Dr Adrian Owen is a renowned neuroscientist and university professor.
The reseacher is best known for his discovery that some patients thought to be in a persistent vegetative state are fully aware and able to communicate using functional magnetic resonance imaging (fMRI).
Dr Owen is now Chief Scientific Officer at Creyos – a company aiming to transform how clinicians assess and manage patient brain health with advanced cognitive testing.
Dr Owen discusses how Creyos is making it faster, easier and cheaper to diagnose ADHD and Alzheimer’s – freeing up clinicians and giving patients peace of mind.
Hi Dr Owen. Cognitive testing isn’t new. How has Creyos advanced the field?
What makes Creyos really unique is that it was really born of the science.
As an 18/19-year-old graduate student, I didn’t set out to form a company and to test people in a commercial sense.
I was just doing the science, and I did that for the first 20 years.
That meant that we built up a really large repository of data: Over 100,000 datasets from about 10 million tests being completed at the time that we spun out the company.
My PhD work involved patients with Alzheimer’s disease and Parkinson’s disease and so on.
Creyos CEO Mark Lipton was smart enough to see that we could use the data in a variety of different contexts.
What makes you different to your competitors?
Our competitors have published one or two papers arguing that [their technology is] the best thing since sliced bread.
But we genuinely have hundreds of peer reviewed papers using these tests in many different contexts.
Another thing is that many of these cognitive assessment systems have basically taken existing paper and pencil tests from the 50s, 60s and 70s, and computerised them to make them simple to administer.
But back when those tests were made, we knew very little about the brain.
So we tried to develop new tests that really assess how different brain regions and brain networks are actually working.
Dr Adrian Owen
The idea was to try to work out, how does the frontal lobe really work? And how does it contribute to our ability to attend to certain things? Which bits of the frontal lobe are most important?
Through that process, we’ve ended up with tests that really assess how well a particular part of the frontal lobe is functioning.
It’s not just a test of whether somebody can sort cards. It’s a test that really gets into the brain.
And from those data, you can deduce which bits of people’s brains are causing the problem.
It seems that cognitive testing has a whole host of applications.
It’s used in many different contexts.
For example, we’re currently using our platform to try to predict which patients with mild cognitive impairment (MCI) will go on to develop dementia.
This is a huge problem because there are cognitive tests that can tell the difference between somebody who’s got full blown Alzheimer’s disease and somebody who is just suffering from the effects of age.
But we’re not very good at predicting who in that group will go on to develop dementia.
You ideally don’t want to put people through unnecessary stress and extensive batteries of cognitive testing if there’s nothing wrong with them.
But similarly, you want to catch these things early, because it can help people with lifestyle planning.
And there are many potential therapies coming around the corner.
While we’re a long way off having a cure, we do know that all of these interventions work best the earlier that they are applied.
We’re running another study looking at the long term effects of brain injury patients who come to the ICU.
Nobody’s ever really looked at what survival really means for these patients.
So we’re using the Creyos platform to follow people up to five years after they leave the ICU to try to map what happens to them and identify whether there’s anything that can be done early on to try to mitigate any long term, cognitive effects.
Please tell me a little bit more about your new dementia protocol.
The idea was to develop a tool that could predict which patients presenting at a neurological clinic with memory and concentration difficulties would go on to develop dementia and who wouldn’t.
This is an interesting problem, because there is nothing out there that can do this.
We had several hundred thousand people in the database at that time, with 14 million tests completed.
We analysed the data to see which tests in the Creyos database would have been best at predicting which patients went on to develop dementia.
We could then take a group of age and gender-matched participants who had also taken the Creyos test but hadn’t presented in neurological clinic, and we trained our classifier to work out the best combination of tests that could tell the difference between the two.
It turned out there were two or three Creyos tests that could divide these two groups of people with 82 per cent accuracy.
We then took it to more datasets from different neurological clinics around the world which again showed about 82 per cent accuracy.
Finally, we tried it on diagosed Alzheimer’s patients in my lab. And of course, with 100 per cent, accuracy, it could tell who has Alzheimer’s disease.
Of course, that’s the least that you would want to do. There are other tools that can also do that. But we wanted to make sure that we were on the right track.
It’s the early cases that I’m most excited about: those patients who present in the real world as feeling like they have a problem.
We need to try to predict whether or not they will develop dementia later down the line.
We’re now working really closely with a group in North America to roll this out much more widely as a tool for predicting which patients with MCI will will go on to get dementia.
Another area of significant unmet needs which you seem to be tapping into is ADHD diagnosis. Can you tell me about that?
For many years, I’ve been saying to Creyos that you can’t diagnose a disorder based on cognition. But they’re gradually proving me wrong.
We have a huge database of several thousands kids who have been seen in ADHD clinics, and many of them had been formally diagnosed as having ADHD.
We looked at four markers of performance. And not just whether the person is good or bad at our tests.
We deconstructed all the tests and realised that if we, for example, take slowness on that test, with accuracy on this other test with, with performance on a third test, we can predict which patients with suspected ADHD actually go on to be diagnosed.
The process is a departure from the old way of doing things. Historically, people have been married to particular tasks.
But conditions like ADHD are really cognitively very complicated. And actually, there isn’t one single test for ADHD.
But it does seem as though there is an interesting pattern, that across various tests, does characterise that disorder.
That’s being rolled out now through North America, and it seems to be working.
The purpose is not necessarily to take the physician out of the equation. It’s to speed up the process and make it simpler.
That leads me on to my next question: What are the wider benefits of these tests?
There are long waiting lists everywhere and it’s also really expensive to get a psychological evaluation.
And then the evaluation itself is an intensive, three-hour neuropsychological assessment.
We can generate the exact same information in under 30 minutes without having an expensive neuropsychologist sitting beside you.
We really want people to be able to do this themselves, or even with a family doctor so it can be carried out really quickly.
Again, it’s not designed to replace the entire [referral and diagnosis] process. But it can identify who it’s worth sending [for an assessment].
If you wait three months and pay all that money to see a trained neuropsychologist for three hours, it could turn out that you’re fine.
Instead, [with Creyos] you could have reached that decision before you’d left the doctor’s surgery.
Everyone benefits.
So what’s coming up next? What are you’re excited about?
One thing the pandemic taught us is that our mental health is quite fragile.
Having widely available, economically feasible tools for measuring this opens up all sorts of opportunities.
There’s tremendous confusion when it comes to Covid, for example. Nobody really knows what brain fog is and everybody’s blaming it on Covid.
I think in many cases, that’s legitimate, but I’m sure in some cases it isn’t and there are other things going on for these people.
Can we tell the difference between who’s really got long Covid and who’s still suffering the consequences of that bash on the head during a rugby game?
It’s areas like these where we’ll be able to apply these tools more widely.
