Aging and the perception of temporally interleaved words

Abstract

Objective: The purpose of this study was to examine differences between older and younger listeners in the ability to sequentially attend to and ignore words.

Design: Participants (n = 13 older adults and 13 younger adults) completed a temporally interleaved word recognition task. On each trial, 10 words were presented, and participants were instructed to repeat back every other word while ignoring the intervening words. Three variables were examined: (1) whether the word strings that were to be attended and to be ignored created syntactically correct sentences; (2) whether the to-be-attended and to-be-ignored words were presented from the same or from different spatial locations; and (3) whether the five target words in each trial (and the five distractor words in each trial) were spoken by a single talker or by five different talkers. In addition, digit-span forward and digit-span backward were measured and used as variables in correlation analyses.

Results: As a group, the younger participants outperformed the older listeners, particularly when the to-be-attended and to-be-ignored words were presented from the same spatial location (versus when they were presented with spatial separation). Compared with the younger participants, older listeners also made more error responses that were to-be-ignored words, although the proportion of errors that were not responses involving masking words did not significantly differ between groups. Scores on the digit-span-forward test (but not digit-span-backward scores or the degree of hearing loss) were associated with older individuals' performance on this temporally interleaved speech-recognition task.

Conclusions: The overall pattern of results suggests that factors other than threshold elevation contribute to speech-understanding problems experienced by older listeners. However, although younger adults outperformed older listeners on this interleaved sentence task, older and younger adults benefited, to a similar extent, from spatial separation of the to-be-attended and to-be-ignored words, and from having a consistent target talker within a trial.

Figure 1

Composite audiograms (means and standard errors) for the younger (left panel) and older (right panel) participants. The dotted lines represent the range of pure-tone thresholds.

Figure 2

Performance on the interleaved sentence tasks. FF = target and masker presented without spatial separation, both from a front loudspeaker; FR = target and masker presented in a spatially separated fashion, with the target words coming from the front and the masker words delivered by a loudspeaker located 60 degrees to the right of the participant. Errors bars represent the standard error.

Figure 3

Difference scores derived for the three within-subjects variables. The upper panel shows spatial benefit (spatially-separated FR trials – spatially coincident FF trials). The middle panel displays talker consistency benefit (talker consistent trials – talker variable trials). The bottom panel represents syntactic benefit (syntactically-correct word order trials – random word order trials). Error bars represent the standard error.

Figure 4

Error responses produced by participants, averaged across all conditions. Masker errors were when participants responded with a word from the to-be-ignored stream. Non-masker errors were responses that were words not played on that trial, or omissions. Error bars represent the standard error.

Figure 5

Scatterplots depicting the associations between digit-span forward and performance in syntactically-correct trials (top panel) and random word order trials (bottom panel) for the 13 older participants.