Speech in noise performance in adults with cochlear implants using a combined channel deactivation strategy with a variable number of dynamic focused channels.

Abstract

Listen

Translate article

Cochlear implant listeners show difficulties in understanding speech in noise. Channel interactions from activating overlapping neural populations reduce the signal accuracy necessary to interpret complex signals. Optimizing programming strategies based on focused detection thresholds to reduce channel interactions has led to improved performance. In the current study, two previously suggested methods, channel deactivation and focused dynamic tripolar stimulation, were combined to create three cochlear implant programs. Utilizing an automatic channel selection algorithm from focused detection threshold profiles, three programs were created with the same deactivated channels but varying proportions of channels employing focused stimulation, monopolar, dynamic focused and a mixed program. Thirteen ears in eleven adult cochlear implant listeners with Advanced Bionics HiRes90k devices were tested. Vowel identification and sentence perception in quiet and noise served as outcome measures, and the influences of listening experience, age, clinical consonant-nucleus-consonant performance, and perceptual thresholds on speech performance were assessed. Across subjects, different degrees of focusing showed individual performance improvements for vowels and sentences over the monopolar program. However, only slight trends and no significant group improvements were observed. Focused listening benefits were shown for individuals with less cochlear implant experience, and clinically poor performers seem to benefit more from focusing than good performers. The current findings suggest that deactivating and focusing subsets of channels improves speech performance for some individuals, especially poor performers, a possible effect of reduced channel interactions. The findings also show that individual performance is largely variable, possibly due to listening experience, age or the underlying detection threshold.