Masashi Yanagisawa’s most immediate commercial opportunity is not another insomnia medicine. It is the measurement layer that tells researchers, doctors, employers and individuals what happens during sleep outside a laboratory. S’UIMIN, the University of Tsukuba spin-out he founded and chairs, has built a home service around electroencephalography, or EEG. Users attach sensors to the head, record sleep over several nights and receive analysis that goes beyond the movement estimates produced by most watches and rings.
In December 2025, S’UIMIN and Japanese research company Macromill launched an online sleep-study package that combines remote participant recruitment, home EEG measurement and data analysis. The offer is aimed partly at companies seeking evidence for foods, health products and other interventions. In June 2026, a separate University of Tsukuba collaboration with JMDC and Omron reported a model that used insurance claims, health-check and personal health-record data to predict sleep apnoea severe enough to require continuous positive airway pressure. Together, the projects point towards a broader business: sleep as a longitudinal data category connected to clinical risk, product trials and preventive care.
Yanagisawa, director of Tsukuba’s International Institute for Integrative Sleep Medicine, has already seen fundamental sleep biology become medicine. His discovery of orexin and its role in wakefulness helped establish drug targets for insomnia and narcolepsy. The home-data strategy tackles a different bottleneck. Sleep disorders are underdiagnosed, subjective reports are unreliable, and laboratory polysomnography is expensive and inconvenient. The 2026 leadership test is whether cheaper measurement produces better decisions rather than a larger market for anxious consumers and weak wellness claims.
Objective sleep moves into the home
Sleep has attracted a crowded device market because the signal appears easy to collect. A wrist or ring can measure movement, pulse and temperature throughout the night. Algorithms infer stages and produce a score. The products are convenient and useful for tracking habits, but they do not directly measure the brain’s electrical activity, the conventional basis for distinguishing sleep stages.
S’UIMIN’s proposition is to bring more clinically relevant measurement into a normal bedroom. Multi-night recording may capture variability that a single laboratory night misses. Remote deployment reduces travel and facility cost. For research sponsors, it can expand recruitment and produce data under real-life conditions. For clinicians, it could identify people who need formal evaluation. For individuals, it can distinguish a perceived problem from an objectively abnormal pattern.
A 2025 study involving Yanagisawa’s group and S’UIMIN data found discrepancies between subjective and objective sleep assessment during real-world home EEG. People were not consistently able to judge sleep depth, brief awakenings or sleep-apnoea risk. That supports the demand for measurement, but it does not mean every person needs EEG. Screening creates value when it leads to an intervention with a clear benefit. Otherwise, it adds cost and may medicalise normal variation.
The commercial advantage over wearables therefore needs to be expressed in decisions, not precision alone. Does home EEG refer the right patients for apnoea testing? Does it change insomnia treatment? Does it show that a food or behavioural programme improves a recognised endpoint? A technically richer sleep graph is not an outcome.
Trials become a service business
The Macromill partnership broadens S’UIMIN from selling measurements to operating studies. Participants can be recruited, equipped and followed remotely; sleep data can be combined with questionnaires and other trial information. This can reduce the friction of running geographically distributed studies and increase sample size. Japan’s market for foods with functional claims gives the service an immediate customer base.
That same market creates reputational risk. Sleep is a popular marketing claim for supplements, beverages, bedding and consumer devices. Sponsors may prefer positive results, endpoints can be selected after data are visible, and statistically significant changes may not matter clinically. S’UIMIN’s scientific credibility becomes valuable only if study design, analysis and reporting remain independent enough to disappoint a paying customer.
The economics of a research service also differ from a scalable software platform. Devices must be shipped, maintained and cleaned; users need support; noisy recordings may require review; every study has a protocol. Automation can improve margins, especially if algorithms score EEG reliably, but human quality control remains important. Expansion outside Japan adds language, logistics and regulatory costs.
A defensible data asset could emerge if consent permits learning across studies. Multi-night EEG linked to demographics, symptoms, interventions and outcomes would be valuable for drug development and diagnostics. Yet sponsors may restrict reuse, and participants may not expect a commercial platform to retain intimate physiological information. Data rights must be clear at collection. Ambiguous consent may create a short-term dataset and a long-term liability.
Sleep apnoea links consumer data to medical spending
The 2026 Tsukuba-JMDC-Omron work shows another route: identify high-risk people before an expensive diagnostic pathway. The model combined claims, routine health checks and lifelog information to predict cases of sleep apnoea treated with continuous positive airway pressure. Such an approach can focus testing on people most likely to benefit, especially when many cases remain hidden.
For insurers and employers, the business case is stronger than a generic sleep score. Untreated apnoea is associated with cardiovascular risk, daytime impairment and accidents; effective identification may reduce downstream cost and improve productivity. Claims data can reveal comorbidities and prior care, while consumer devices add behavioural context. A targeted home EEG or oxygen assessment could then refine referral.
Prediction from administrative records has limitations. A model trained on people who received treatment may learn access and prescribing patterns as much as disease. It can miss undiagnosed patients whose records contain no obvious signal. Performance may change when reimbursement rules or health-check practices change. External validation across employers, regions and care systems is essential before automated outreach.
The partnership landscape also raises questions about who owns the customer. JMDC has claims and health data; Omron has devices and distribution; universities contribute methods; S’UIMIN offers sleep expertise and EEG. The company that controls the integrated workflow—from risk identification to measurement and referral—will capture more value than a component supplier. Yanagisawa’s influence can convene the ecosystem, but commercial agreements must define where scientific collaboration ends and proprietary advantage begins.
Asia is the natural market and the hardest comparison
Japan’s ageing population, intense work culture and mature employer-health programmes make it an attractive launch market. Sleep problems intersect with hypertension, metabolic disease, cognitive decline and workplace safety. A trusted home test can serve hospitals that cannot bring every concerned patient into a sleep laboratory. Corporate programmes can reach people before they seek care.
Across Asia, the opportunity expands with urbanisation, shift work and rising chronic disease. Singapore and South Korea have strong digital-health infrastructure; China offers scale and device manufacturing; India and Southeast Asia have large underdiagnosed populations but fewer sleep specialists per patient. Remote measurement can extend capacity, although price and internet access shape who benefits.
Algorithms will not travel automatically. Sleep schedules, housing, noise, occupation, comorbidities and healthcare access differ. Skin, hair and user behaviour can affect sensor attachment and signal quality. A model trained on Japanese participants needs local validation before it is used to reassure or refer people elsewhere. The same applies within Japan across age and disease groups.
Regulatory classification may vary depending on whether a product offers wellness information, screening or diagnosis. Crossing that boundary can expand clinical value while increasing evidence, quality and surveillance obligations. S’UIMIN must be precise about which claims each service supports. The temptation to use medical language for consumer marketing is particularly dangerous in sleep, where people are vulnerable to simple explanations for complex fatigue.
Measurement needs an intervention
Yanagisawa’s orexin work offers a useful comparison. The discovery created value because it identified a mechanism that medicines could change, and clinical trials connected that change to wakefulness or sleep. Home EEG is an enabling technology. Its impact depends on what follows: continuous positive airway pressure for apnoea, behavioural treatment for insomnia, medication review, schedule changes or a validated therapeutic trial.
S’UIMIN should therefore be judged on a funnel, not the number of nights recorded. How many tests are interpretable? How many identify a clinically important problem? How many users reach appropriate care? How many improve? What is the cost per useful intervention compared with questionnaires, wearables or laboratory testing? Those measures will tell health systems whether the platform expands capacity or adds another layer of screening.
Privacy belongs in the same scorecard. Sleep data can reveal routines, health conditions and periods of vulnerability. When combined with employment or insurance records, it can influence decisions beyond care. Employers should not receive individual results without explicit, freely given consent. Insurers should not turn preventive screening into discriminatory pricing. Strong separation between wellness programmes and employment decisions is part of the product’s licence to operate.
Yanagisawa has moved sleep science from a neuropeptide to a pharmaceutical field and now towards a home-data industry. The next proof is not that subjective sleep can be wrong; that is already clear. It is that objective measurement at scale sends the right people towards effective care, improves outcomes and protects the intimacy of the night. If S’UIMIN can demonstrate that chain, home EEG becomes healthcare infrastructure. Without it, the company risks producing the most accurate version yet of a sleep score that changes little after breakfast.
