Evaluation of a New Algorithm to Optimize Audibility in Cochlear Implant Recipients.
Level 3 - non-randomized controlled study
Prospective within-subject comparative cohort study with acute laboratory and take-home evaluations without reported randomization
PubMed 30418283 · doi:10.1097/AUD.0000000000000680
What was done
The study evaluated the SoftVoice algorithm (Advanced Bionics [AB]), designed to remove system noise and improve soft-speech audibility for adult cochlear implant (CI) recipients. In Study 1A, 16 adult unilateral AB CI recipients were tested acutely using a Harmony processor to compare speech recognition between their standard everyday program and the same program with SoftVoice across soft to conversational presentation levels in quiet (35 to 60 dB SPL) and noise (60 dB SPL). In Study 1B, 10 participants compared both programs at home. In Study 2, 10 participants completed acute testing with a Naída CI Q70 processor and evaluated different electrode threshold level (T-level) programming settings.
What was found
In Study 1A, speech recognition scores were significantly higher with SoftVoice than Everyday at soft levels (35, 40, 45, and 50 dB SPL), with no significant differences at conversational level (60 dB SPL) in quiet or noise. In Study 1B, 5 of 10 participants preferred SoftVoice at home, while 6 reported bothersome environmental sound. In Study 2, soft-speech performance was similar between Harmony and Naída processors; setting T-levels at the default (10% of most comfortable levels [Ms]) reduced bothersome environmental sounds during take-home use, whereas soft-speech perception was highest when T-levels were behaviorally set above 10% of Ms. Specific numerical percentage scores and statistical values were not reported in the abstract.
Why it matters
SoftVoice improves soft-speech perception in cochlear implant users without degrading performance at conversational levels or in noise. However, clinical implementation requires balancing electrode threshold levels to optimize soft-speech audibility against patient comfort in daily acoustic environments.
Limits
Sample sizes were small (16 participants in Study 1A, 10 in Study 1B, 10 in Study 2) and restricted to adult users of a single implant brand. The abstract does not report randomization or blinding in testing orders, nor does it provide quantitative speech recognition scores, confidence intervals, or p-values.
Cited by
- supports Normal conversational speech occurs at approximately 60 decibels sound pressure level.