Fig. 4.

Experiments 5a and 5b: Foreground detection is robust to background interruptions. (A) Experimental task. Stimuli were like those from Experiment 1 but were modified by replacing the Middle 500 ms (Experiment 5a) or 1,500 ms (Experiment 5b) of background noise with either silence (orange) or white noise (yellow). Participants were asked to ignore this interruption and judge whether the stimulus contained one or two sound sources. (B) Experiment 5a results. Average foreground detection performance (quantified as d′) is plotted as a function of interrupter type and foreground onset time. Shaded regions plot SE. The gray region denotes the temporal position of interruption in background noise. * indicates statistical significance, P < 0.001. (C) Experiment 5b results. Same conventions as (B).