Predicting the speech reception threshold of cochlear implant listeners using an envelope-correlation based measure

Area of Science:

• Audiology
• Biomedical Engineering
• Speech Processing

Background:

• Cochlear implants (CI) aim to restore hearing but performance varies significantly among users, especially in noisy environments.
• Accurate prediction of speech intelligibility is crucial for optimizing CI device fitting and improving patient outcomes.

Purpose of the Study:

• To propose and validate a novel modulation-based index for predicting speech intelligibility in CI users.
• To assess the index's correlation with objective speech-reception thresholds (SRTs) in various listening conditions.

Main Methods:

• Speech envelopes were extracted using individual CI users' daily fitting strategies.
• A modulation-based index was developed incorporating parameters like active electrodes and electrical dynamic range.
• The proposed index was correlated with SRTs measured in steady and speech-masker conditions.

Main Results:

• The proposed index demonstrated a high correlation (r = 0.96) with SRTs obtained from CI users.
• Electrodogram information proved sufficient for reliably predicting CI user performance in noise.
• The index effectively incorporates individual CI fitting parameters.

Conclusions:

• The developed modulation-based index is a reliable tool for predicting speech intelligibility in CI users.
• This index can assist clinicians in optimizing CI fitting parameters for enhanced performance in noise.
• The findings suggest electrodogram-based metrics are valuable for assessing CI user capabilities.