Earbuds that read the brain: review maps the rise of in-ear EEG

Electroencephalography has been one of medicine’s most trusted windows into the living brain for decades, but it has always come with a price: a cap studded with electrodes, conductive gel rubbed into the hair, specialised amplifiers and a trained technician to put it all in place. A systematic review published in Frontiers in Neurology argues that a quieter revolution is underway at the other end of the scale. Researchers from Imperial College London and Luton & Dunstable University Hospital analysed 24 peer-reviewed studies published from 2018 onwards and concluded that in-ear EEG, in which electrodes sit inside or around the external auditory canal, has matured into a credible platform for monitoring brain activity during sleep, seizures, cognitive workload, fatigue and even emotional states, while still falling short of conventional scalp recordings in signal quality and spatial coverage.

The review, conducted according to PRISMA guidelines with searches of PubMed and Google Scholar completed in January 2026, focused on four questions: how in-ear EEG devices have been developed, how their signal quality compares with scalp EEG, what clinical and cognitive applications have been reported, and what limitations remain. The 24 included studies were analysed thematically, and each was graded with a framework adapted from the Oxford Centre for Evidence-Based Medicine levels of evidence. The authors note that the field traces back to the early 2010s, when a study showed that EEG signals could be recorded from electrodes embedded in custom ear moulds positioned in the ear canal. Since then, research into ear-EEG architectures, electrode materials and recording configurations has expanded rapidly, driven by the same miniaturisation wave that produced modern earbuds and hearing aids.

Technologically, today’s in-ear systems are typically small, lightweight and discreet, often integrated into earbud-like earpieces or modified earphones. Compared with conventional scalp EEG they use far fewer electrodes, commonly one to three channels, and primarily capture activity from brain regions close to the temporal lobe. Design innovation has centred on fit and wearability: custom-moulded earpieces have been developed to improve electrode-skin contact and reduce discomfort caused by variation in ear canal anatomy, while other groups have pursued more ergonomic earbud designs intended to hold the sensors stable during daily activities. Portability is the decisive advantage. Unlike traditional systems requiring extensive setup and conductive gel, in-ear devices can often be inserted like ordinary earphones without specialist assistance, and the stable positioning of the ear canal may help maintain consistent electrode contact and reduce certain types of motion artefact.

The trade-off is signal fidelity. Because in-ear systems use fewer electrodes and record from a restricted anatomical region, their signals are generally weaker than scalp recordings, yielding lower signal-to-noise ratios. The review highlights this as the most frequently reported limitation. Moumane and colleagues, in a study of 30 healthy adults, found that an in-ear device showed spectral patterns comparable to a 64-channel scalp cap during rest and naps, indicating feasibility for monitoring wakefulness and sleep, but the device was more susceptible to movement-related artefacts during sleep recordings and showed lower signal-to-noise ratios, particularly during more dynamic or demanding tasks. Crétot-Richert and colleagues similarly reported lower signal quality and reduced sensitivity for certain signal components, particularly in the time domain, although their study of 15 healthy adults found that ear-centred EEG could reliably detect differences in mental focus and cognitive workload, with results similar to traditional scalp EEG.

Spatial resolution is the second structural constraint. Conventional scalp EEG covers broad cortical territory, whereas in-ear EEG primarily captures activity near the temporal lobe, which may reduce performance in tasks involving distributed brain networks or more complex cognitive processes. Tremmel and colleagues, testing a commercial in-ear device in 17 healthy adults, found statistically significant performance in estimating cognitive workload, but conventional EEG achieved higher accuracy. Nevertheless, the review concludes that in-ear EEG remains sufficiently informative for targeted applications such as workload estimation, sleep analysis, fatigue monitoring and emotion detection. Some teams have attacked the coverage problem directly: Meiser and colleagues reported that higher-density ear-EEG systems improved spatial sampling and enhanced performance in measuring auditory evoked and event-related potentials, suggesting that multi-electrode ear arrays may eventually narrow the gap without sacrificing wearability.

The most mature application area, according to the review, is clinical monitoring, particularly sleep and epilepsy. Nakamura and colleagues validated a standardised in-ear EEG platform in 22 healthy adults, showing that unattended overnight sleep monitoring with automatic sleep-stage prediction reached substantial agreement with polysomnography, potentially offering a less intrusive way to assess sleep in community settings. Henao and colleagues showed that in-ear EEG can pick up slow oscillations during sleep in 24 adults, suggesting ear-EEG might suit sleep intervention systems, and other work has suggested ear-EEG may be suitable for long-term home-based monitoring outside traditional laboratory environments. On the epilepsy side, Zibrandtsen and colleagues reported in a proof-of-concept study of epilepsy patients that ear-EEG derived signals detected all three recorded generalised tonic-clonic seizures with 100 percent sensitivity and no false positives, supporting nocturnal seizure monitoring. Joyner and colleagues, in a validation study of 20 analysed patients with refractory focal epilepsy, found that an in-ear device detected focal-onset seizures with 86.4 percent sensitivity compared with simultaneous video EEG, and patients generally tolerated prolonged use well.

Beyond these headline uses, the review catalogues a striking breadth of applications. Long-term outpatient ear-EEG monitoring detected interictal epileptiform discharges in 80 percent of patients with dementia with Lewy bodies in one observational study of 10 patients and 15 controls, with most discharges occurring at night and unseen on short scalp EEG recordings; a companion study in mild to moderate Alzheimer’s disease detected epileptiform discharges in most of 24 patients but none on short scalp recordings in the same format. Athavipach and colleagues reported that a low-cost generic in-ear prototype could classify emotional states based on valence and arousal with performance comparable to temporal scalp sites. Other studies have explored drowsiness detection during simulated driving, stress monitoring with multimodal headsets combining in-ear EEG and ECG, attention-state classification using machine learning, objective hearing threshold estimation, and even three-factor biometric authentication using custom-fit in-ear electrodes.

The authors are careful about what in-ear EEG cannot do. Because of its limited spatial coverage, they argue it is unlikely to replace conventional scalp EEG for comprehensive neurophysiological assessment; seizures arising from frontal, occipital or other extra-temporal regions may be less readily detected, which matters in extra-temporal or multifocal epilepsy. Current systems also generally lack the synchronised video of conventional video-EEG, limiting the distinction between epileptic, nonepileptic and physiological events. Motion artefacts from jaw movement, speech and head motion remain a significant obstacle, and early studies have shown that physiological factors including earwax can substantially increase dry-contact impedance and degrade signal quality. Anatomical variability affects fit and reproducibility, and custom-moulded devices, while helpful, add manufacturing complexity that may limit scalability. Prolonged wear may still cause discomfort, pressure or irritation, and battery life and durable lightweight hardware remain practical concerns.

A further weakness lies in the evidence base itself. Most of the 24 studies recruited small cohorts of young, healthy adults, often with gender imbalance, while relatively few included older adults or patients with neurological disease, and several studies did not report participants’ age or gender at all, limiting assessment of generalisability and selection bias. The review’s own search was restricted to PubMed and Google Scholar, to English-language articles and to studies from 2018 onwards, and the included studies varied so substantially in device design, protocols and populations that quantitative meta-analysis was not possible. Looking forward, the authors call for larger, more diverse and better-characterised cohorts, head-to-head comparisons with scalp EEG, multicentre prospective validation, standardised reporting protocols, direct comparisons of commercial and research-grade devices, and longer follow-up of user comfort, adherence and durability. Advances in flexible electrode materials such as graphene and conductive polymers, adaptive filtering, independent component analysis, wavelet-based denoising and machine learning, they conclude, may carry in-ear EEG from a promising experimental technology to a practical tool for continuous brain monitoring in clinics and everyday life.

Subject of Research: In-ear electroencephalography technology, applications and challenges examined in a systematic review

Article Title: In-ear electroencephalography: a systematic review of technology, applications, challenges, and future directions

Article References: Haria, C., Harris, B. H. L., Mortimer, L., Platek-Johnson, K., Freilich, S., Koizia, L. J., & Fertleman, M. B. (2026). In-ear electroencephalography: a systematic review of technology, applications, challenges, and future directions. Frontiers in Neurology, 17, 1914159. https://doi.org/10.3389/fneur.2026.1914159

Image Credits: AI Generated

DOI: 10.3389/fneur.2026.1914159

Keywords: in-ear EEG, ear-EEG, wearable neurotechnology, systematic review, epilepsy monitoring, sleep monitoring, cognitive workload, signal quality, motion artefacts, scalp EEG comparison, brain-computer interface, digital health