Towards an Open-Source Precision Audiological Diagnostics Core for Large-Scale Data Analysis

Authors: Andrew N Sivaprakasam *1; Samantha Hauser 1 ; Natalie Seidl 1 ; Delaney Hake 2 ; Hari Bharadwaj 3; Michael G Heinz 2

Affiliations:
Purdue University
2 Watson Clinic, LLP
3 University of Pittsburgh

Background: Studies of human auditory perception and physiology can be bolstered by the collection and efficient pooling of audiological, imaging, and other data from thousands of participants across the human lifespan. Studies of such magnitude often require the help of costly or closed-source software tools owned by individual laboratories or institutions. As a result, the potential for researchers to systematically contribute data and metrics to answer questions requiring a big-data approach remains untapped. Here, we demonstrate progress towards an open-source collection of software and a data-analysis framework compatible with commonly used clinical equipment.

Methods: A software framework and data-analysis pipeline was developed by the Audiology Research Diagnostics Core (ARDC) at Purdue University as a collaborative effort between faculty, students, and clinicians with engineering and hearing-science backgrounds. Audiometry, QuickSIN, acoustic reflexes, tympanometry, and OAEs were collected using clinical-grade equipment (Grason-Stadler AudioStar and Interacoustics Titan) with added customizations and research software. Our analysis software (https://github.com/ARDCPurdue) condenses these measures into a singular data structure facilitating big-data analyses and targeted subject recruitment.

Results: To date, two laboratories have integrated the core and data structure for the collection of baseline audiological and survey measures necessary for EEG and hearing-aid research. The ARDC dataset is growing, with regular data collection from several labs each week.

Conclusion: The ARDC is facilitating the large-scale pooling of standard hearing measures collected at Purdue across numerous human hearing-science labs. A major hope of our open-source data-analysis toolkit is to enable other institutions collecting audiological measures to review, improve, and utilize these tools to answer current and future research questions requiring a large-scale approach.”