Logo image
Back
AudioChip: A Deep Phenotyping Approach for Deconstructing and Quantifying Audiological Phenotypes of Self-Reported Speech Perception Difficulties
Journal article   Peer reviewed

AudioChip: A Deep Phenotyping Approach for Deconstructing and Quantifying Audiological Phenotypes of Self-Reported Speech Perception Difficulties

Ishan Sunilkumar Bhatt, Raquel Dias, Nathan Wineinger, Sheila Pratt, Jin Wang, Nilesh Washnik, O'neil Guthrie, Jason Wilder and Ali Torkamani
Ear and hearing, Vol.43(3), pp.1023-1036
05/2022
DOI: 10.1097/AUD.0000000000001158
PMCID: PMC9010350
PMID: 34860719
url
https://www.ncbi.nlm.nih.gov/pmc/articles/9010350View
Open Access

Abstract

About 15% of U.S. adults report speech perception difficulties despite showing normal audiograms. Recent research suggests that genetic factors might influence the phenotypic spectrum of speech perception difficulties. The primary objective of the present study was to describe a conceptual framework of a deep phenotyping method, referred to as AudioChipping, for deconstructing and quantifying complex audiometric phenotypes. In a sample of 70 females 18 to 35 years of age with normal audiograms (from 250 to 8000 Hz), the study measured behavioral hearing thresholds (250 to 16,000 Hz), distortion product otoacoustic emissions (1000 to 16,000 Hz), click-evoked auditory brainstem responses (ABR), complex ABR (cABR), QuickSIN, dichotic digit test score, loudness discomfort level, and noise exposure background. The speech perception difficulties were evaluated using the Speech, Spatial, and Quality of Hearing Scale-12-item version (SSQ). A multiple linear regression model was used to determine the relationship between SSQ scores and audiometric measures. Participants were categorized into three groups (i.e., high, mid, and low) using the SSQ scores before performing the clustering analysis. Audiometric measures were normalized and standardized before performing unsupervised k-means clustering to generate AudioChip. The results showed that SSQ and noise exposure background exhibited a significant negative correlation. ABR wave I amplitude, cABR offset latency, cABR response morphology, and loudness discomfort level were significant predictors for SSQ scores. These predictors explained about 18% of the variance in the SSQ score. The k-means clustering was used to split the participants into three major groups; one of these clusters revealed 53% of participants with low SSQ. Our study highlighted the relationship between SSQ and auditory coding precision in the auditory brainstem in normal-hearing young females. AudioChip was useful in delineating and quantifying internal homogeneity and heterogeneity in audiometric measures among individuals with a range of SSQ scores. AudioChip could help identify the genotype-phenotype relationship, document longitudinal changes in auditory phenotypes, and pair individuals in case-control groups for the genetic association analysis.
 Auditory Threshold Evoked Potentials, Auditory, Brain Stem Female Humans Male Phenotype Self Report Speech Perception

Details

Metrics

19 Record Views
Logo image