Using the Coordinate Response Measure (CRM) paradigm, recognition of target speech in the presence of competing speech has been shown to depend upon both the rhythmic context of target and background speech and fundamental frequency differences between the target and background speakers (McAuley et al., 2021). In the present study, two experiments examined the effects of the relative timing of target and background key words and the presence or absence of a background rhythmic context on target key word recognition using the same male talker for both target and background sentences. Experiment 1 varied the onset asynchrony between target and background key words when background rhythmic context was removed (i.e., the background consisted only of the competing key words) and Experiment 2 manipulated the rhythm of background speech leading up to key words, but left the key words intact with an onset asynchrony of ℗ł50ms. Experiment 1 revealed an asymmetric U-shaped performance curve where (1) target recognition improved with increasing deviation of background key words from the expected onset timing of target keywords, and (2) target key words were better recognized when they began prior to the onset of background key words, compared to when they began after. With the reintroduction of the background context in Experiment 2, performance was reduced to chance both when the background rhythm was intact and when it was altered to be rhythmically irregular, suggesting that listeners were unable to distinguish target and background sentences and could not develop expectations for target key word timing.
Read
