Improving Spatial Perception and Speech Understanding in Multitalker Mixtures

The purpose of this study is to investigate several approaches for improving spatial perception and speech intelligibility in multitalker listening situations for hearing-aid users. The hypotheses are that spatial perception and speech intelligibility will be improved by (1) increased high-frequency audibility, (2) speech envelope enhancement, and/or (3) appropriate sound image externalization.

The purpose of this study is to investigate several approaches for improving spatial perception and speech intelligibility in multitalker listening situations for hearing-aid users. Aim 1 will investigate how well different amplification strategies provide access to the spectrotemporally sparse speech information that is available in multitalker mixtures. The hypothesis is that subtle differences in high-frequency audibility across different strategies may have much greater effects on the intelligibility of sparse speech than intact speech. This Aim will involve 40 subjects with hearing loss across 2 experiments. Aim 2 is an exploratory aim that will investigate the potential of a speech enhancement algorithm to improve spatial hearing in multitalker mixtures. The hypothesis is that speech enhancement, in addition to its established monaural benefits, will enhance the salience of interaural time differences and thus provide binaural benefits in multitalker mixtures. This Aim will involve 30 subjects with normal hearing and 60 subjects with hearing loss across 3 experiments. Aim 3 will investigate how different hearing-aid styles affect sound externalization. The hypothesis is that hearing aids can disrupt sound externalization, which in turn impacts speech intelligibility in multitalker mixtures. This Aim will involve 30 subjects with normal hearing and 60 subjects with hearing loss across 3 experiments. This is a prospective design involving adult human subjects. While the investigators will recruit subjects with normal and impaired hearing, all subjects will be assigned to the same set of experimental conditions. Therefore, for the purposes of intervention (or prospective assignment), the subjects can be treated as a single group completing the same intervention (or set of interventions). The behavioral experiments involve listening to speech sounds (either unaided, with hearing aids, or with simulated hearing aids), and providing responses related to spatial perception or speech intelligibility. The order of experimental conditions will be randomized for each subject in order to control for order effects. All of the experimental tasks and outcome measures have been previously established, have good reliability and validity, and will be appropriate for informing future studies.

Hearing Aids, Speech Intelligibility, Spatial Hearing, Cocktail Party, Audibility, Sound Image Externalization, Speech Envelope

All participants will be tested under a number of experimental conditions that simulate different hearing aid conditions. They will act as their own control.

While participants will not be informed about the specific experimental condition being tested, certain acoustical changes may be perceptible to them and thus strict masking is impossible.

Participants will be presented with sound stimuli under a number of experimental conditions that simulate different hearing-aid conditions.

The intervention will be a hearing aid or a modification to the sound stimulus that simulates a hearing aid configuration, style, or processing strategy.

Speech intelligibility scores will be measured in a baseline condition (e.g. unaided) and in a number of hearing-aid conditions (post-intervention). Scale is in percent correct (0-100%).

Speech intelligibility scores will be measured in a baseline condition (e.g. unaided) and in a number of hearing-aid conditions (post-intervention). Scale is in percent correct (0-100%).

Speech reception thresholds will be measured in a baseline condition (e.g. unaided) and in a number of hearing-aid conditions (post-intervention). Scale is in decibels (the level of the target speech relative to the masker speech needed to identify 50% of the words in a sentence).

Speech reception thresholds will be measured in a baseline condition (e.g. unaided) and in a number of hearing-aid conditions (post-intervention). Scale is in decibels (the level of the target speech relative to the masker speech needed to identify 50% of the words in a sentence).

Thresholds for the discrimination of interaural time differences in speech will be measured in a baseline condition (e.g. unaided) and in a number of hearing-aid conditions (post-intervention). Scale is in microseconds.

Thresholds for the discrimination of interaural time differences in speech will be measured in a baseline condition (e.g. unaided) and in a number of hearing-aid conditions (post-intervention). Scale is in microseconds.

Externalization judgments for speech will be measured in a baseline condition (e.g. unaided) and in a number of different hearing-aid conditions (post-intervention). Scale is in percent (0-100%).

Externalization judgments for speech will be measured in a baseline condition (e.g. unaided) and in a number of different hearing-aid conditions (post-intervention). Scale is in percent (0-100%).

Inclusion Criteria: Normal-Hearing Subjects 18 to 35 years of age Audiometric thresholds that do not exceed 20 dB HL at any frequency from 250 to 8000 Hz Able to provide informed consent and understand experimental instructions Normal or corrected-to-normal vision Inclusion Criteria: Hearing-Impaired Subjects 18 to 80 years of age Documented sensorineural hearing loss Able to provide informed consent and understand experimental instructions Normal or corrected-to-normal vision Exclusion Criteria (some experiments) Non-native speakers of English