Figure 4

Decreased compensation for larger f_o shifts. (A) Schematic drawing of normalized distributions for the spontaneous f_o variability (upper) and shifted versions of its feedback after the introduction of + 50 and + 100 cent shifts (lower). Given a certain amount of variability, originally intended f_o will not overlap well with the distribution for large f_o shifts, i.e., the large f_o shifts will be outside of the motor exploration range. This can be expected to reduce the compensation ratio for that condition. (B) Percent amount of compensatory responses against different amounts of f_o shift (25, 50, and 100 cents). The vocal responses to positive f_o shifts were sign-inverted and averaged with that to negative shift conditions. Each dot indicates one individual participant. Error bar shows the standard error (n = 38). Asterisk (*) indicates statistically significant difference (p < 0.05 in post hoc comparisons with Tukey–Kramer correction). (C) Correlation between the compensation ratio for the 100-cent shift amount and the variability of the slow component. r shows Pearson’s correlation coefficient. Two participants who showed negative values in the compensation ratio were excluded as outliers. Asterisk (*) indicates statistically significant correlation (p < 0.05).