Transient Noise Reduction in Cochlear Implant Users: a Multi-Band Approach

Area of Science:

• Audiology
• Biomedical Engineering
• Signal Processing

Background:

• Cochlear implant (CI) users often struggle with speech perception in noisy environments.
• Previous transient noise reduction (TNR) algorithms offered limited benefits.
• Transient noises are spectrally distinct from speech, suggesting potential for targeted noise reduction.

Purpose of the Study:

• To evaluate a modified multi-band TNR algorithm for CI users.
• To compare the multi-band TNR against a single-band TNR and no TNR.
• To assess improvements in speech intelligibility, sound quality, and warning signal recognition.

Main Methods:

• Fifteen experienced CI users participated in the study.
• A modified TNR algorithm was developed to process noise across multiple frequency bands.
• Speech reception thresholds (SRTs) were measured in various noise conditions (cafeteria, office) with three processing settings: TNR off, single-band TNR, and multi-band TNR.

Main Results:

• Multi-band TNR (TNR_mult) improved SRTs by up to 3 dB compared to both no TNR (TNR_off) and single-band TNR (TNR_sgl).
• TNR_mult resulted in higher subjective ratings for comfort and clarity in cafeteria noise.
• No significant differences were observed in warning signal recognition across conditions.

Conclusions:

• Multi-band TNR is more advantageous than single-band TNR for CI users.
• The developed TNR algorithm shows significant potential to enhance speech perception and listening comfort.
• Further research into TNR applications for CI users is warranted.