Computer-Based Connected-Text Training of Speech-in-Noise Perception for Cochlear Implant Users.

Only nine participants completed the training, so the result is a small single-group study and cannot show how much of the improvement was due to the training itself. There was no control group or randomisation described, so practice effects, expectation or other changes over time cannot be ruled out. Sentence recognition was tested with the same 20-talker babble used in training, while other tests used speech-shaped noise, so learning to cope with that specific babble may have affected the sentence result. The proportion of key words that occurred at least once during training was around 0.6 for BKB sentences and 0.5 for IEEE sentences, so overlap between training words and test words may have helped sentence scores. Two participants withdrew during pretesting, S8 did not have follow-up testing, and some secondary tests were completed fewer times for a few participants. Participants' attitudes, motivation and expectation were not formally assessed, and no blinding to training is described. The study did not measure everyday communication outcomes, so it is unclear whether the laboratory improvement would help patients in real-world listening.

Funded by Action on Hearing Loss; no other funding or author conflict details are given in the supplied text.

Do not conclude that this training improves all speech perception in noise or works for every cochlear implant user. Only nine experienced unilateral users were studied, and gains were found for sentence recognition in babble while vowel identification, consonant identification and digit span did not show clear improvement.