Brain-computer interfaces, long associated with restoring movement and communication, are being repositioned as tools for monitoring and treating mental health conditions. In an editorial published in Frontiers in Human Neuroscience, Bettina Habelt of Technical University Dresden and Andrei Dragomir of the National University of Singapore survey a collection of research articles on brain-computer interfaces, or BCIs, and argue that technologies built around electroencephalography, or EEG, could eventually extend beyond traditional sensorimotor rehabilitation toward personalised mental healthcare. The editorial frames the field at an inflection point: the underlying signals are increasingly well characterised, machine learning methods are becoming more capable, and yet awareness of BCIs among clinicians remains strikingly low.
The core scientific premise rests on EEG-based biomarkers. According to the editorial, such biomarkers have been proposed for integration into routine clinical practice because of their potential to characterise the pathophysiology of psychiatric disorders, identify vulnerable individuals, and predict treatment response and clinical trajectories. The goal is to support personalised treatment and improve therapeutic outcomes, moving psychiatric care away from trial-and-error prescribing toward decisions grounded in an individual’s brain activity patterns.
That promise, the editors note, has not yet translated into routine clinical awareness. Despite successful applications of BCIs in various clinical settings, clinicians remain largely unfamiliar with the technology, a gap that motivated one of the collected articles to provide a comprehensive overview of available BCI technologies.
One of the highlighted studies illustrates how even an adjunctive, non-pharmacological therapy can be probed with these tools. Belkacem and colleagues used EEG to investigate the effects of music therapy, a treatment used alongside standard care for psychiatric disorders. Applying computational modelling based on a deep residual multilayer perceptron, a type of deep neural network, to EEG recordings of neural oscillations captured during music and no-music sessions, the researchers demonstrated that music may modulate brain activity patterns in ways that improve the discriminability of mental fatigue states.
The editorial places this work in a broader tradition of oscillation-based diagnostics. Brain oscillation-based diagnostics and the assessment of treatment efficacy have been extensively investigated in motor disorders, particularly Parkinson’s disease, where reduced beta-band power modulation is associated with impaired voluntary movement initiation. Beta oscillations, rhythmic brain activity in a specific frequency range, have accordingly been explored as biomarkers for guiding targeted brain stimulation intended to alleviate motor symptoms such as tremor and rigidity.
Extending that logic from movement to rehabilitation, Su and colleagues proposed using the dynamics of sensorimotor beta-band activity as a basis for BCI-guided therapy in stroke rehabilitation. The aim is twofold: improving BCI performance, so that the interface can decode the user’s intent more reliably, and enhancing upper-limb functional recovery in stroke survivors. Because the protocol is a planned investigation rather than a completed trial, its value lies in the framework it establishes rather than demonstrated therapeutic gains.
Notably, the editors argue that the protocol’s relevance may reach beyond motor rehabilitation. They suggest it could provide a basis for exploring the broader applicability of BCI-based approaches to post-stroke non-motor dysfunction, given the interdependence of motor, cognitive, and emotional recovery.
A second highlighted study pushes EEG features directly toward psychiatrically relevant symptoms. Meng and colleagues used machine learning to investigate whether EEG features could predict non-motor symptoms in patients with Parkinson’s disease while also distinguishing those patients from healthy controls. They assessed EEG-derived functional connectivity, a measure of how strongly activity in different brain regions is coordinated, across four conditions: eyes closed, eyes open, and viewing happy or sad video clips. Overall, the patients showed a tendency toward increased functional connectivity strength across multiple frequency bands and conditions, a pattern that differed from the healthy comparison group.
The diagnostic and predictive results were condition- and frequency-specific. Gamma-band connectivity during eyes-open recordings achieved the highest accuracy among the individual conditions, whereas beta-band connectivity performed best when data from all four conditions were combined. Additionally, theta-band connectivity during the sadness condition predicted scores on the Pittsburgh Sleep Quality Index, a widely used self-report measure of sleep quality and disturbances that commonly accompany psychiatric disorders such as depression and anxiety.
None of this, the editors caution, will translate into psychiatric practice without confronting what Chandra and colleagues call the “three Es” of BCIs: Ethics, Efficiency, and Effectiveness. Ethical considerations include patient autonomy, privacy, and informed consent, and they extend beyond ordinary data breaches. Because BCIs communicate bidirectionally and in real time with external devices, they may enable unauthorised manipulation or actions that conflict with users’ intentions, and open questions remain about safeguarding mental privacy and determining who owns and controls neural data. In response, scholars have advocated recognising neuro-specific human rights to protect individuals from emerging neurotechnological risks. Efficiency concerns technical performance, reliability, usability, and practical feasibility, including resource requirements and suitability for clinical implementation, while effectiveness refers to the ability to achieve meaningful and sustained clinical benefits, improve patients’ quality of life, and minimise adverse effects. The editorial’s conclusion is measured: EEG-integrating BCIs offer promising opportunities, and emerging research highlights the potential of neural oscillations and functional connectivity to characterise disease-related alterations, while advanced machine learning may enable more accurate symptom prediction, patient classification, and identification of treatment-relevant neural patterns. But further research is needed to validate these approaches and establish their clinical effectiveness, reliability, and practical feasibility, and successful translation into psychiatric clinical practice will also require careful consideration of patient safety, autonomy, data privacy, and other ethical challenges.
Subject of Research: Brain-computer interfaces and EEG-based biomarkers for monitoring and treating mental health conditions
Article Title: Editorial: BCIs for monitoring and treating mental health conditions
Article References: Habelt, B., & Dragomir, A. (2026). Editorial: BCIs for monitoring and treating mental health conditions. Frontiers in Human Neuroscience, 20, Article 2039121. https://doi.org/10.3389/fnhum.2026.2039121
Image Credits: AI Generated
DOI: 10.3389/fnhum.2026.2039121
Keywords: brain-computer interfaces, EEG, mental health, neural oscillations, functional connectivity, machine learning, biomarkers, psychiatric disorders, neuroethics, stroke rehabilitation, Parkinson's disease, music therapy
News Source: Cassandra Pierce. (November 6, 2026). Brain-computer interfaces move toward mental health monitoring and treatment, editorial says. Scienmag.
