.

Figure 5.

(left) Correlation plot of Praat's and HPT's $f_{\text{0}}$ estimates of all three voicing conditions (reading, speaking and corrected vowel) across participants. (right) Bland–Altman LOA plot with 95% CIs. The difference between HPT and Praat for a given participant in a given condition is plotted on the y-axis against the average of both estimates on the x-axis.

Figure 6.

(left) Correlation plot of Praat's & HPT's $f_{\text{0}}$ estimates of the corrected vowel a: condition. (right) Bland–Altman LOA plot with 95% CIs. The ordinate and abscissa represent the same as in figure 5.

Figure 7.

Fixed-effects estimates and their interactions of the model $m{5}_{RS2}$ with 95 and 90% CIs, indicated in thin and thick lines, respectively.

Figure 8.

Random-intercept and random-slope variance per pitch condition within each subject, referenced to the control pitch condition.

Figure 9.

The estimated marginal means as computed from the final model $m{5}_{RS2}$. The plot shows the estimates for each pitch condition in every deviation magnitude. The bars indicate asymptotic 95% CIs. Both the estimates and the intervals are shown on the response scale (hit rate), back-transformed from the logit scale with bias-adjustment.

Figure 10.

(left) Example of the drops in Praat-generated pitch contours of the first vowel recording from participant no. 2 (before manual correction). (middle) The same vowel after manual contour correction. (right) Example of multiple contour drops form in the third vowel recording from participant no. 9.

Figure 11.

(left) Correlation plot of Praat's & HPT's $f_{\text{0}}$ estimates of all three voicing conditions (reading, speaking and vowel) across participants. (right) Bland–Altman LOA plot with 95% CIs. The ordinate and abscissa represent the same as in figure 5.

Figure 12.

(left) Correlation plot of Praat's and HPT's $f_{\text{0}}$ estimates of the reading condition. (right) Bland–Altman LOA plot with 95% CIs. The ordinate and abscissa represent the same as in figure 5.

Figure 13.

(left) Correlation plot of Praat's and HPT's $f_{\text{0}}$ estimates of the speaking condition. (right) Bland–Altman LOA plot with 95% CIs. The ordinate and abscissa represent the same as in figure 5.