Fig. 3.

Mean performance across the twenty-six listeners who could perform the task in terms of d’ scores for the within- (left panel) and the across- (right panel) sequence tasks in Experiment 2. In the within- sequence task, high d’ values indicate a greater tendency to segregate the sequences apart; in the across- sequence task, high d’ values indicate a greater tendency to integrate the interleaved sequence into one single stream. The error bars correspond to ±1 standard error of the mean.