Behavioral assessment of adaptive feedback equalization in a digital hearing aid.

An evaluation was made of the efficacy of a digital feedback equalization algorithm employed by the Central Institute for the Deaf Wearable Adaptive Digital Hearing Aid. Three questions were addressed: 1) Does acoustic feedback limit gain adjustments made by hearing aid users? 2) Does feedback equalization permit users with hearing-impairment to select more gain without feedback? and, 3) If more gain is used when feedback equalization is active, does word identification performance improve? Nine subjects with hearing impairment participated in the study. Results suggest that listeners with hearing impairment are indeed limited by acoustic feedback when listening to soft speech (55 dB A) in quiet. The average listener used an additional 4 dB gain when feedback equalization was active. This additional gain resulted in an average 10 rationalized arcsine units (RAU) improvement in word identification score.

Properties of an adaptive feedback equalization algorithm.

This paper describes a new approach to feedback equalization for hearing aids. The method involves the use of an adaptive algorithm that estimates and tracks the characteristic of the hearing aid feedback path. The algorithm is described and the results of simulation studies and bench testing are presented.

Adaptive feedback stabilization of hearing aids.

A method is described for adaptively identifying and equalizing the feedback path of a hearing aid in order to stabilize the system. The algorithm utilizes an LMS adaptive filter and is implemented in digital form. An additional 10 to 15 dB of stable gain margin for a BTE hearing aid configuration has been demonstrated in laboratory experiments with hearing-impaired subjects using an interim custom VLSI circuit implementation.

High-frequency thresholds: sound suite versus hospital room.

Benefits of high-frequency audiometry in monitoring hearing sensitivity of patients administered ototoxic medications are well established. Thresholds obtained within a sound suite have been proven reliable. It may, however, often be necessary for the audiologist to evaluate the patient at bedside. The primary purpose of this study was to determine if significant differences are present between high-frequency thresholds measured in a sound suite versus thresholds measured in a hospital room. In addition, the test-retest reliability of high-frequency thresholds was determined when measured in a hospital room. For 25 normal hearing subjects, results revealed that significant differences were not observed between thresholds measured in a sound suite versus those measured in a typical hospital room. In addition, differences between the initial and repeated thresholds obtained in the hospital room were not significant, and the differences were, for the most part, within +/- 10 dB at all test frequencies.

Behavioral assessment of CID's benchtop (version 2) digital hearing aid: noise reduction.

Perceptual rating and word identification procedures were used to evaluate the efficacy of several digital hearing aid adaptive and nonadaptive noise reduction algorithm parameters. The study included 15 hearing-impaired adults listening to speech at soft, moderate, and loud input levels in quiet, fan noise, and multitalker babble noise. The results suggest that hearing-impaired listeners can make clarity judgments that result in the selection of algorithms that optimize word intelligibility at soft and moderate speech levels. The spectral adaptation requirements, however, were not constant with signal level.

Role of predictability of sequence in auditory stream segregation.

This study evaluates the role of item predictability in auditory temporal coherence. Thirteen normal-hearing subjects were required to hold together long tonal sequences as single strings of notes. Temporal and spectral predictability of successive notes in a sequence varied as a function of experimental condition. As the frequency separation of the notes in the sequence increased, the subjects found it more difficult to hold the sequence together as a single stream. There was no significant difference in subjects' abilities in performing this task as a function of experimental condition. That is, the predictability of successive notes appeared not to have a role in temporal coherence.

Inductive coupling of hearing aids and telephone receivers.

The present investigation sought to determine the effects of hearing loss, signal level, telephone receiver type, and telecoil location on the aided word-discrimination abilities of hearing impaired subjects using electromagnetic coupling of their hearing aids to the telephone. As expected, significant deterioration in discrimination scores was observed with increasing hearing loss. In addition, significant improvement in discrimination scores was observed, for all subject groups, as the signal level increased from 80 dB SPL to 105 dB SPL. Of the four receiver conditions, the U1 and BARC receivers resulted in the best discrimination scores. Also, the results suggest that the U1 and BARC receivers are indeed equivalent in terms of word discrimination scores. Finally, the telecoil location in individual hearing aids appeared to have little influence on the speech discrimination capabilities of the subjects in this study. However caution must be used in interpreting that result, since subjects were allowed to adjust the telephone handset position to maximize the signal level in any given condition. This occasionally resulted in inappropriate handset locations for conventional telephone use.

Acoustic reflex measures of cochlear damage--a normative study.

It has been argued that certain reflex measures may indicate the presence of end-organ hearing loss. It has been noticed that the difference between wide band noise and pure tone thresholds is diminished, that the reflex threshold duration function is flattened and that the reflex relaxation index, as specified by Norris et al. (1974b), is reduced in cases of end-organ hearing loss. These aspects of the acoustic stapedial reflex, together with its growth with stimulus intensity, have been investigated in 25 normally hearing subjects. The overriding feature of the measures taken of the acoustic stapedial reflex, was the high inter-subject variability. The differences between the acoustic reflex thresholds for pure tone and WBN was found to be normally distributed about a mean value of 12 dB. The threshold-duration functions again were highly variable and encompassed the entire range of results from normal and end-organ impaired groups as reported by Woodford et al. (1975). The variability encountered in the RRI measures was found to increase with frequency. The growth of the acoustic stapedial reflex with intensity was found to be frequency dependent: the gradient of the function increasing with frequency for the long duration (500 ms) pulses and decreasing with the short duration (50 ms) pulses. It is concluded that considerable caution must be exercised in extrapolation from any one of these measures in isolation.