Self-Guided Beta-Responsive Visual Neurofeedback in a Public Exhibition: Feasibility and Usability Randomized Controlled Trial

Published November 3, 2026. In a public-exhibition test, 79 of 110 people who started an autonomous consumer-EEG neurofeedback session completed the task with a valid recording. The 52 people retained for analysis rated usability favorably on average. Adding a fixed slow-breathing guide did not improve usability, relaxation or mindfulness. This brief nonclinical study examined first-contact use, not symptom relief, neurofeedback learning or treatment effectiveness.

Baumgartner M, Weibel R, Hoesli T, von Wangenheim F, Ollier J.
JMIR Neurotechnology · Publication date: 2026-11-03
Added to site: 2026-11-08
DOI: 10.2196/89256

Published November 3, 2026. Added to Neurofeedback.com November 8, 2026 after comparison with fresh and reserve candidates. The source date is separate from the site addition date.

Modalities

Neurofeedback, EEG, Consumer neurotechnology

Clinical focus

Autonomous first-contact use, Usability and feasibility, Signal quality and onboarding

Population

Self-selected adults trying autonomous consumer-EEG neurofeedback at a public science exhibition in Bern, Switzerland

Sample

110 sessions initiated; 96 randomized; 79 completed with a valid EEG recording; 52 met final analytic criteria

Intervention / method

A tablet guided consent, self-fitting of a four-electrode consumer EEG headband, a 30-second eyes-closed baseline, randomization, a 120-second beta-responsive visual task and questionnaires without direct experimenter assistance.

Comparator

Visual neurofeedback alone versus the same visual neurofeedback plus a fixed open-loop breathing pacer; there was no sham-feedback, breathing-only or no-intervention control

Clinician summary

Of 96 randomized sessions, 17 failed the app's EEG contact-quality threshold and allocation could not be recovered; 79 completed with valid recordings and 52 met the final analytic rules. Mean System Usability Scale score was 72.3. The breathing-pacer comparison was null for usability, relaxation and state mindfulness. The exploratory resting-beta association with total usability did not survive correction across the broader EEG-outcome family.

How to interpret the findings

The trial was not prospectively registered, had no formal power calculation or prespecified progression criteria, and excluded many randomized participants from the final analysis. It used self-selected nonclinical visitors, one 120-second task and predominantly self-reported outcomes. There was no sham-feedback condition, raw EEG was not retained, manufacturer-derived band-power estimates were not independently validated and the eyes-closed baseline differed from the visual task. Harms and transient discomfort were not assessed systematically. Results do not establish beta suppression, learned self-regulation, durable home use, clinical benefit or superiority of autonomous delivery.

Equipment and disclosures

The study used a Muse 2 consumer EEG headband with four dry electrodes and manufacturer-processed Unity SDK features. It did not retain raw EEG, validate the features against research-grade recording or establish superiority for this or any other device. No equipment is attributed to BrainMaster.

Conducted at the Mobiliar Lab for Analytics at ETH Zurich with no dedicated external grant. The authors declared no conflicts. They disclosed supervised generative-AI assistance for code optimization, proofreading, translation and reformatting, with authors responsible for the final work. No correction, retraction or expression-of-concern relation appeared in the checked JMIR, Crossref and bounded DOI/title records on November 8, 2026. AI-assisted source review for Neurofeedback.com, not independent human clinical review.

Related evidence

Compare findings from the same clinical area, including mixed results.