Confirmation and rejection of the “precedence effect”

Abstract

A series of experiments designed to investigate the “precedence effect” in lateralization space were conducted using a two-alternative, forced-choice paradigm. Stimuli consisted of 40-μs transients separated by inter-click intervals (ICIs) of 0.38f–30 ms. In a two click paradigm (experiment 1) interaural differences of time (IDT) thresholds were determined independently at click positions 1 and 2. Thresholds measured at click 2 ranged from 12 μs at extended (over 5 ms) and short duration (less than 1 ms) ICIs to a peak of 250 μs at ICIs of 1.75–2.35 ms. Thresholds measured at click 1 were uniform across ICIs, constant at approximately 15–25 μs. Experiment 2 consisted of a three click paradigm where the ICI between clicks 2 and 3 was held constant at a value where subjects in experiment 1 had peaked (1.75–2.35 ms), while the ICI between clicks 1 and 2 was varied from 0.38 to 10 ms. IDT thresholds were measured for click 3. Results show a U-shaped function for which thresholds drop below 30 μs for click 3 at ICIs between 2–8 ms, with the effect returning at long and short duration ICIs. Experiment 3 consisted of click trains of n = 2, 3, 4, 5, 6, and 12 where ICI was held constant at subjects' peak value in experiment 1 and IDT thresholds were measured at the last click in the train. Results show a negatively decelerating function as n increases, wtih IDT thresholds falling from 250 down to 20 μs. The most dramatic find in these series of experiments, however, was the surprising yet compelling dissipation of “suppression” in the temporal window of interest (1- to 5-ms ICIs) over multiple runs. Thresholds fell by 5 to 10 folds, matching values reported by Perrott et al. (in press) in the free field. Across variations of stimulus duration, form, intensity, or frequency, highly trained subjects showed no recovery of the precedence effect. [Work supported by NSF and NIH.]