Interpreting the role of suppression in psychophysical tuning curves.

We describe two models which make different assumptions about the factors determining threshold for psychophysical tuning curves (PTCs) determined in simultaneous masking. The models make different predictions about the amount of activity evoked by off-frequency maskers in the frequency region of the signal. The predictions were tested by using a forward-masking technique to estimate the amount of this activity. The data support a model which assumes that threshold in simultaneous masking corresponds to a "swamping" of the signal activity by that of the masker, rather than a suppression of the low-level signal to its absolute threshold. To explain the fact that PTCs in forward masking are sharper than those in simultaneous masking, the model requires that suppression acts to sharpen the excitation pattern of the masker. This sharpening is not seen in simultaneous masking, since suppression in a given frequency region affects both signal and masker.