Revisiting superoptimal perceptual integration for pitch at high frequencies

Abstract

In a frequency discrimination task, Lau et al. [J. Neurosci. (2017)] reported superoptimal integration of information from individual components in a very high frequency region, where phase locking to the temporal fine structure is presumably absent, when the components were combined into a harmonic complex. We tried to replicate this finding using stimuli identical to those of Lau et al., with some additional conditions. Using an adaptive two-alternative forced-choice procedure, we measured fundamental frequency difference limens (F0DLs) for complex tones containing harmonics 6–10 with F0s of 280 and 1400 Hz, and frequency difference limens (FDLs) for each harmonic of the complex presented alone. Stimulus duration was 210 and 1000 ms. All tones had random phases, a ±3 dB level rove, and were presented in a continuous threshold-equalizing noise that was either diotic or dichotic. Observed DLs were lower overall than in the study of Lau et al. As in their study, for the low F0, observed F0DLs were worse than predicted assuming optimal combination of frequency information from the individual harmonics and assuming that performance is limited by peripheral noise. However, for the high F0, observed and predicted F0DLs did not differ significantly, in contrast to the finding of Lau et al.