Hearing Aid Algorithms for Sudden Sounds

Participants will first undergo a comprehensive audiologic test battery and questionnaires. The audiologic test battery will consist of (a) pure tone audiometry: basic and extended high frequency, (b) tympanometry, (c) wideband acoustic immittance, (d) acoustic reflexes, and (e) otoacoustic emissions. They will also complete online questionnaires at home: (a) loudness and annoyance, (b) noise exposure, (c) general survey concerning hearing in difficult situations, medical and otologic history, noise exposure, tinnitus, balance, sensory sensitivity (lights, sound, touch, smell, taste), and amplification. In addition, they will complete online tests of temporal gap detection.

Participants who meet the study inclusion criteria (hearing aid candidacy based on air conduction thresholds between 2000 and 6000 Hz) will return to the lab for two additional sessions. During the first session, before starting the main experiment, they will complete the Working Memory Questionnaire (Vallat-Azouvi et al., 2012), which evaluates short-term storage, attention, and executive control. The first main experiment evaluates sentence recognition, storage, and retrieval. Participants will hear two sentences in a row: one by a male or female talker and the other by a person of the opposite gender. Participants are asked to remember the first sentence while listening to the second sentence. They will be asked to repeat the last sentence aloud (a measure of sentence recognition). Then, they will be asked to type the first sentence (a measure of sentence storage and retrieval). Most, but not all, sentences will have sudden soft sounds and one sudden loud sound. Participants will hear each pair of sentences only once; they will not be allowed to replay the audio.

Sentence pairs were recorded through hearing aids in an anthropomorphic manikin with one of four treatments for sudden sounds: (1) off, (2) low, (3) high, and (4) maximum. Participants will listen to 36 sentence pairs for each of the four treatments. All will contain sudden soft sounds (e.g., keyboard typing, footsteps, chewing, etc.) throughout the two sentences; half will contain a sudden loud sound (e.g., gunshot, door slamming, iPhone notification, etc.) at the onset of a keyword in the first sentence and the other half in the second sentence. Control stimuli do not contain sudden sounds. Conditions assigned to each sentence list (18 pairs) will be randomly assigned to each sentence participant. In addition, all stimuli were recorded using one of eight standardized audiometric configurations. Participants will listen to the recordings made with the audiogram that best matches the hearing loss in one of their ears. Participants can scale the overall output up or down 15 dB to accommodate differences in loudness preferences. Finally, all recordings were filtered before playback to remove the resonance associated with the external ear properties of the anthropomorphic manikin using the test earphones.

Participants will return to the lab for a second session in which they will rate their preferences for the different sudden sound reduction settings. Before beginning the experiment, they will complete the Visual Letter Monitoring (VLM) test (Gatehouse et al., 2003). The VLM test assesses attention, reaction time, continuous performance, and working memory capabilities. Participants will observe letters displayed one at a time in large font on a computer screen. Letters will switch between vowels and consonants in an unbroken sequence with intervals of 2 ("slow") or 1 second ("fast"). Participants will press the space bar on the computer keyboard or a button on the touchscreen whenever three sequential letters (Consonant-Vowel-Consonant) form a word.

For the subjective preference portion of the experiment, participants listen to sentences recorded with two different sudden sound reduction settings (6 combinations comprised of off, low, high, or maximum). Each sentence will be played twice, in succession, but with different pairs of settings from trial to trial. Following each trial, they will select which setting they prefer overall. Additionally, they will choose the setting they favor more for the speech (identified as the clearest, most natural, and least distorted) and the setting they favor for the noise (deemed the most comfortable, least annoying, and softest). To indicate their preference, participants will press the "1" button if the first setting was more to their liking or the "2" button if they found the second setting superior. Participants are permitted to revise their choice and switch their selection. They can re-listen to the sentence pairing as many times as needed. Participants will listen to 72 sentences for each combination of settings to provide sufficient power to conduct single-subject statistical analyses.