Comparison of Frequency Transposition and Frequency Compression for People With Extensive Dead Regions in the Cochlea.

Only 10 participants completed the trial, and three of the 13 enrolled dropped out, so the design was not fully balanced. Dead-region testing was often inconclusive because high-frequency hearing thresholds were very high, and some participants had middle-ear dysfunction or eardrum perforations. Speech-in-noise testing could be completed by only three participants, so it gives little information about everyday listening in noise. The frequency-lowering settings were not chosen using audibility estimates, and the source band often missed the frequency range where /s/ has most energy. Conditional frequency lowering occurred inconsistently in background noise, so participants may not have had stable exposure to the lowered speech sounds. The participants had long-standing severe high-frequency hearing loss and extensive dead regions, so the results may not apply to people with smaller dead regions.

Funding included Phonak AG, whose hearing aids were used, plus the Medical Research Council, Action on Hearing Loss, the H.B. Allen Trust and the Engineering and Physical Sciences Research Council.

Do not read the improved audibility of /sh/ or the more consistent error patterns as evidence that frequency lowering improves speech recognition. The consonant identification, /s/ and /z/ detection and self-rated hearing outcomes did not show a significant difference from control hearing aids, and speech-in-noise testing was possible for only three participants.