Artificial Intelligence in Audiology: Computational Auditory Neuroscience and Neural Decoding

BackgroundArtificial intelligence (AI) and modern machine learning are rapidly reshaping how we study, model, and measure hearing, from mechanistic questions in auditory neuroscience to clinically relevant problems in audiology. Large-scale neural recordings, advanced auditory neuroimaging, speech and sound datasets, and real-world device data are enabling computational approaches that were previously impractical. At the same time, progress in neural decoding and representation learning is opening new ways to link sensory stimuli and behavior to neural activity across the auditory pathway.Despite these advances, key gaps remain. Models that perform well in controlled settings often fail to generalize across subjects, devices, environments, and clinical populations. There is also a need for more interpretable and biologically grounded AI approaches that can generate testable hypotheses about auditory processing, rather than acting as “black boxes.” Finally, translation to clinical decision-making and hearing technologies requires robust validation, careful attention to bias and uncertainty, and reproducible pipelines that can be adopted beyond a single lab.GoalThis Research Topic aims to bring together cutting-edge work at the intersection of AI, auditory neuroscience, and audiology, with a focus on computational frameworks that (i) advance mechanistic understanding of auditory processing, and (ii) enable practical decoding, prediction, and decision support relevant to hearing research and clinical care. By consolidating methods, benchmarks, and translational examples, this collection will help clarify where AI is already impactful in audiology and where the next breakthroughs are needed for reliable, interpretable, and clinically meaningful deployment.Scope (non-exhaustive)We welcome submissions including (but not limited to) the following themes:o Neural decoding for auditory stimuli and perception, including decoding of speech, pitch, timbre, attention, and auditory scene analysis from invasive and non-invasive recordings (e.g., EEG, MEG, ECoG, intracranial recordings)o Computational auditory neuroscience, including biologically inspired models, encoding models, normative models, and hypothesis-driven AI that links neural dynamics to behavioro Self-supervised and foundation models for audio and speech, and their use in modeling auditory representations or predicting neural responseso Multimodal modeling combining acoustics, neural data, behavior, and clinical measures (e.g., audiometry, speech-in-noise outcomes, device logs)o Clinical translation and decision support in audiology, including AI for diagnosis support, outcome prediction, rehabilitation planning, and patient stratification (with appropriate validation)o Model interpretability, uncertainty, and robustness, including explainable AI, calibration, domain shift, cross-site generalization, and reliability across populations and recording conditionso Benchmark datasets, evaluation standards, and reproducible pipelines, including open datasets, shared code, standardized preprocessing, and reporting guidelines for AI in auditory researcho Ethics, fairness, and privacy considerations for AI applied to neural and clinical auditory data, including bias assessment and responsible data governanceArticle types: Original Research, Methods/Technology and Code, Brief Research Report, Mini Review, Review, Systematic Review, Perspective, Hypothesis and Theory. Topic editors: Mehdi Abouzari, Hovannes Kulhandjian, Neeraj Singh Thakur Source: https://www.frontiersin.org/research-topics/77984/artificial-intelligence-in-audiology-computational-auditory-neuroscience-and-neural-decoding