AI and Eye Signals Revolutionize Diagnosis | with Paul Constable #157

You're listening to the Autism Weekly Podcast. Each week, we share community

voices and bring light to stories that increase awareness, acceptance, equity, access, and inclusion. If you haven't already, subscribe to join the Autism Weekly family. I'm your host, Jeff Skibitsky. This week, we're joined by Paul Constable, an optometrist and researcher. Paul's journey into autism research began with his son, Miles, who was diagnosed at age three. Faced with testing challenges, Paul explored retinal signals, inspired by Ed Ritvo's pioneering work.

His dedication led to groundbreaking studies and advanced signal analysis methods. Today, we'll delve into a research on utilizing eye signals and AI for quicker and more precise diagnoses of autism in children. Welcome to the podcast, Paul. Thank you. Thank you. So, Paul, one thing I'd love to be able to do is to give our listeners a little bit of a flavor of who you are and what brought your passion to the field.

Because a lot of people who get into research and get into practice around autism, there's a personal experience that led to this. And I'd love to hear your story. Sure. Yeah, it's quite a long story, but it all began with, certainly, I was living in London, just started the PhD, and I was looking at eye diseases in a completely different field.

And our first son was born, Miles, and he was a very sort of quiet boy, and we thought he didn't complain very much, but he had this fixation with opening and closing doors and switching lights on and off. And one day, the health visitor came around at the age of two and said, look, he doesn't really talk very much. and we were like, oh, he's a boy. Boys are always a bit slow with their language. What do you expect? And she said, no, no, no, he really doesn't talk very much.

And, of course, she sort of suggested we get a referral and send us off to some counselling to try and develop his language and so forth. And, you know, it was true. In a sense, he made very little eye contact. He never spoke our name. He referred to us as this one or that one.

And so the signs were always there. but we we didn't know how our parents were overseas we had no grandparents we had no sort of reference point for what's what's normal in a child and so at the age of you know two and a half three we finally got to see the the the the pediatrician and the psychologist and the the assessment and we walked into the assessment naive and went for a coffee and came back and I said, we're going to do some stuff with Miles.

And they came back and just said, oh, look, your son has autism. And it sort of hit us very much out of the blue. It wasn't on the radar. We didn't know what autism was, really. We didn't have any understanding of what the future might hold for Miles and what this all meant.

And so that then sort of began, I guess, the search for me personally into what is autism and how do we recognize it? and the question I had in my mind was, well, what happens next? Will he ever read? Will he ever sing? Will we ever play football together? You know, what's... Is this the end of the world, so to speak? Everything sort of came crashing

down. So I wandered into... My PSD supervisor's office in the next week, probably looking a bit glum, and said, oh, look, you know, Jeff, you know, my son that I've shown you the last, you know, photos of the last two or three years, he's apparently he's got this thing called autism. And Jeff was 83 at the time. He'd been working in vision science for centuries, it seemed. And he just turned to me and said, well, dear boy, maybe there's something to do with the brain.

We must look at the electroretinogram because you know he knew that people with sort of cortical or neural problems in the brain also often had sort of signs in the retina and this was perhaps you know patients with parkinson's disease they have changes in the retina and so this was sort of the link and so he then introduced me to a lady called dorothy thompson at great ormond street hospital in london and and we began in a sense trying

to to record ergs or electro retinograms which are the signals from the retina, in in people with with autism but that time back in 2005 to record one of these tests you needed to put an electrode on the eye you needed to put props in the eye to dilate the eyes this was all very uncomfortable so it took a while to get ethics and we could look at sort of adults with the own only at that at that stage and so things sort of stopped there because we had a very small study in 2016.

It took us to get everything going in about 12 adults, and we replicated some of the very early findings of a guy called Ed Ritvo. Who way back in 1988 did the very first study at UCLA into a population of children, I think in Utah and LA, and discovered some differences in the signals from the retina.

And that was the only study that had been done up until that point, and it had just been sort of gathering dust in the wilderness, I guess, because no one had been able to do these tests in children since then. So that was sort of the beginning of the studies into the ERG. It was really, you know, the questions about how do we improve the diagnostic process better for a parent, and just understanding, you know, we sort of left the consulting room days. Here's a leaflet.

Here's a phone number. Here's a phone number to go and call the National Autistic Society. And that sort of led to then, you know, the search for a school. What does it mean? Nursery school? Primary school? And, you know, we sort of, you know, you realise, I guess, then as a parent that you're on that different journey. You know, the other kids were growing up and getting toilet trained and learning to eat and going out and having parties.

And we were sitting at home with Miles. Suddenly the invitations dried up. There was no more invites to the school, to the French birthday parties, because Miles was just sort of, you know, his behavior was a little bit different. So I sort of thought that that then led to sort of, I guess, a sort of a separate area of interest in sort of how do we help these children. And as an optometrist, I was sort of thinking, well, and you might recognize

I've got a big prescription. I wear glasses and I think any parent who wears glasses probably thinks, I must get my son's eyes tested and maybe they need glasses as well. And I'm the optometrist in the family, I should be able to test my son's eyes.

And but it was just it was just a very fearful experience just because i couldn't consider taking my son to somebody who would be able to communicate with him or understand you know what his needs would be in the consulting room and i and and around this time we were looking at at schools and there was a a special school i guess you'd say in london called queensville and we thought mars that was an autism dedicated school

we thought mark would fit in there and they came came to look at Myles and said, well, he's not bad enough for us. And we thought, oh, okay, that's probably good in a way. So he sort of ended up in sort of mainstream school, which was good with support. But when I visited this school in Queensland or in London, it was sort of the children there would have had two-to-one carers and were very non-verbal.

Myles was quite non-verbal. I think he started to speak when he was about the age of eight and got out of nappies about then as well and started to develop some skills then. So he was always about five or six years behind the pack, but not as bad, perhaps, as these children that were in this school full time.

And so I began doing eye tests on these children and then started developing strategies for testing children with poor communication because that was sort of – they care about – because I finally took Miles in for an eye test at the age of eight or nine or so to the optometry clinics at the university where I sort of worked and thought. I had some colleagues there who'd be able to manage the eye test quite well, but they were terrible. It was just a complete disaster.

They were just a complete disaster. You know, Marth was wanting to come down in the chair and sit still, put the glasses on. They wanted to put the drops in his eyes and put the drops in his eyes. And I was like, don't put the drops in his eyes, don't put the drops in his eyes, no. And so of course, he's got the drops in his eyes he can't see. And we're on the 19 bus going home, and of course, he's full of tears because he can't.

Where his vision's gone, he's sort of flying out. He likes to count the numbers

down on the bus as we go home and he knows and he can't see the numbers and he's just a mess. And I realised then, as a parent, it's like, this has got to change. So I started working and sort of trying to sort of educate optometrists as well about how to approach a child with ASD in the consulting room and did some guidelines and stuff like that. So that was, I guess, the early years.

And, you know, there was a lot of support I had, I think, from a guy called deva bowler who was at the university where i was working he had an autism research lab and i went to see dermot and said look our son's called autism what does it mean and he was working in adults and he was um really a bit of a godsend because he said look paul.

Things get better you know they say every phase of life is hard you know it's going to be hard in their fight in their preschool years and then they'll get to teenage years and they'll get to adult and they're all going to bring different challenges and you just need to um have hope but work on language communication which we did and and it gave a bit of a ray of sunshine there because at the time we were pretty down

um well you went through i mean it just kind of go back a step i mean with miles miles went through the typical diagnostic process it sounds like and what you're looking at is groundbreaking in the fact that you're hoping to bring different information, more information to the table.

So, what is it that maybe a family is going to learn, or a clinician may learn by utilizing eye signals or AI through the diagnostic process that's different than the psychological diagnostic assessment or neuropsych assessment?

Yeah well i think the big thing the big question i had as a parent was like well how bad is it and what's the future so it's just prognostic it's a it's a biological marker like it's a physiological thing it's not oh his language is on a two out of five scale or he's you know his social reciprocation score is this score or that score it's very it's very abstract and.

Although you know the ados is very good and it's a standardized test and it does give you know normalized data but as a parent maybe it's my scientific background but i wanted a number i wanted i wanted to know i wanted to know the cause or the reason you know something like perhaps diabetes or hypertension or something which you can measure and sort of monitor and and and so that was sort of i think the the you know the attraction perhaps of an objective.

Test that taps into the retina which is connected to the brain which we know in other conditions which affects the brain has different signaling properties. And we thought that those different properties would help us to categorize better or give some sort of indication as to what.

Chemical maybe imbalanced in in in this particular individual and give some sort of i guess i guess i guess that's sort of you know the question you have is what causes autism you know why why does my son have or daughter have autism what's the what's the biology behind this this condition and of course the genetics is all out there the there's no answer to what causes autism and i think that was sort of part of the question for me is like well what's you know

what's the the main driver for mars's behavior is it too much of this or not enough of this or which pathway is affected in a sense and that and that's i think what where the erg may help to give some assurance to parents as well it's like it's not their fault i think that's sort of that the question we all have is like you know why me why us what have we what did we do wrong you know what's what's what's happened whereas if there's a perhaps oh it's

because of you know there's a little bit too much of this or a little bit less of that then it perhaps makes that makes it makes that the condition more understandable for a parent as like it's not it's not you it's causing this it's their biology in a sense or their physiology and that and that makes a lot of sense i mean that just looking at what you're describing there it's that it's giving more of a medical understanding to you know this is this is the causality are

you also potentially looking at or it's their future research that's looking at, you know, maybe there's some treatment ramifications that are coming.

Or maybe a clinician would look at putting more effort into these sorts of skill building or into this sort of medication path because they're seeing this through the ERG.

Yeah, that was really the goal, that this would help with the management and diagnostic, diagnostic you know sort of categorization perhaps you know there are looking at the genetic studies now there's lots of different sort of there seem to be lots of different flavors of autism lots of you know based on pathways and immune pathways and you know which i get lost in as well and and maybe this sort of you know the erg can help sort of at least help to to

clarify some of those those categorizations so look this group of children are really going to benefit from this medication or this group of children don't even touch medication that's not going to help you all these children are going to get a bit better because it's actually the retinal look the retinal signals are pretty normal and there's a good prognosis or there's a retinal signals are really bad and you know maybe we need to put more effort into into the communication really

so you know give you more support those those sort of things now and that and that that makes i mean just it gives hope to being able to say you know we can we can focus our care better and that we can and hopefully expedite our care models to be able to be more efficient in what we're trying to be able to do to empower the kids or the adults. I think it's also, I think, I guess it also gives that sort of.

Market that you can measure like blood pressure because is are things getting better by doing this you know are things changing and uh in this in the balance of the the neurotransmitters in the brain perhaps are they are they having some effects we can detect in the retina quite easily without doing a brain scan or an eug or you know those those sort of things so that's perhaps the the attraction the party attractions to doing this sort of study so if we start at the very

and beginning and kind of work our way forward and look at the diagnostic process alone. Is this something that you're looking at diagnostics could potentially start earlier? Or, I mean, is this something that it still would be, we're looking at 18 months as the starting point? And what's the start? Where do we start with it? Yeah, that's the idea.

I mean, I guess, you know, I dream one day, how many children do you have the neonatal screening where they do the hearing test and it'd be great to do the eye test as well the eye screen as well and say oh look you know the hearing is good but the retina is also good and or whatever and so that that that would be the idea and in theory it's possible we haven't got to those younger children yet that's for future studies but in

theory you you can record an erg in very young children when they're sleeping or when they're breastfeeding or so forth and that's where i think you know once they get to the age of two three four five they're a little bit harder harder to test.

So it's sort of looking at that sort of at the neonatal age would be the, you know, the pathway and bringing it in as part of that neuro screen, I guess, for all children, along just along with a hearing screen, you do an eye screen, you know, you do an ERG, you do an auditory test as well. And I think that would be the, you know, that's what we're hoping to get to.

And I mean, you had mentioned that there were ethical concerns back in the 80s about this, but I mean, to calm people's fears, I mean, a lot of those things have been solved.

It's now a lot easier to do the test. I mean, I guess we have to wait for the technological advances. A company called LKC Technologies, based in the US, developed a handheld device called the Redaval, which uses a sticker electrode on the cheek, which means that, and you don't need to use eye drops, so it measures the pupil diameter, so we don't need to dilate the pupil.

So no drops, nothing on the eye, portable, do it in your office, do it in any way you like, you know, not a very invasive, invasive. And that really enabled us then to do the further studies with the children. And it was really Ed's help in the end who sort of really helped because I sort of, I'd met Ed at IMFAR once in Boston, although I don't think he recognised me. But I was sort of like in awe of Ed Whitburn. And,

And we published the adult study. And I thought, oh, look, you know, let me send this off to, you know, I couldn't find Ed's email address because he'd retired. And so I think I sent it to an Ari Ritbo, which I assume might have been his daughter, but I think it turned out to be his ex-wife or something anyway. But she was at Yale. And I sent this paper off to Ari and said, oh, you know, you might be related to Mr. Ritbo, you know, you might be interested in this study that we just sort of

replicated. And I heard nothing.

And at that stage, we moved back to Australia. Australia because I sort of had to jog here in Australia left left the UK and I was sitting in my desk one day and I got this phone call and was like hey Paul it's Ritvo here how are you, it's like oh Ed how are you he said yeah man I got your paper really cool stuff like I've got some money at UCLA let's do a study and you know Ed's 88 at this stage but I'm like yeah Ed let's do the study so it sort of gave him a lot of um he really

drove the next phase of the study and I I think it was part of his – it was nice for him to see some of the culmination of his early work sort of being replicated and repeated and expanded. And that was always his dream. He was quite a visionary guy to sort of say, you know, we want to use this ERG to help diagnose and to understand autism more. And so from 1988 to 2016, there was just this dead space. If you look at PubMed, there's no publications in the field.

And then we started to get stuff going. And so I sort of feel a little bit carrying on the baton from Ed's pioneering work to try and sort of answer those questions. That sort of drives some of the research as well. So yeah so so moving on now we've really expanded the. The potential of the of the of.

The electroretinogram and so i guess to put simply what we've been doing i guess if you think of the electroretinogram a bit like the electrocardiogram it's like it measures the electrical signals of the of the eye and the heart so i guess it's the heartbeat of the of the eye if like you could say and and you know we have when we look at the erg on a very simple level you could say well the you categorize ergs based on like you think of like like a song is it

sort of r b or is it classical or is it jazz or whatever and they're very sort of broad broad groups that you can put put things into and and when we just do those sort of very broad categorizations looking at the peaks and the troughs of the of the signal we can't we can sort of classify the groups quite well but it's when you go into sort of the the analysis of the of the signal you know what are the what are the frequencies what's the

high frequency what's the low frequency what's the beat what's the rhythm what's the tempo and that's really what we're doing the signal analysis we're sort of really pulling apart the song if you like the the signal from the retina and and and looking at the what the structure if you like and and that's where the the the excitement i guess is coming from now we can start to really sort of untangle very small changes in the signal in these children with asd or children with adhd

and then you know the third group the.

The when we were doing the study with with ed we we created children with asd and of course but some of the children would turn up and they'd have asd oh and they've also got a bit of adhd as well the parent would say oh well don't worry about that you know we'll i'm sure it won't have any effect and i and i remember it you know at the time when i was diagnosed and oh he's got asd but there's and there's a little bit of adhd and they're like well what's this adhd

stuff i don't know great he's got a bit of adhd as well anyway so but we started to notice that these children with this little bit of adhd this you know other thing they'll be having quite differently to the children just with the ASD diagnosis. And so we thought, oh, well, maybe there's something in the ADHD group as well. We hadn't really thought about that. So now we looked a little bit at children just with ADHD and then COVID came and we sort of stopped.

But we found that the children with ADHD had a very altered signal to the children with ASD. And the children with ASD plus ADHD, they were somewhere in between. So it was sort of this sort of spectrum of signals that we were finding in these three populations. And so the question so what we're looking at now is children with ASD plus that little bit of ADHD and children with the ADHD diagnosis to try and sort of see if we can sort of you know.

Pass out these these three groups and that's where the ai comes in you know to help us with the.

We're deciding which signal characteristic belongs to which group and so sort of combining that sort of the ai plus that deeper signal analysis looking at the tempo and the frequencies and stuff that we can do in the waveforms building that into a computer program yeah into a testing tool for clinician they can say to the test ping here's the the likelihood of the probability of your child being in this group and house and how

severe going back that severity you know where do they fall in this in this full update that we have you know what does it mean to the parent and how then what might be then be the best way to manage that and so where we're sort of pushing again with the university candidate with with um hugo's group there with bioengineering we're sort of pushing now to develop a new device if you like which will run off a smartphone which do all the calculations so that the

smartphone will run the device which is sort of like a pair of glasses which will measure the signal and then the question or the parent or the psychologist or the school teacher or the nurse or kindergarten teacher anybody who can sort of you know do this sort of test and the parent perhaps could do that at home and sort of monitor things as well So that's, you know, that's, yeah.

I mean, the value of this just sounds astounding. I mean, just to have that much information and to be able to continuously build upon it, because it sounds like you, once you start storing this data. Is that now you have computer learning. You can start answering more and more questions through it because it sounds like the eye is capturing a lot more information than what we're able to extract right now. Is it a matter of we need more participants to be able to kind of build a larger database?

Is it a matter of we need somebody to take this sort of technology to market, to be able to bring it to clinicians, to hospitals, to – where's the block? Yeah, well, they're the next steps. I don't think they're blocks, but we'd love to have someone to come along and help. We're building a prototype. We'll have a prototype to test in sort of middle of the year or so.

And so then it's okay. Okay, so then, you know, getting, yeah, looking at industry partners and trying to sort of, you know, develop it and get a patent and produce it and then get a large clinical, multi-centre clinical trial. Because that's always the thing is like, well, populations are all different, the diagnostic standards, although they're similar or very standardised, there's perhaps some variation.

Variation i guess the other important and new thing that i've been working on well with a group in in germany at siemens ai you know they'll be in the media about how ai is destroying the world or whatever but i think i see i see this really sort of positive and useful thing and and some colleagues in germany they've been sort of generating fake erg signals so based on the population signals the ergs that we have they can use ai to

synthetically i guess it's deep fake making making that just started sort of showing me some of the data and i can't tell the difference between a fake erg from an autistic kid from a real one so you know we can generate.

More data to use in analysis based on real data and so there's what there's a potential to sort of to to to push ai even further and to increase our sample through the generation of synthetic or artificial signals as well but it's obviously nothing but but but not nothing beats a real clinical population but it's sort of but it helps the machine learning it helps the computer understand it's like well we can you have two thousand signals we can give you two and a half salesman.

And when you're talking to a parent, because you went through this process, I mean, you didn't have any of this opportunity when you went through it with Miles, but talking to a parent, talking to a clinician, talking to researchers, I would imagine it's different stories. And I'd like to hear kind of where the benefit is for each of these populations so that they can be on board with understanding how they can contribute.

So when you're talking to parents who are trying to figure out you know why is it that this would be where i would want to put my time energy and resources what what is the story that that they need to hear to be able to understand you know this is this is worth my my efforts yeah i think it's the.

It's the it's the question of you know why why and what's going to happen you know i said look look at miles now i look back to what he was like in his early years and now he's he's on his xbox and he talks a bit he's still very at home but you know the the one thing i remember from from a woman called gillian baird because being a a doubting thomas parent i got a second opinion from gillian baird he was sort of an expert in in autism and she and she.

Great advice from him was don't ever underestimate him and I think that's probably sort of what I took home and even though things are bad now you know you don't, underestimate their potential or his potential or what what's going to happen what they can become and i think you know the erg is going to help perhaps to give some certainty to that there is potential here i think there is a cause and an effect and and we can monitor and we can program and we can we can see how this will change

in the future and i think for parents the other important thing i think is for parents there's so many parents i saw obviously have come in with one child and then you know they get diagnosed at three or four and then have another child and then the second child has autism and it's not too late but it's like so it would be it would be it would be nice i think to have a parent if they have a newborn child or younger child the second one to have you know you have one with autism or is

a second one kind of clear or not clear or whatever and so that that would sort of have perhaps help so interventions can be brought earlier to perhaps you know the the second or third child in in the family as well so that's that's sort of another aspect to to the testing not only for the the child that might be affected Absolutely.

I think for a lot of parents, it's just understanding creates compassion, intention. It creates some of that idea of this is just part of a journey, but I now understand my journey. Yeah. How about for the clinician? I mean, you look at precision medicine right now, really taking a path.

And when you were talking about ADHD versus autism and maybe them creating a differentiation through some of this understanding is that it probably has something to do with some of the psychopharmaceutical direction that you might have with some of this.

Is there is there benefits that maybe researcher physicians would be looking at saying you know yeah i should probably be supporting or investing more into something like this because this is going to lead to some really wonderful results in the future for me yeah well certainly certainly my help with help with drug targeting i guess because we can sort of the retina is a very interesting because you can think of the retina as a very simplified model of the brain

it has neurons it has the same chemical signals and if we and from the the pattern of the erg we can sort of tell which which which chemicals are aren't working or imbalanced if you like and so that gives a target for the for the medication and also perhaps for the for the dosing as well so you know say well how much drugs should i be using well okay do the erg is the signal normalizing stabilizing is do you need more do you need less though those sort

of things would sort of help i I think they work with the management, I guess, that therapeutic management and for parents to say, well, you know, am I doing too much? Am I not doing enough? You know, what do I... You know, medication's always sort of the question, I guess, parents need to ask. It's like, you know, is this good for my child or what are the long-term effects? And, you know, those sort of questions. And I think just having the...

You know, the ERG in your pocket, I call it, you know, the smartphone in your pocket that you can just pull out and show your parents and say, look, here's a normal signal. Here's your child. They're clearly different. There is a problem. And I think that perhaps, again, for parents who may be confused or doubt the diagnosis, and I think perhaps, well, necessarily in our case, but I can imagine, you know, perhaps parents, you know, blaming the other one perhaps, but it's actually no one's fault.

It's that sort of it's fault free. It's like it is just the way it is, you know, But this is a biology. It's not anyone's or anything you've done that's caused this outcome. You know, you just got to, you know, you're on a different journey. And that makes, I mean, it makes it so that you can almost take that next step and start kind of that continuity onto the journey, as you described. A lot easier versus being stuck in that, where do I go?

And almost going in this circular path of, I'm going to hover here until I have certainty. And that leaves me not really understanding what's going to help make my life better or help empower my child to feel more comfortable or to be able to engage in more of what they're hoping to be able to accomplish in their lives. And I think that that's important.

I think that acceptance is the important step, you know, that's sort of like, because we didn't understand why, you know, what does it mean? Again, what does it mean? What's going to happen? What's causing this difference? Why us? But I think having that sort of, and so for us, having a biological marker, something that you could sort of, you know, look at and go, okay, I see there's a difference. I understand now why things are the way they are, a little bit.

I don't like it necessarily, but I understand it. Absolutely. So where can people follow along with what's coming out in this technology? What resources are available? How I can learn more and dive deeper into what this actually means? Because we're on a 30-minute podcast. We're not going to everything that could be describing all the benefits. Where can people turn? Yeah, well, we're just setting up a website now called retinobiomarkers.com, and that should be live in the next day or two.

And so people could subscribe to the website and leave their details, and we'll sort of send updates to anybody that's interested in following our work. And if we get research projects going in the US and Australia, the UK, Europe, they're our sort of main sites. sites, then, you know, we'd love to have everybody involved with developing things further and helping the lives of not only the children, but also the parents.

I think parents, families, the whole, and it's not just the parents, it's the grandparents, it's the cousins, it's the whole, everyone's involved once the child has the diagnosis to support that child. So we want to get everybody involved if we can. Well, thank you so much for giving us some of your time today,

day, Paul, and for the work and passion that you have, because you took something and basically stopped in the research field. Like you said, there's a black hole, and you rejuvenated it. Now it has this momentum where it's going to empower so many people in the future. And I encourage folks to read more about it, invest some of their time and resources to really understand and hopefully be able to build upon the technology that's out there. Thank you so much. I'm glad to share it.

Thank you for listening to Autism Weekly. We hope you tune back in next week to learn more about autism in the real world. Autism Weekly is now found on all the major listening apps, including Apple Podcasts, Google Podcasts, Stitcher, Spotify, Amazon Music, and more. Subscribe to be notified when we post a new podcast. Autism Weekly is produced by ABS Kids. ABS Kids is proud to provide diagnostic assessments and ABA therapy to children with developmental delays like Autism Spectrum Disorder.

You can learn more about ABS Kids and the Autism Weekly Podcast by visiting abskids.com. Thanks for tuning in. See you again next week.