• PreSeed Now
  • Posts
  • The health secrets hidden in your voice

The health secrets hidden in your voice

Psyrin wants to catch severe mental illness early, using vocal biomarkers

Hello there,

One thing I love about writing PreSeed Now is how it gives me opportunities to learn about areas of tech that aren’t generating tons of hype but that are still packed with potential.

Today’s startup is a good example of that, coming as it does from the emerging field of vocal biomarker tech. Scroll down to read all about Psyrin.

We don’t usually don’t cover two startups in the same broad field of tech back-to-back; variety is our thing, but as chance would have it, this is our second healthtech startup in a row. Thursday’s startup is going to be a very different proposition though!

– Martin

PS: Don’t forget, you can keep up with updates about all 100+ startups we’ve already profiled in our Startup Tracker, a great benefit for our paying subscribers.

Psyrin uncovers mental health secrets in patients’ voices

What can your voice reveal about you? Maybe more than you think.

‘Severe mental illness’ is a technical term for conditions like schizophrenia and bipolar disorder. As Public Health England puts it, the term refers to “people with psychological problems that are often so debilitating that their ability to engage in functional and occupational activities is severely impaired.”

Spotting these conditions early is obviously important, so Psyrin is harnessing a new way of doing just that: by analysing a patient’s speech. The London-based startup is part of an emerging field in tech, using ‘vocal biomarkers’ for health assessment.

“We're building a new standard of clinical assessment for these serious mental illnesses,” is how co-founder and CEO Edwin Wong pitches his startup to me. “We do this by leveraging some proprietary A.I. algorithms to analyse just five minutes of patient speech, facilitating quick evaluation of mental health conditions.”

How can that possibly work? How could speech be a reliable indicator of a mental illness? The answer comes from the PhD work of co-founder and CTO Julianna Oláh

While she doesn’t want to give away the full ‘secret sauce’, she explains that the algorithms she has developed analyse linguistic parameters (the words a patient uses) and paralinguistic parameters (such as intonation, volume, and pitch) to identify specific conditions a patient may have.

The tech looks for “changes in articulation that can signal many problems in the brain,” Oláh explains.

Wong says that Psyrin analyses more than 150 different features of speech. Patients have a specific five-minute task where they are given a range of prompts, which are used to collect the necessary data. He says the approach can both identify current mental health conditions and predict future conditions before they emerge.

It’s this preemptive approach that Psyrin hopes will be transformative to mental health care.

“Clinical care for these illnesses is an inherently reactive system. Diagnosis happens too late and therefore treatment happens way too late,” explains Wong. 

“Obviously, the solution is a preventive clinical system with early detection and early diagnosis, but that process is extremely constrained right now because assessment, as of right now, requires extremely specialised training. It's highly subjective, and extremely time consuming,” 

Psyrin is specifically focused on identifying schizophrenia and bipolar disorder at present.

“Eventually, we want to look into severe depression, as well as other mental conditions that may be relevant to the clinical care system. But our initial focus is on psychotic disorders, given the cost and the burden that they have on clinical systems as of right now,” says Wong.

How it began

Oláh says she began the research that led to Psyrin because there were people in her life who suffered severe mental illness. This drove her to look into developing new methods of preventative care.

During her masters degree, she explored neuroimaging but decided something cheaper and more cost effective was required so she turned to speech analysis for her PhD. But she realised that if she stayed in academia, her work could take a long time to have an impact in the real world.

“We started doing a startup just to shorten the time frame from academic research into actual clinical implementation,” says Oláh.

And so she launched Psyrin with Wong and third co-founder Raheem Chaudhry, who has a clinical background. The three met last year and bonded over a passion for bringing this technology to market.

Next steps

Wong admits that there is a “steep path” to Psyrin’s technology being deployed in a clinical setting, thanks mainly to the certification required. 

He says to date the startup has trained and tested its algorithms in a proof-of-concept study.

The next step is the development of a risk assessment product for use in “highly specialised” research settings, which they plan to launch in around 12 months’ time.

Four years from now, Psyrin aims to have received medical device certification and have a product deployed in clinical care settings.

“We have a couple of NHS studies running in the background, but our primary initial market will probably be the United States, given the timelines involved,” explains Wong.

Go deeper on Psyrin

More information on their funding, vision, competition, and challenges ahead:


Subscribe to Premium Membership to read the rest.

Become a paying subscriber of Premium Membership to get access to this post and other subscriber-only content.

Already a paying subscriber? Sign In

A subscription gets you:
Full profiles of early-stage startups every Tuesday & Thursday with investment info, their future plans
Access to our Startup Tracker database
Support our work to bring you the best startups