Individual control of input rate improves recall of spoken discourse by adult users of cochlear implants: An exploratory study

• Published in: Quarterly journal of experimental psychology (2006) 2024-12, p.17470218241301415
• Main Authors: O'Leary, Ryan M, Capach, Nicole Hope, Hansen, Thomas A, Kinney, Alex J, Payne, Taylor A, Wingfield, Arthur, Svirsky, Mario A
• Format: Article
• Language: English
• ISSN: 1747-0226; 1747-0226
• DOI: 10.1177/17470218241301415

Although cochlear implants (CI) successfully replace the sense of hearing, they do not restore natural hearing. Still, CI users adapt to this novel signal, reaching meaningful levels of speech recognition in clinical tests that focus on repetition of words and short sentences. However, many patients who score above average in clinical speech perception tests complain that everyday speech interactions are both difficult and cognitively draining. In part, this difficulty may be due to the naturally rapid pace of everyday discourse. We report a study in which 12 CI users aged 23 to 77, recalled multi-sentence discourse presented without interruption, or in the condition of interest, when passages were paused at major linguistic boundaries, with participants given control of when to initiate the next segment. Comprehension of the discourse structure was based on a formalised representational system that organises discourse elements hierarchically to index the relative importance of different elements to the overall understanding of the discourse. Results showed (a) better recall when CI users were allowed to control the discourse pace; (b) an overall effect of aging, with older CI users recalling discourse less accurately; (c) better recall for passages with higher average inter-word predictability; (d) a "semantic hierarchy effect" reflected by better recall of main ideas versus minor details; (e) an attenuation of the semantic hierarchy effect for low predictability passages. Results underscore the benefits of extra processing time in addressing CI listening challenges and highlight the limited ecological validity of single-word or single-sentence speech recognition tests.