Categorization of Environmental Sounds Based on Pitch-Spectral Correlation and Preliminary Findings in Clinical Population

Authors: Megha Sasidharanl1, Rashmi J Bhat1, Kishan Madikeri Mohan2, Anna Rajan1

1Dr S R Chandrasekhar Institute of Speech and Hearing

Background: Environmental sounds are commonly encountered in our daily lives and have been studied in hearing loss and cochlear implant users. The aim of the study was to adapt an environmental sound test for open set task in an Indian context. The subjective pitch and acoustic spectrum of the sounds were correlated and they were categorised based on frequency. It also explored the preliminary findings of the test in normal hearing, sensorineural hearing loss (SNHL), and auditory neuropathy spectrum disorder (ANSD).

Method: Familiarity rating and content validity of the sounds were performed. Subjectively these sounds were rated as high pitch, low pitch or uncategorized by speech and hearing professionals. Objectively the first five spectral peaks were identified for the same sounds using PRAAT software and categorised as low or high frequency. The pitch and spectral analyses correlated for the selected sounds based on which they were divided as low and high frequency sounds. The sounds that did not fall into either were labelled as uncategorized. Acceptable naming was obtained for the final list of sounds in Kannada language. The sound identification scores were obtained for individuals with normal hearing, SNHL and ANSD.

Results: From a total of 185 exemplars, the final list had 37 exemplars each representing one sound source, with 14 low frequency and 10 high frequency sounds. The results showed a statistically significant difference between the three groups, with the best performance in normal hearing, followed by SNHL and ANSD. SNHL had significantly poorer scores for low frequency sounds than high frequency. Similar results were obtained in ANSD but were not statistically significant.

Conclusion: The tool can assess auditory disorders in the Indian context. It demonstrated the effect of poor phase locking of low frequency sounds and differentiated SNHL and ANSD. Thus, it can be a potential diagnostic test as non-linguistic material with natural stimuli.