Explainable machine learning reveals the relationship between hearing thresholds and speech-in-noise recognition in listeners with normal audiograms

Some individuals complain of listening-in-noise difficulty despite having a normal audiogram. In this study, machine learning is applied to examine the extent to which hearing thresholds can predict speech-in-noise recognition among normal-hearing individuals. The specific goals were to (1) compare the performance of one standard (GAM, generalized additive model) and four machine learning models (ANN, artificial neural network; DNN, deep neural network; RF, random forest; XGBoost; eXtreme gradient boosting), and (2) examine the relative contribution of individual audiometric frequencies and demographic variables in predicting speech-in-noise recognition. Archival data included thresholds (0.25-16 kHz) and speech recognition thresholds (SRTs) from listeners with clinically normal audiograms (n = 764 participants or 1528 ears; age, 4-38 years old). Among the machine learning models, XGBoost performed significantly better than other methods (mean absolute error; MAE = 1.62 dB). ANN and RF yielded similar performances (MAE = 1.68 and 1.67 dB, respectively), whereas, surprisingly, DNN showed relatively poorer performance (MAE = 1.94 dB). The MAE for GAM was 1.61 dB. SHapley Additive exPlanations revealed that age, thresholds at 16 kHz, 12.5 kHz, etc., on the order of importance, contributed to SRT. These results suggest the importance of hearing in the extended high frequencies for predicting speech-in-noise recognition in listeners with normal audiograms.

Effect of Sensory Modality on Reaction Time in Individuals with Auditory Neuropathy Spectrum Disorder

PURPOSE

To investigate and compare the reaction time of individuals with auditory neuropathy in three modalities, auditory, visual, and audio-visual. The reaction time of individuals with auditory neuropathy was also compared with those with normal hearing. The relationship between reaction time across modalities and the duration of hearing loss in auditory neuropathy was also investigated.

METHODS AND MATERIALS

The reaction time of adults with auditory neuropathy and those with normal hearing was measured in the three modalities using the Choice reaction time task.

RESULTS

The auditory neuropathy group significantly had a longer reaction time than the normal hearing in all modalities. The trend of the mean reaction time differed across groups. Further, a significant difference in reaction time of the auditory neuropathy group was noted between auditory and visual mode, auditory and audio-visual mode. However, no significant difference between visual and audio-visual modalities was noted in reaction time.

CONCLUSION

Significantly longer reaction time in auditory neuropathy is presumed to have resulted from neural conduction delay and impaired processing. The auditory neuropathy group can utilize visual cues for faster processing, and the study recommends an audio-visual mode for their management. In addition, the duration of hearing loss in auditory neuropathy had no relationship with reaction time across all modalities.

Dynamics of Speech Perception in the Auditory-Visual Mode: An Empirical Evidence for the Management of Auditory Neuropathy Spectrum Disorders

BACKGROUND AND OBJECTIVES

The present study probed into the relative and combined contribution of auditory and visual modalities in the speech perception of individuals with auditory neuropathy spectrum disorders (ANSD). Specifically, the identification scores of consonantvowel (CV) syllables, visual enhancement (VE), and auditory enhancement in different signal to noise ratios (SNRs) were compared with that of the control group. Subjects and.

METHODS

The study used a repeated measure standard group comparison research design. Two groups of individuals in the age range of 16 to 35 years participated in the study. The clinical group included 35 participants diagnosed as ANSD, while the control group had 35 age and gender matched individuals with typical auditory abilities. The participants were assessed for CV syllable identification in auditory only (A), visual only (V), and auditory-visual (AV) modalities. The syllables were presented in quiet and at 0 dB SNR.

RESULTS

The speech identification score was maximum in AV condition followed by A-condition and least in V condition. This was true in both the groups. The individuals with ANSD were able to make better use of visual cues than the control group, as evident in the VE score.

CONCLUSIONS

The dynamics of speech perception in the AV mode is different between ANSD and control. There is definite benefit of auditory as well as visual cues to individuals with ANSD, suggesting the need to facilitate both the modalities as part of the audiological rehabilitation. Future studies can focus on independently facilitating the two modalities and testing the benefits in the AV mode of speech perception in individuals with ANSD.