Computational audiology is a branch of

audiology Audiology (from Latin , "to hear"; and from Greek , '' -logia'') is a branch of science that studies hearing, balance, and related disorders. Audiologists treat those with hearing loss and proactively prevent related damage. By employing vario ...

that employs techniques from mathematics and computer science to improve clinical treatments and scientific understanding of the auditory system. Computational audiology is closely related to computational medicine, which uses quantitative models to develop improved methods for general disease diagnosis and treatment.

Overview

In contrast to traditional methods in audiology and hearing science research, computational audiology emphasizes predictive modeling and large-scale analytics (" big data") rather than inferential statistics and small-cohort hypothesis testing. The aim of computational audiology is to translate advances in hearing science, data science, information technology, and machine learning to clinical audiological care. Research to understand hearing function and auditory processing in humans as well as relevant animal species represents translatable work that supports this aim. Research and development to implement more effective diagnostics and treatments represent

translational Translation is the communication of the meaning of a source-language text by means of an equivalent target-language text. The English language draws a terminological distinction (which does not exist in every language) between ''translat ...

work that supports this aim. For people with hearing difficulties,

tinnitus Tinnitus is the perception of sound when no corresponding external sound is present. Nearly everyone experiences a faint "normal tinnitus" in a completely quiet room; but it is of concern only if it is bothersome, interferes with normal hearin ...

hyperacusis Hyperacusis is the increased sensitivity to sound and a low tolerance for environmental noise. Definitions of hyperacusis can vary significantly; it can refer to normal noises being perceived as: loud, annoying, painful, fear-inducing, or a combina ...

, or balance problems, these advances might lead to more precise diagnoses, novel therapies, and advanced rehabilitation options including smart prostheses and e-Health/mHealth apps. For care providers, it can provide actionable knowledge and tools for automating part of the clinical pathway. The field is interdisciplinary and includes foundations in

auditory neuroscience Auditory means of or relating to the process of hearing: * Auditory system, the neurological structures and pathways of sound perception ** Auditory bulla, part of auditory system found in mammals other than primates ** Auditory nerve, also known ...

computer science Computer science is the study of computation, automation, and information. Computer science spans theoretical disciplines (such as algorithms, theory of computation, information theory, and automation) to practical disciplines (includin ...

data science Data science is an interdisciplinary field that uses scientific methods, processes, algorithms and systems to extract or extrapolate knowledge and insights from noisy, structured and unstructured data, and apply knowledge from data across a bro ...

machine learning Machine learning (ML) is a field of inquiry devoted to understanding and building methods that 'learn', that is, methods that leverage data to improve performance on some set of tasks. It is seen as a part of artificial intelligence. Machine ...

psychology Psychology is the scientific study of mind and behavior. Psychology includes the study of conscious and unconscious phenomena, including feelings and thoughts. It is an academic discipline of immense scope, crossing the boundaries betwe ...

signal processing Signal processing is an electrical engineering subfield that focuses on analyzing, modifying and synthesizing '' signals'', such as sound, images, and scientific measurements. Signal processing techniques are used to optimize transmissions, ...

natural language processing Natural language processing (NLP) is an interdisciplinary subfield of linguistics, computer science, and artificial intelligence concerned with the interactions between computers and human language, in particular how to program computers to proc ...

, and vestibulology.

Applications

In computational audiology,

models A model is an informative representation of an object, person or system. The term originally denoted the plans of a building in late 16th-century English, and derived via French and Italian ultimately from Latin ''modulus'', a measure. Models c ...

and

algorithms In mathematics and computer science, an algorithm () is a finite sequence of rigorous instructions, typically used to solve a class of specific problems or to perform a computation. Algorithms are used as specifications for performing ...

are used to understand the principles that govern the auditory system, to screen for hearing loss, to diagnose hearing disorders, to provide rehabilitation, and to generate simulations for patient education, among others.

Computational models of hearing, speech and auditory perception

For decades, phenomenological & biophysical (computational) models have been developed to simulate characteristics of the human auditory system. Examples include models of the mechanical properties of the basilar membrane, the electrically stimulated cochlea, middle ear mechanics, bone conduction, and the central auditory pathway. Saremi et al. (2016) compared 7 contemporary models including parallel filterbanks, cascaded filterbanks, transmission lines and biophysical models. More recently, convolutional neural networks (CNNs) have been constructed and trained that can replicate human auditory function or to reproduce complex cochlear mechanics with high accuracy. Although inspired by the interconnectivity of biological neural networks, the architecture of CNNs is distinct from the organization of the natural auditory system.

e-Health / mHealth (connected hearing healthcare, wireless- and internet-based services)

Online pure-tone threshold audiometry (or screening) tests, electrophysiological measures, for example distortion-product OAEs (DPOAEs) and speech-in-noise screening tests are becoming increasingly available as a tools to promote awareness and enable accurate early identification of hearing loss across ages, monitor the effects of ototoxicity and/or noise, and guide ear and hearing care decisions and support to clinicians. Smartphone-based tests have been proposed to detect middle ear fluid using acoustic reflectometry and machine learning. Smartphone attachments have also been designed to perform

tympanometry Tympanometry is an acoustic evaluation of the condition of the middle ear eardrum (tympanic membrane) and the conduction bones by creating variations of air pressure in the ear canal. Tympanometry is an objective test of middle-ear function. It is ...

for acoustic evaluation of the

middle ear The middle ear is the portion of the ear medial to the eardrum, and distal to the oval window of the cochlea (of the inner ear). The mammalian middle ear contains three ossicles, which transfer the vibrations of the eardrum into waves in ...

eardrum. Low-cost earphones attached to smartphones have also been prototyped to help detect the faint otoacoustic emissions from the cochlea and perform new-born hearing screening.

Big data and AI in audiology and hearing healthcare

Collecting

large numbers Large numbers are numbers significantly larger than those typically used in everyday life (for instance in simple counting or in monetary transactions), appearing frequently in fields such as mathematics, cosmology, cryptography, and statistical ...

of audiograms (e.g. the NIOSH or NHANES databases) provides researchers opportunities to find patterns of hearing status in the population or to train AI systems that can classify audiograms. Machine learning can be used to predict the relationship between multiple factors e.g. predict depression based on self-reported hearing loss or the relation between genetic profile and self-reported hearing loss. Hearing aids and wearable provide the option to monitor the soundscape of the user or log the usage patterns which can be used to automatically recommend settings that are expected to benefit the user.

Computational approaches to improving hearing devices and auditory implants

Methods to improve rehabilitation by auditory implants include improving music perception, models of the electrode-neuron interface, and an AI based Cochlear Implant fitting assistant.

Data-based investigations into hearing loss and tinnitus

Online surveys processed with ML-based classification have been used to diagnose somatosensory tinnitus. Automated NLP techniques, unsupervised and supervised Machine Learning have been used to analyze social posts about tinnitus and analyze the heterogeneity of symptoms.

Diagnostics for hearing problems, acoustics to facilitate hearing

Machine learning has been applied to audiometry to create flexible, efficient estimation tools that do not require excessive testing time to determine someone's individual's auditory profile. Similarly, machine learning based versions of other auditory tests including determining dead regions in the cochlea or equal loudness contours, have been created.

e-Research (remote testing, online experiments, new tools and frameworks)

Examples of e-Research tools include including the Remote Testing Wiki, the Portable Automated Rapid Testing (PART), Ecological Momentary Assessment (EMA) and the NIOSH soundlevel meter. A number of tools can be found online.

Software and tools

Software and large datasets are important for the development and adoption of computational audiology. As with many scientific computing fields, much of the field of computational audiology existentially depends on open source software and its continual maintenance, development, and advancement.

Related fields

Computational biology Computational biology refers to the use of data analysis, mathematical modeling and computational simulations to understand biological systems and relationships. An intersection of computer science, biology, and big data, the field also has fo ...

, computational medicine, computational pathology are all interdisciplinary approaches to the life sciences that draw from quantitative disciplines such as mathematics and

information science Information science (also known as information studies) is an academic field which is primarily concerned with analysis, collection, classification, manipulation, storage, retrieval, movement, dissemination, and protection of information. ...

References

{{reflist Audiology Auditory system Computational fields of study Computational science