Spectral contrast enhancement improves speech intelligibility in noise for cochlear implants

Area of Science:

• Audiology
• Biomedical Engineering
• Signal Processing

Background:

• Cochlear implant (CI) users often struggle with speech intelligibility due to spectral smearing, requiring higher signal-to-noise ratios (SNRs).
• Spectral smearing arises from the limited number of CI electrodes and electrical field overlap, distorting the spectral information of sound.
• Existing CI strategies may not fully compensate for these limitations, impacting users' ability to understand speech in noisy environments.

Purpose of the Study:

• To evaluate a novel spectral contrast enhancement (SCE) algorithm as an add-on feature for standard CI coding strategies.
• To determine if SCE can mitigate the effects of spectral smearing and improve speech reception thresholds (SRTs) in CI users.
• To assess the impact of SCE on electrical stimulation patterns and loudness perception.

Main Methods:

• Developed and implemented an SCE algorithm that enhances spectral peaks while attenuating valleys in speech signals.
• Conducted speech reception threshold (SRT) tests with 12 CI users comparing a standard CI strategy with and without the SCE algorithm.
• Performed a second evaluation with balanced loudness perception between SCE and non-SCE conditions to isolate the intelligibility benefits.

Main Results:

• Initial SRT tests showed no significant difference between the standard strategy and SCE.
• Analysis revealed reduced stimulation current with SCE, suggesting potential efficiency gains.
• In the loudness-balanced condition, SCE significantly improved SRT by 0.57 dB (p < 0.0005), indicating enhanced speech intelligibility.

Conclusions:

• The spectral contrast enhancement (SCE) algorithm shows promise for improving speech intelligibility in cochlear implant (CI) users, particularly when loudness perception is matched.
• SCE partially compensates for spectral smearing, a known limitation in CI processing.
• Further research and optimization of SCE could lead to more effective CI sound processing strategies, enhancing communication for CI recipients.