Relating physiology to perception: The case of the notched-noise masker

Abstract

Sharpness of peripheral tuning is an essential factor in auditory signal processing. The notched-noise (NN) masking paradigm (Patterson, 1976, JASA) is an ingenious strategy used in numerous studies to estimate peripheral tuning bandwidths, filter shapes, and level-dependence in normal and impaired ears. Interpretation of NN results is based on the power-spectrum model of masking; however, the fact that NN thresholds are little influenced by a roving-level paradigm calls this interpretation into question (Lentz et al., 1999, JASA). Here, we explore neural cues for detection in NN in auditory-nerve (AN) and midbrain (inferior colliculus, IC) responses. Strong low-frequency fluctuations in AN discharge patterns associated with sharp spectral edges provide effective inputs for midbrain neurons tuned to low-frequency fluctuations. Addition of a tone target reduces the fluctuations in some frequency channels, creating a strong contrast in the fluctuation profile along the frequency axis. We explore the NN paradigm using computational models for AN and amplitude-modulation (AM) tuned IC neurons. Recordings from band-pass AM tuned neurons in the IC of awake rabbit support the hypothesis that the midbrain response profile can explain perception in the NN paradigm. Interpretation of NN masking results should include not only peripheral tuning but also central processing mechanisms.