Psst… We received a tip…

You are sitting in a waiting room with a clipboard and questionnaire. How often have you felt hopeless? How has your sleep been? Are you avoiding certain places, people, or memories? Are there thoughts you cannot shake?

You answer as honestly as you can. The clinician reads it back, asks follow-up questions, and starts building a picture.

That picture may take time. Repeated assessments, symptom thresholds, clinical judgment, and the patient slowly finding the words for what is actually happening.

A small startup out of South Korea thinks a headset and two audio tones can help. This week’s patent from Bwave Corporation describes a system that reads your brain’s electrical response to short beeps, then uses AI to flag whether you are more likely living with PTSD, depression, both, or neither.

This patent originates from Dr. Seung-Hwan Lee, a practicing professor of psychiatry at Inje University’s Ilsan Paik Hospital in Goyang-si, South Korea. A clinician with over two decades of experience treating patients with depression, PTSD, and other mood disorders.

After more than 20 years accumulating brainwave data from psychiatric patients, he founded Bwave in April 2019 to commercialise what he had been building in the clinic. The company is small, seed-funded, and operating squarely at the intersection of academic psychiatry and commercial software. The company’s public ambition is to reduce a brainwave-based depression assessment from hours to minutes. That claim needs care. A faster assessment tool is not the same thing as saying a full psychiatric diagnosis is neatly settled in one sitting. 

The patent describes a system that reads a patient’s brainwave response to a simple sound, then uses AI to identify which mental disorder they are likely experiencing. Think of it as a biological clue for the clinician, especially when two conditions look similar on the surface.

Bwave’s system works like this. The patient puts on an EEG headset and listens to two tones played in random order. A tone here just means a short beep at a particular pitch. One beep is common. One beep is rare. The patient presses a button when they hear the rare one.

That is it. But the brain’s electrical response to that rare beep, measured in milliseconds, produces what researchers call a P300 response. In Dr. Lee’s research, those P300 patterns looked measurably different across patients with PTSD, patients with major depressive disorder, and healthy controls. (PubMed)

The system captures that difference, maps where in the brain the activity is coming from, and feeds it into a classification model trained on data from real patients across both conditions. A result gets pushed to a mobile device within the same session.

Mental health has a measurement problem. More than 1 billion people are estimated to be living with mental health disorders globally, and depression is one of the major contributors to that burden. But diagnosis still depends heavily on interviews, questionnaires, and what a patient can explain in the room. Useful, yes. Fast, cheap and scalable? Not really. (World Health Organization)

That gets especially messy when conditions overlap. A meta-analysis of 57 studies found that 52% of people with current PTSD also had major depressive disorder, which is the kind of blur a biological signal could help cut through. (PubMed)

The opportunity is diagnostic discernment. Bwave is trying to make assessment faster and help clinicians separate conditions that can look almost identical at the symptom level but may require different treatment decisions.

Take the example of a veteran returning from active service. They are struggling. Sleep is broken, concentration is gone, mood is flat. A GP refers them to a psychiatrist, who immediately faces a hard problem, that PTSD and depression share many symptoms on the surface. Distinguishing between them can take careful interviews, repeated assessments, and time.

This is important for patients who cannot narrate their symptoms cleanly. A person with PTSD may report broken sleep, flat mood, and poor concentration, which can sound very similar to depression. The missing detail may be that their sleep is being broken by trauma dreams, that they are avoiding specific reminders, or that the conversation itself is triggering.

A tool that adds biological context could be valuable if it helps clinicians see what the patient cannot yet explain.

The broader category is called digital biomarkers, the idea that mental health symptoms can be captured as measurable signals rather than relying solely on what a patient reports in a clinical setting.

The closest comparison is Thymia, a London-based startup with an unusually personal origin story.

Its founder, neuroscientist Dr. Emilia Molimpakis, started the company after her best friend’s depression went undetected by multiple clinicians. Her response was to build better tools. Thymia’s platform runs patients through AI-powered video games while reading voice patterns, facial expressions, movement, and in-game behaviour to help assess depression, anxiety, and ADHD. It has raised $3.5 million in seed funding and is working toward medical device approval.

Other companies are chasing the signal through voice.

Ellipsis Health and Sonde Health both offer mental health screening built around short voice samples. Easy to scale, and cheap to deploy. The approach has real appeal, though the regulatory path has proven harder than it looks.

Alto Neuroscience is playing a different game entirely. Rather than screening for conditions, it uses EEG activity, cognitive assessments, wearable data, and other biomarkers to match patients with the right psychiatric drugs. It has raised $75 million in total funding, with Novartis among its backers. That is a different value proposition from a diagnostic tool, but the underlying logic rhymes: psychiatry is moving from “what symptoms do you report?” toward “what measurable biological pattern do you fit?” (Alto Neuroscience Investors)

The EEG devices market sits at around $1.2 billion today and is projected to nearly double to $2.4 billion by 2030, with disease diagnosis as a major application segment. (Grand View Research)

Zoom out to digital biomarkers broadly and the numbers are considerably larger. The global digital biomarkers market was estimated at $5.55 billion in 2024 and is projected to reach $35.8 billion by 2035, growing at 18.5% annually. The mental health technology market sits above that, valued at $15.22 billion in 2024 and expected to reach $30.98 billion by 2030. (Roots Analysis)

An EEG-based PTSD-versus-depression assessment tool is a small slice of the broader EEG, digital biomarker, and mental health technology markets. The same pressure sits underneath all of them, that mental health systems are overwhelmed, clinician supply is stretched, and health systems are looking for tools that can support faster, cheaper, and more consistent assessment.

One geography worth watching is Asia-Pacific. The region is expected to show considerable growth in digital biomarkers, and some forecasts put the category at around 20% annual growth through the end of the decade. That matters considerably for a Korean company looking to scale regionally before taking on the US and European markets. (GlobeNewswire)

Mental health investment attracted $2.7 billion in venture capital in 2024, a 38% rise year-on-year, making it one of the fastest-growing segments in digital health.

The money is not spread evenly.

It is concentrating in companies that can show clinical outcomes and a plausible path to reimbursement. (Galen Growth)

Beacon Biosignals is the clearest example of where that capital is going in the EEG space. The Boston company pairs FDA-cleared wearable EEG devices with AI analytics for drug development and clinical diagnostics. It raised an oversubscribed $86 million Series B in November 2025, backed by Google Ventures, General Catalyst, Takeda, and Catalio Capital Management, bringing total funding to over $121 million. Samsung Next joined a subsequent extension, pushing the figure above $132 million. (Medical Device Network) (Beacon)

Alto Neuroscience uses EEG biomarkers to match patients with psychiatric drugs rather than diagnose conditions. It raised a further $50 million in October 2025, led by Perceptive Advisors, with Eli Lilly, Novartis, and the Wellcome Trust among its backers over time. (Alto Neuroscience) (Alto Neuroscience Investors)

Thymia remains at the seed stage with $3.5 million raised, working toward medical device approval across the UK, US, and several other markets.

The pattern is two-tiered. Companies with pharma partnership angles and clear clinical trial applications are attracting serious late-stage capital. Earlier-stage diagnostic tools without a defined FDA pathway or reimbursement model are finding it considerably harder. Bwave sits firmly in the second camp for now.

The first question worth asking is whether this can work. 

The P300 response Bwave relies on is one of the more studied markers in psychiatry research. But no EEG-based biomarkers are currently used clinically for psychiatric disorders, and their validity still needs to be established through large-scale clinical validation. The gap between a promising research finding and something a clinician can act on is wide. (Journal of Yeungnam Medical Science)

There is also the false positive problem.

More diagnoses is not automatically better. When one health plan screened patients the traditional way, 3% showed signs of depression. Run through an AI screening tool, that figure jumped to 33%. (Prism News)  Some of those are genuine catches. Others are not. The downstream costs of acting on a wrong result are real.

Then there is the bias problem.

Once a company starts turning biological signals into psychiatric labels, the dataset matters enormously. Age, sex, medication, hearing ability, neurological history, culture, and familiarity with sound-based tasks could all shape the result. A professional musician and someone who struggles to distinguish tones may not experience the same “simple beep” test in the same way.

That does not make the patent useless. It means Bwave has to prove the system works across different people. Psychiatry has a long and ugly history of over-reading biology, appearance, and behaviour. Any new biological classifier has to avoid dressing old bias in new software. Digital biomarker researchers have already warned clinicians against using models before they are assessed for equitable prediction across diverse patients, behaviours, devices, and data types.(VIVO)

Then there is the FDA.

Kintsugi, a voice-based depression detection startup, spent seven years and $30 million navigating the De Novo clearance process, and still shut down when approval did not arrive in time. (Healthcare IT News) Bwave is earlier in that journey and operating from South Korea. But US or European market access means facing the same wall. (Kintsugi Health)

In the United States, brain data also sits in a different privacy category to most health data, and who owns a patient’s brainwave signature is still an unsettled question. State-level neural data laws are already emerging, and the proposed MIND Act would direct the FTC to study neural data governance and recommend a federal framework. (RMMagazine)

Psychiatry has spent decades searching for a biological signal it can trust. EEG has been in that conversation for a long time without delivering one that holds up consistently outside a research setting. Bwave’s bet is that combining brainwave data with source-level brain mapping gets you closer. The clinical evidence at scale will settle it.

The timing is worth noting though. Mental health systems are under pressure, capital is moving fast, and regulators are slowly working out what evidence they need to clear these tools. The window is real. Whether this small startup is the one to climb through it is a different question.

This week’s patent is US 12,629,071 B2, authored by Bwave Corporation.

Enjoyed this one? Forward it to someone in health tech or investing. And follow us on Instagram and LinkedIn for more through the week.