Vocal Biomarkers: Helping Clinicians Detect What Patients Hesitate to Share

April 2, 2026

By Henry O’Connell, Canary Speech

Five years into caring for a patient, a primary care physician saw her again shortly after she delivered her first child. The visit was scheduled as a routine six-week newborn evaluation. As is common in primary care, the encounter extended beyond the infant to include the mother’s well-being.

During the appointment, the new mother appeared upbeat. “I’m doing great — it’s such a joy to have my baby at home and to be through the process of pregnancy,” she said. On the surface, there were no red flags. A vocal biomarker analysis conducted during the visit told a different story: her depression levels were as high as the system could measure.

Despite her positive demeanor, the patient was experiencing postpartum depression. When asked by her doctor, she could not bring herself to admit it. While the baby was technically the only patient scheduled to be seen, the real-time insight generated by the mother’s vocal biomarkers created an opening for the physician to gently explore her mental health.

What followed was a pivotal conversation. The physician reassured the woman that she was not alone, that postpartum depression is common and treatable, and that seeking help did not reflect her failure as a mother. Initially embarrassed and concerned she was not living up to expectations, the patient began to acknowledge what she had been masking.

As real-time vocal biomarker analysis becomes available to more clinicians in more offices around the world, its potential to overcome previously unknown barriers — like patient embarrassment — is just beginning to reveal itself.

Postpartum depression is just one example.

Consider the elderly patient first beginning to show signs and symptoms of memory loss. A standard question like “how are things going?” might not prompt the introspection a provider needs to uncover the underlying problem. A patient who couldn’t find their car keys, forgot their spouse’s name, or got lost on the way home from the grocery store, might feel like they’re having a good day overall. Even those who recognize their forgetfulness as a possible symptom of something bigger might be too proud, or embarrassed, to tell their doctor.

Fortunately, vocal biomarkers can offer a clue to the presence of mild cognitive impairment (MCI), an Alzheimer’s precursor. Today’s technology allows unobtrusive, ambient listening technology to pick up on these hallmarks of MCI in real time. A referral to a specialist can then facilitate a formal diagnosis if warranted, giving the patient (and their family) fast access to resources that can help manage the condition as it progresses. Prescription medications, when taken appropriately, can delay cognitive decline.

Even the most experienced gerontologist might miss a potential Alzheimer’s or MCI case in their office. Consider the 65-year-old child of a 90-year-old patient who drives mom or dad to their appointment. It isn’t the physician’s job to probe the younger person for potential warning signs of their own memory loss. Through vocal biomarker analysis, however, a normal conversation can be used to detect MCI in not only the 90-year-old patient, but also their caregiving child simultaneously with a high degree of accuracy. For caregivers — whether they are supporting an aging parent or caring for a newborn — their own symptoms may feel like the least pressing concern, even when they warrant medical attention.

Vocal biomarker plugins for video appointments are gaining adoption as well, giving patients who meet virtually with their doctors access to the same degree of accurate, unobtrusive, early detection. Distance  was among the early obstacles that modern technology has helped to overcome as virtual appointments became more common. Patient embarrassment is a less obvious, but still prevalent, barrier for many who could benefit from earlier clinical intervention for their cognitive or behavioral condition. Thanks to vocal biomarker-based detection, it’s become a smaller obstacle than ever.

Across each of these encounters – the new mother insisting she is fine, the older adult minimizing lapses in memory, the caregiver too focused on someone else’s needs to acknowledge their own–the pattern is unmistakable. People do not always say what they are experiencing, and clinicians cannot treat what remains unspoken. Vocal biomarkers are revealing what demeanor, pride, stigma, or simple lack of awareness can conceal. They do not replace clinical judgment; they sharpen it, giving providers timely insight into conditions that too often surface only after they have progressed. Taken together, these stories point to a clear reality: the future of care will depend not only on what patients choose to share, but on what their voices can objectively reveal — enabling earlier intervention, more honest conversations, and better outcomes for the people who need care most.

Henry O’Connell is the CEO and co-founder of Canary Speech, the leading AI-powered health tech company that uses real-time vocal biomarkers to screen for mental health and neurological disorders. Henry has more than twenty years of experience in technology company leadership, both private and public. He has served on the board of directors for several technology companies in the U.S. and internationally.  Among his medical diagnostic and technology experience, Henry worked for Hewlett-Packard, Gibson and the National Institutes of Health in neurological research.