Amplitude mapping and phoneme recognition in cochlear implant listeners

Area of Science:

• Audiology
• Biomedical Engineering
• Speech Science

Background:

• Cochlear implants (CIs) require sound compression due to limited electric dynamic range.
• Understanding the impact of compression and dynamic range reduction on speech recognition is crucial for CI users.

Purpose of the Study:

• To investigate how amplitude compression and dynamic range reduction affect phoneme recognition in quiet and noise for CI listeners.
• To compare CI listener performance with normal-hearing listeners.

Main Methods:

• Four CI users of the Nucleus-22 SPEAK processor participated.
• Amplitude compression was varied by Q-value; dynamic range was reduced by adjusting threshold and comfortable levels.
• Phoneme recognition was tested with vowels and consonants in quiet and noise across various signal-to-noise ratios (SNRs) and electrode counts (20, 10, 4).

Main Results:

• Amplitude compression did not significantly affect phoneme recognition.
• Dynamic range reduction marginally impacted quiet recognition but significantly degraded noise recognition.
• Performance with 20 and 10 electrodes was similar; 4 electrodes yielded poorer results.
• CI users required higher SNRs and scored lower than normal-hearing listeners.

Conclusions:

• Amplitude compression has minimal effect, but dynamic range reduction significantly impairs speech recognition, particularly for vowels in noise.
• Findings may be specific to spectral peak-extracting processors like SPEAK.
• More than four electrodes are likely needed for optimal speech recognition in noise.
• A performance gap persists between CI and normal-hearing listeners, necessitating improved CI designs and fitting strategies.