School-age children’s development in sensitivity to voice gender cues is asymmetric

Speakers’ perceived voice gender is characterized by a number of acoustic features, but it is primarily defined by fundamental frequency (F0), related to the speaker’s glottal pulse rate, and vocal tract length (VTL), related to the speaker’s size. At a young age, children are already able to distinguish voices based on clear voice gender differences, but they are still less sensitive than adults to less salient differences in voice cues, such as prosody or accents. It remains unclear how children’s sensitivity to differences in voice cues develops over time. In the present study, we investigated whether children’s sensitivity to F0 and VTL differences and their weighting of F0 and VTL cues for voice gender categorization develop in parallel or if there is an asymmetry. We hypothesized that there may be a potential hierarchy in development as children generally focus more on global, dynamic acoustic cues, such as variations in F0, but may require more experience to make use of fine-grained, static acoustic information, such as VTL (Nittrouer, Manning, & Meyer, 1993). Furthermore, we investigated how children’s sensitivity to F0 and VTL relates to their weighting of F0 and VTL for voice gender categorization, as categorization may depend more on underlying representations of voice gender rather than sensitivity.

We tested fifty-eight children between the ages of 4 to 12 and fifteen adults. In experiment 1, we measured participants’ just noticeable differences (JNDs) in F0 and VTL via a 3-AFC procedure. In experiment 2, we studied participants’ weighting of F0 and VTL cues for voice gender categorization by manipulating the F0 and VTL parameters of an originally female speaker’s voice. Our results of experiment 1 showed that children’s sensitivity to VTL becomes adult-like around the age of 8, but their sensitivity to F0 still differs from adults at the age of 12. On the contrary, children’s weighting of F0 for voice gender categorization was more adult-like than their weighting of VTL but continued to differ from adult’s weighting at all ages. After correcting for a general effect of age, children’s weighting of F0 and VTL cues was weakly to moderately related to their sensitivity to F0 and VTL. Hence, there is asymmetry in children's development of sensitivity and weighting of voice cues related to voice gender, but both abilities seem to develop at a different pace, implying that perceptual sensitivity and categorization of voice cues rely on different underlying mechanisms.