Psychophysical tuning curves in forward masking

Abstract

Using a two-interval forced-choice detection paradigm with fixed signals and variable maskers, we have obtained psychophysical analogs of neural tuning curves. The signal was a 20-msec burst of a 1000-Hz sinusoid, presented at several fixed levels and temporal positions relative to a 200-msec pure-tone masker. Masker intensity was varied according to the PEST procedure to determine the masker intensity required for 75% correct detections of the signal. This required masker intensity was measured as a function of masker frequency to produce the so-called “tuning curves.” With a low-level signal (10 dB SL) presented 10–30 msec after termination of the masker, the tuning curves we obtained were quite similar to those from other psychophysical and physiological experiments, showing the usual asymmetry and values of Q10 dB from 5 to 7 depending on the subject. As signal level was increased (to 30 dB SL) or as the masker-signal interval was increased, the tuning curves become very broad (Q10 dB ≈ 2). When the signal was presented simultaneously with the masker (30 msec before masker termination), the high-frequency skirt of the tuning curve became considerably shallower, but the low-frequency skirt was unaffected. This latter result, we believe, reflects lateral suppression of the signal by the masker. [This research was supported by Grant NS 12045 from NIH.]