Figure 1.

Task design. A, Trial sequence. Subjects were presented with an auditory scene consisting of two overlapping streams. They were instructed to detect and respond to a repetition in one of the two streams. A cue indicated whether the repetition would appear in the speech stream (S), the environmental sound stream (E), or either stream (S or E). B, Example auditory scene (bottom), consisting of a speech signal (top) and an environmental signal (middle). Time–frequency plots represent the spectral content of the streams; time series represent the normalized amplitude. Full and dashed green outlines indicate the first and second instance of an example repetition, respectively. Note that the temporal location of the repetition within the sound scene was randomized across trials.