Author: Dr Raul Sanchez-Lopez1,2
1 Interacoustics Research Unit, Kgs. Lyngby, Denmark
2 Hearing Systems Section, Dep. Health Technology, Technical University of Denmark, Kgs. Lyngby, Denmark
Abstract: The hearing deficits are the perceptual consequences of different impairments in the auditory system. The sensorineural hearing loss is complex, and it is difficult to connect the perceptual deficits with specific impairments such as cochlear amplifier dysfunction or disruptions in the neural encoding of the incoming acoustic signals. Moreover, a detailed characterization of the auditory perceptual deficits might be more valuable than a detailed characterization of the impaired mechanisms, at least, for personalized audiology.
Previous studies have relied on correlations or regression analyses to shed light to the association between speech perception and other perceptual measures. In contrast, data-driven approaches, where an entire dataset formed by different of auditory perception is analyzed as a whole, can effectively provide patient subpopulations with distinct differences in terms of hearing abilities. This is the definition of “auditory profiling” used here.
In this contribution, the concept and principles of knowledge discovery from databases are revisited and applied on the scope of two recently published studies about auditory profiling. Here, the importance of tailoring the data-driven methods to a hypothesis towards solving a well imposed research question is stressed. The main aims were 1) to identify and characterized four clinically relevant subpopulations of hearing-impaired listeners, and 2) to explore hidden patterns of benefit in data from questionnaires about activity limitations and participation restrictions. Furthermore, these studies are analyzed in terms of findings, limitations, considerations, and perspectives towards personalized treatments.
As a final contribution, general opportunities and possible scenarios for personalized audiology are presented in association with auditory and/or audiological profiling, patterns of benefit, auditory ecology, genetics, preference profiles and other possible insights that data-driven approaches can offer for a better hearing rehabilitation.