Michael F. Dorman - US grants

The long term objectives of our research are to solve the puzzle of why hearing-impaired listeners achieve, variously, good or poor word recognition scores when signals are presented at suprathreshold levels and to define signal modifications, other than amplification, which will benefit the hearing-impaired listener. The short-term objectives of the proposed research are: (1) to assess the contributions of spectral tilt at signal onset, burst/voice onset time (VOT) duration, and burst/formant frequency information to the identification of /bdgptk/ when combined with /iau/ and created to simulate male and female voices presented both in quiet and against background noise; and (2) to assess how the listeners' age, hearing impairment, temporal resolution capacity, and prior hearing experience (i.e., age of onset of impairment) affect usage of the several cues to place of articulation. Our methodology involves creating stimuli by factorily combining the frequency, tilt, and temporal cues for labial, alveolar, and velar place of articulation. For some signals all of the cues will be appropriate for a given place, while for others, one or more of the cues will be inappropriate for that place. Our pilot data demonstrate that the identification of these signals differentiates normal and hearing-impaired listeners. For example, the information provided by the tilt of the onset spectrum and by burst/VOT duration affects the performance of hearing-impaired listeners more than normal hearing listeners. This outcome, in turn, suggests that strategies for signal modification, other than simple amplification, may be of benefit to hearing-impaired individuals.

The long term goals of the research proposed here is to account for three aspects of speech recognition by patients who use a multichannel cochlear prosthesis: (1) Different initial (1 month) levels of performance; (2) different rates of improvement during the first year of implant use and (3) different terminal levels of performance. Preliminary studies suggest that two factors may account in large measure for the differences in terminal performance: the range of pitch sensation available through the prosthesis and the dynamic range of intensity available to the patient. We propose to assess the relationship between measures of auditory function (pitch scaling, frequency DLs, dynamic range and intensity DLs) and speech recognition (vowel and consonant recognition, spondee recognition, monosyllabic word recognition and recognition of words in sentences) in patients who use the Symbion 4-channel prosthesis. Two experiments are proposed. In one experiment, measures of auditory function and measures of speech recognition will be assessed for patients who have reached an asymptotic level of speech recognition. Four groups of 10 patients will be tested-- two groups with poor speech recognition scores, one group with average speech recognition scores and one group with good speech recognition scores. In the second experiment, the measures of performance will be taken at intervals of 1 month, 6 months and 1 year after patients (n =12) have been fit with a prosthesis. If we can relate differences in speech recognition to differences in auditory function, then we will have gained a major step in understanding speech recognition by patients who use cochlear prostheses.

DESCRIPTION (provided by applicant): The goal of this project is to conduct a study of the speech, voice and music perception abilities of patients who hear with combined electric and acoustic stimulation (EAS) and to relate speech, voice and music perception to underlying psychophysical abilities in the region of acoustic hearing. The goal will be met through experiments and analyses relative to 6 aims. Aim 1 is to characterize low frequency acoustic hearing in patients with residual hearing before and after implantation by assessing auditory thresholds and by obtaining estimates of nonlinear cochlear processing, frequency selectivity and temporal resolution. Aim 2 is to assess the effects of acoustic stimulation alone, electric stimulation alone and combined acoustic and electric stimulation on tests of speech, voice and music recognition. Aim 3 is to determine the relationships among the psychophysical measures of low-frequency acoustic hearing and the measures of speech, voice and music recognition. Aim 4 is to compare the performance of EAS patients, patients with unilateral implants and patients with bilateral implants on measures of speech, voice and music recognition. Aim 5 is to assess the pitch of stimulation delivered to electrodes relative to that of pure tones in the region of acoustic hearing and to relate electrical pitch to computerized radiography assessment of electrode position in the cochlea. Aim 6 is to conduct studies of filter-band-to-electrode assignments in order to assess (a) whether assigning filter-bands to electrodes based on acoustic-electric pitch matching (as in Aim 5) produces better speech recognition than a standard assignment of filter bands and (b) whether overlap in representation of low frequencies by acoustic and electric stimulation is necessary for maximum levels of performance.

DESCRIPTION (provided by applicant): In the near future most cochlear implant patients will hear with two 'ears' -- either with two cochlear implants (CI) or with a CI and a hearing aid in the opposite ear (combined electric and acoustic hearing or EAS). The goal of the research proposed here is to develop a tool for clinical decision making relative to these two interventions. Clinical decision making will depend critically on the nature of the tests and environments used to assess the benefit of having two ears participate in speech understanding. Standard clinical test environments can only approximate the real world environment of having sound surround the listener. Laboratory environments with multiple, spatially separated speakers can simulate real world environments but cannot be duplicated in the clinic for reasons of time, space and cost. In this project we will test EAS and bilateral CI patients in both standard and realistic test environments with the goal of creating a decision matrix that links data that can be easily collected in the clinic, e.g., CNC scores in quiet and the amount of residual hearing, with data that cannot be collected in the clinic, i.e., data collected with multiple, spatially separated loudspeakers. We propose to test bilateral CI patients and EAS patients in two realistic test environments using an 8-speaker 'surround sound system' which can simulate, with high fidelity, (i) a restaurant environment with speech babble as the noise, and (ii) a cocktail party environment with competing sentence material as the noise - a situation of 'informational' masking.

DESCRIPTION (provided by applicant): Many deaf or severely hearing-impaired individuals can understand speech in quiet environments using a cochlear implant (CI), which stimulates the auditory nerve directly with electrical current. However, their speech understanding typically declines significantly in even small amounts of background noise. For those who have some residual low-frequency hearing, the combination of electric and acoustic stimulation (EAS) can significantly improve speech understanding in background noise. Fundamental frequency (F0) variation and low-frequency amplitude envelope of the target talker are important cues for EAS benefit. The broad long-term goals of the proposed research are to advance the understanding of how low-frequency acoustic stimulation combines with electric stimulation to enhance speech understanding in difficult listening situations, and to enhance EAS benefit for individuals who might otherwise receive limited or no benefit. One long-term goal is to develop a wearable real-time processor that can deliver low-frequency speech cues to CI users more effectively. The specific aims are to (1) to increase the amount of EAS benefit to CI patients who already show a benefit; (2) To provide EAS benefit other CI patients who show little or no benefit typically; and (3) understand why some CI patients do not benefit from EAS, even when their audiometric results suggest they might. This work has the potential to extend the benefits of EAS to those CI users who do not possess enough residual hearing to show an EAS benefit typically, and to enhance the EAS benefit for those who do.