Word Categorization of Vowel Durational Changes in Speech-Modulated Bone-Conducted Ultrasound.

Ultrasound can deliver speech information when it is amplitude-modulated with speech and presented via bone conduction. This speech-modulated bone-conducted ultrasound (SM-BCU) can also transmit prosodic information. However, there is insufficient research on the recognition of vowel duration in SM-BCU. The aim of this study was to investigate the categorization of vowel durational changes in SM-BCU using a behavioral test. Eight Japanese-speaking participants with normal hearing participated in a forced-choice behavioral task to discriminate between "hato" (pigeon) and "haato" (heart). Speech signal stimuli were presented in seven duration grades from 220 ms to 340 ms. The threshold at which 50% of responses were "haato" was calculated and compared for air-conducted audible sound (ACAS) and SM-BCU. The boundary width was also evaluated. Although the SM-BCU threshold (mean: 274.6 ms) was significantly longer than the ACAS threshold (mean: 269.6 ms), there were no differences in boundary width. These results suggest that SM-BCU can deliver prosodic information about vowel duration with a similar difference limen to that of ACAS in normal hearing.

Audiological evaluation of infants using mother's voice.

OBJECTIVES: Hearing loss is a serious problem in infants and children because it may interfere with the development of typical speech, verbal language, and auditory and communication skills. By measuring hearing ability (thresholds) as early as possible, even as early as during infancy, effective treatment can be administered. These treatments may significantly reduce the handicap associated with hearing loss. However, at times during behavioral auditory tests, observers cannot determine whether or not an accurate threshold was obtained. To support the use of infant audiometry for accurate diagnosis, audiologic behavioral responses may be obtained by selecting stimuli that interest infants, e.g., their mothers' voices. METHODS: We evaluated 30 infants who were presented to our hospital for behavioral auditory assessment in 2016. The infants' ages ranged from 4 months to 3 years and 6 months. The mean age was 1 year and 10 months ±10 months (±standard deviation). The infants' hearing thresholds for their mothers' voices and warble tones at 250-4000 Hz were measured. Auditory brainstem response (ABR) had already been evaluated in 24 infants. Relationships between mother's voice and warble tone or ABR thresholds as well as responses to the initial stimuli and stimuli at the threshold were investigated. These responses were classified into four grades (no response, uncertain response, possible positive response, and positive response), and the response to mother's voice and warble tone were subsequently compared. RESULTS: Mother's voice thresholds significantly correlated with all warble tone thresholds. In the relationship between the mother's voice threshold and average hearing levels of 500, 1000, and 2000 Hz, two infants were outliers. In these infants, the average hearing levels were relatively higher than the mother's voice thresholds. Judging from their ABR thresholds, the mother's voice thresholds were valid and the average hearing levels were worse than their original assessed hearing ability. The responses to mothers' voices were more distinct than those to warble tones, both for initial stimuli presentation and the determined threshold. CONCLUSIONS: Audiologic behavioral responses to mothers' voices were clearer than those for warble tones. Evaluations that use the mother's voice threshold are useful for estimating hearing levels in infants.

Temporal window of integration estimated by omission in bone-conducted ultrasound.

Bone-conducted ultrasound (BCU) can be heard for both normal-hearing and some profoundly deaf individuals. Moreover, amplitude-modulated BCU can transmit the speech signal. These characteristics of BCU provide the possibility of the developing a bone-conducted ultrasonic hearing aid. Previous studies on the perception mechanism of speech-modulated BCU have pointed to the importance of temporal rather than frequency information. In order to elucidate the perception of speech-modulated BCU, further investigation is need concerning the processing of temporal information. The temporal processing of air-conducted audible sounds (ACASs) involves the integration of closely presented sounds into a single information unit. The long-temporal window of integration was estimated approximately 150-200 ms, which contribute to the discrimination of speech sound. The present study investigated the long-temporal integration system for BCU evaluated by stimulus omission using magnetoencephalography. Eight participants with normal hearing took part in this study. Ultrasonic tone burst with the duration of 50 ms and frequency of 30 kHz was used as the standard stimulus and presented with steady onset-to-onset times or stimulus-onset asynchronies (SOAs). In each sequence, the duration of the SOAs were set to 100, 125, 150, 175, 200, or 350 ms. For deviant, tones were randomly omitted from the stimulus train. Definite mismatch fields were elicited by sound omission in the stimulus train with an SOA of 100-150 ms, but weren't with an SOA of 200 and 350 ms for all participants. We found that stimulus train for BCUs can be integrated within a temporal window of integration with an SOA of 100-150 ms, but are regarded as a separate event when the SOA is 200 or 350 ms in duration. Therefore, we demonstrated that the long-temporal window of integration for BCUs estimated by omission was 150-200 ms, which was similar to that for ACAS (Yabe et al. NeuroReport 8 (1997) 1971-1974 and Psychophysiology. 35 (1998) 615-619). These findings contribute to the elucidation and improvement of the perception of speech-modulated BCU.

Suppression of Subsequent N1m Amplitude When the Masker Frequency is Different from the Signal.

When two tones are presented in a short interval of time, the presentation of the preceding tone (masker) suppresses the response evoked by the subsequent tone (signal). To address the processing in forward suppression, we applied 2- and 4-kHz maskers, followed by a 1-kHz signal at varying signal delays (0 to 320 ms) and measured the signal-evoked N1m. A two-way analysis of variance revealed a statistically significant effect for signal delay in both masker presentation conditions. The N1m peak amplitude at the signal delay of 320 ms was significantly larger than those of 10, 20, 40, and 80 ms (p < 0.05). No significant enhancement for the very short signal delay was observed. The results suggest that the enhancement of N1m peak amplitude for short signal delay conditions is maximized when the frequency of the masker is identical to that of the signal.

Evaluation of prosodic and segmental change in speech-modulated bone-conducted ultrasound by mismatch fields.

Speech-modulated bone-conducted ultrasound (BCU) can transmit speech sounds for some profoundly deaf individuals. Hearing aids using BCU are considered to be a novel hearing system for such individuals. In our previous study, the speech discrimination for speech-modulated BCU was objectively confirmed using a magnetoencephalography. Moreover, in our previous behavioral study, prosodic information for speech-modulated BCU could also be discriminated in the normal hearing. However, the prosodic discrimination for speech-modulated BCU has not objectively been studied. In order to evaluate the prosodic discrimination for speech-modulated BCU, mismatch fields (MMFs) elicited by prosodic and segmental change were measured for speech-modulated BCU and air-conducted speech. Ten Japanese participants with normal hearing took part in this study. Stimuli re-synthesized from the speech of a native Japanese female adult were used. Standard stimulus was /itta/ with a flat pitch pattern, and two deviant stimuli were /itta?/ with a rising pitch pattern and /itte/ with a flat pitch pattern. All and nine participants elicited the prominent MMF elicited by the prosodic and segmental change for the speech-modulated BCU, respectively. The moment of MMF components for speech-modulated BCU was significantly smaller than those for air-conducted speech, while no difference in the MMF latency elicited by the prosodic and segmental change were observed between both stimulus conditions. Comparing the MMFs elicited by prosodic and segmental change, no significant differences were observed for both stimulus conditions. Thus, it is suggested that the prosodic change can be discriminate to the same degree as segmental change even for speech-modulated BCU. However, discrimination capability for speech-modulated BCU is slightly inferior to that for air-conducted speech.

Magnetoencephalographic study on forward suppression by ipsilateral, contralateral, and binaural maskers.

When two tones are presented in a short time interval, the response to the second tone is suppressed. This phenomenon is referred to as forward suppression. To address the effect of the masker laterality on forward suppression, magnetoencephalographic responses were investigated for eight subjects with normal hearing when the preceding maskers were presented ipsilaterally, contralaterally, and binaurally. We employed three masker intensity conditions: the ipsilateral-strong, left-right-balanced, and contralateral-strong conditions. Regarding the responses to the maskers without signal, the N1m amplitude evoked by the left and binaural maskers was significantly larger than that evoked by the right masker for the left-strong and left-right-balanced conditions. No significant difference was observed for the right-strong condition. Regarding the subsequent N1m amplitudes, they were attenuated by the presence of the left, binaural, and right maskers for all conditions. For the left- and right-strong conditions, the subsequent N1m amplitude in the presence of the left masker was smaller than those of the binaural and right maskers. No difference was observed between the binaural and right masker presentation. For left-right-balanced condition, the subsequent N1m amplitude decreased in the presence of the right, binaural, and left maskers in that order. If the preceding activity reflected the ability to suppress the subsequent activity, the forward suppression by the left masker would be superior to that by the right masker for the left-strong and left-right-balanced conditions. Furthermore, the forward suppression by the binaural masker would be expected to be superior to that by the left masker owing to additional afferent activity from the right ear. Thus, the current results suggest that the forward suppression by ipsilateral maskers is superior to that by contralateral maskers although both maskers evoked the N1m amplitudes to the same degree. Additional masker at the contralateral ear can attenuate the forward suppression by the ipsilateral masker.

Hearing aids reduce overestimation in pre-fitting self-assessment.

OBJECTIVE: When asking about hearing disability, the self-reported answers are not always equivalent to those of their family and neighbors. It is often experienced that family and neighbors indicate more severe hearing disability. Hearing difficulty itself may prevent hearing impaired subjects from understanding the degree of their own hearing disability. If their hearing impairment interferes with self-assessment of unaided hearing ability, it may change after a non-hearing aid user begins using hearing aids. METHODS: Thirty-four adults who had almost no experience with using hearing aids participated in this study. Unaided hearing disability was assessed with visual analog scale (VAS) and ten 5-point category scales. The assessment was performed not only by the subject (pre-fitting self-assessment) but also by his/her family members, presumed to be understanding persons regarding the degree of his/her hearing disability (family-assessment). For evaluating the effect of amplification on self-assessment, re-assessment was performed more than three months later (post-fitting self-assessment). The overestimation in the pre-fitting self-assessment and the effect of amplification were investigated in comparison to the family- and post-fitting self-assessments. RESULTS: The pre-fitting self-, family- and post-fitting self-VAS values for total hearing abilities were 50.5±15.6, 45.1±14.7 and 34.3±18.8, respectively. The pre-fitting self-VAS value was significantly higher than the family- and post-fitting self-VAS values (p<0.05 and p<0.01, respectively). For the 5-point category scale, the pre-fitting self-assessment was better than the family- and post-fitting self-assessments. Significant differences were observed in 1 and 2 situations, as compared with the family- and post-fitting self-assessments, respectively. These findings suggest the involvement of overestimation in pre-fitting self-assessment and its reduction after amplification with hearing aids. Although the pre-fitting self- and the family-VAS values were independent of the results of audiometric tests, the post-fitting self-VAS value was significantly related to the pure tone threshold and maximum speech recognition score (p<0.05). Regarding the changes in self-VAS values after amplification, the higher the pre-fitting self-VAS value, the larger the decrease in the self-VAS value. Thus, amplification with hearing aids may reduce overestimation and change self-assessment to reflect the audiometric assessments. CONCLUSION: The pre-fitting self-assessment of hearing disability involves over estimation which is exacerbated by hearing difficulty. A high pre-fitting VAS value may reflect a large overestimation in hearing ability. This overestimation can be improved by the amplification with hearing aids.

Duration-dependent growth of N1m for speech-modulated bone-conducted ultrasound.

Bone-conducted ultrasound (BCU) modulated by speech sound is recognized as speech sound and activates the auditory cortex similar to audible sound. To investigate the mechanisms of perception, the effects of stimulus duration on N1m were compared among air-conducted audible speech sound (AC speech), AC speech with carrier BCU and speech-modulated BCU in eight native Japanese with normal hearing. The Japanese vowel sound /a/ was used as a stimulus with durations of 10, 15, 20, 30, 40 and 60 ms. Comparison between AC speech with and without carrier showed that the presentation of carrier had no effect on N1m evoked by AC speech. Comparison among the three conditions showed that N1m amplitude for speech-modulated BCU differed from that for the two AC speeches. Moreover, N1m amplitude growth saturated at 40 ms for speech-modulated BCU, and at 20 ms for two AC speeches. These results suggest a difference in temporal integration of N1m between speech-modulated BCU and AC speech. Considering these results, it is reasonable to conclude that N1m evoked by speech-modulated BCU is influenced mainly by the ultrasonic component rather than demodulated audible sound. Given this finding, the notion needs to be considered that the mechanisms underlying perception and recognition of speech-modulated BCU depend on the ultrasonic component to some extent.

Peripheral perception mechanism of ultrasonic hearing.

Ultrasound can be perceived by bone conduction, and its characteristics differ from those of air-conducted audible sound (ACAS) in some respects. Despite many studies on ultrasonic hearing, the details have not yet been clarified. In this study, to elucidate the perception mechanism, the masking of bone-conducted ultrasound (BCU) produced by ACAS and the sensitivity of BCU in hearing impaired subjects were evaluated. We found that BCU was masked by high frequency ACAS, especially in the frequency range of 10-14 kHz. The most effective masker frequency depended on masker intensity. For hearing impaired subjects, the pure tone thresholds at 1-8 kHz and the maximum audible frequencies at cut-off intensities of 70-100 dB HL were significantly associated with the BCU threshold (p < 0.01 or p < 0.05). No subjects with estimated total loss of the inner hair cell system in the cochlear basal turn could hear BCU. These results suggest the peripheral perceptual region to be located in the cochlea. The results of masking show the faster excitation spread to the lower frequency range, depending on the intensity. This faster excitation spread may be due to nonlinearity in cochlear mechanics, which may work even without cochlear amplifier, and induce unique characteristics of BCU.