Imaging the ADHD Brain. Webinar

– [Facilitator] Welcome
to theDeakin Alumniand School of Psychology Webinar
with Dr.Tim Silk. Dr. Tim Silk is a cognitive
neuroscientist specializing in pediatric neurodevelopmental imaging in order to
understand the brain-behavior interface, when and where
that goes awry. Specifically, he is focused on identifying
neuroimaging markers that can be used to distinguish children with
neurodevelopmental disorders, monitor progression, and predict likely
outcome or treatment response. Dr. Silk studied his Bachelor
of Behavioral Neuroscience and Ph.D. at Monash University and
the Howard Florey Institute. He then completed two years of
post-doctoral research at the Queensland Brain Institute in
Brisbane before returning to Melbourne to commence at the
Murdoch Children’s Research Institute where he was
for nine years. Dr. Silk joined Deakin in April 2018,
where he heads the Brain and Cognitive Development Lab
in the School of Psychology. Thanks so much for joining us today, Tim. I’ll now pass it over to you to begin your
presentation. – [Tim] Okay. Well, thank you very much for the
invitation. So, yeah. In the Brain and Cognitive Development
Lab, we’re very much about using neuroimaging techniques to try and
understand how the typical brain develops and then also when and
where that can go wrong. And today’s presentation is going to be
talking about a large cohort of children with ADHD that
we’ve been following. And I’ll really be describing the cohort,
and then I will give you one of…present some of the findings from one of the key
baseline papers from this cohort. So, when we talk about neuroimaging,
there are a number of neuroimaging techniques, but I’m principally
going to be talking about MRI today or magnetic
resonance imaging. But just to start off with, I’ll give you
a quick introduction to ADHD. I mean, everyone has sort of a kind of an
idea in their mind of what ADHD is, but sometimes that can be a little bit
misrepresented. So, I want to try and
make that quite clear. So, ADHD is
characterized by sort of inappropriate levels of inattention
and/or impulsivity and hyperactivity. So, not all children will be hyperactive,
and not all would be inattentive. There’s a mixture of these two sort of
symptom domains. And there are
different subtypes or presentations that the
children could present with. So, some may not have hyperactivity at all
and purely have problems in maintaining attention,
and the reverse can also be true. It affects around 5% of
school-age children and that drops to about sort of
2% or 3% in adults. And despite what some of the media says
that that percentage is sort of worldwide as well as being quite consistent for the
last 30 years. There is somewhat of a sex ratio so that
it’s more commonly diagnosed in males and that’s, it can be up to about five
times more common in a clinical sample and maybe two to three in a
community-based sample. And the reason those are different is that
often those with inattentive symptoms, without that hyperactivity, don’t get
picked up quite as much as well as females don’t get picked up
quite as much. And so, they get captured in a community
sample where whereas they’re not getting referred
to the clinics. Above and beyond the symptoms that an
individual needs to meet criteria for that diagnosis, ADHD can really have a
big functional impact on many areas of functioning
for these individuals. This can be academic, in their academic
achievements, in cognitive domains. There’s kind of some common areas where
there’s cognitive deficits, but it’s not necessarily the same for all
individuals. Also, it has implications for their social
functioning with their peers as well as broader mental health functioning and
different comorbidities. And a majority
of these will have impairments that
do persist into adulthood. Even if they do officially remit out of
their ADHD diagnosis, many of those functional impairments will
still persist. This is just a nice
figure from a key paper, which just sort of highlights these key
sort of long-term outcomes. So, there are a lot of sort of risk-taking
type behaviors that can lead to sort of premature mortality or other, you know,
high frequency of car, motor accident, vehicle accidents, and unplanned
pregnancies. In terms of their academic achievement
individuals are more likely to fail grades at school, or be expelled,
or drop out. And this also then
goes through into adulthood. So, a few’ll go on to tertiary education
or have more problems with
finding employment. In terms of the biological basis for ADHD,
there are a number of studies that have found clear
environmental effects. And this table just highlights a review of
some of the studies that have found some association,
but there is some clear, robust evidence of genetic
susceptibility to ADHD. This is through a range of sort of family
studies looking at the genetics of parents as well, sibling studies where there’s a
higher rate of ADHD among siblings as well as
within twin studies. So, a higher rate within monozygotic
identical twins. And this overall heritability is really
quite high. It’s up around 7 and 7.75,
which if you consider something like height
is around 0.9. So, even though there’s this strong sort
of heritability, there’s been no single gene that’s been identified that that has
a strong causative role, but rather many, many genes that have been identified that
might contribute a small risk
towards ADHD. Kind of a hallmark of ADHD is its
heterogeneity, its variability in the way
that it presents. This is both in its clinical presentation. So, I’ve already mentioned different
subdomains of symptoms like the inattention or the hyperactivity,
but also the pattern of co-morbidities. So, different mental health problems that
they might have, I’ve already mentioned. So they can have different profiles
where there’s cognitive deficits and even whether and how they
respond to treatment. Now, this is not unique necessarily to
ADHD, but across many behaviorally-defined psychiatric disorders,
there has been this sort of shift to, at least in a research realm,
to sort of shift away from looking at these two supposedly homogenous groups,
a clinical and a healthy control. So, rather looking at how some of these
traits might vary across the population. So, what can your imaging tell us about
mental health? I’ll just sidestep briefly just to
introduce what MRA is for those who don’t know, just so we’ve got a clear
picture. Magnetic resonance imaging is basically
uses a large magnet and the idea that water in our body and particularly in the
brain, it constitutes the greatest sort of mass
within the brain. So, nearly up to 80%,
but that concentration is obviously different in different tissue types such
as the cerebral spinal fluid versus the brain tissue
versus the bone. Now water molecules are kind of like a
little magnet. They’re polar, they have a
negative and positive side. So, when we put someone in a large
magnetic field, these water molecules line up with that
magnetic field. We can knock them out of that alignment
with some radio frequencies. And as they return back to that alignment,
we can sort of have measurements out of there that allow us to look at the
concentration within those different tissue types which give us
the images that we see. Now, there are lots of different types of
images or photos if you like, that we can actually acquire from magnetic
resonance imaging. So, the most common will be the sort of
structural images. And the most common we can see there on
the top layers, the T1 and a T2, but they have different contrasts
based on, you know, or are used to look at different pathologies, depending on
what you’re trying to look at. And there are a range of others. And these sort of images can be used to be
able to look at the cortical thickness, like gray matter that
coats around the brain. And we can divide that into different
particular regions or look at the subcortical type, you know,
segment [inaudible] structures out and you
look at the volumes. We can also have a look at function. So, we can give individuals a task to do
in the scanner and see what parts of their brain they’re using to do that task as
well as what’s been used a lot more recently is what’s
known as a resting state. So, even if that individual is not doing a
task, looking at a sort of fluctuations of, of blood flow and what
parts of the brain are kind of functionally
connected. We can also use the properties of how
water diffuses to infer structure about the white matter,
which is basically the telephone cables that connect
parts of the brain together. And we can use this to sort of look at the
whole brain tracks or particular or segment particular
anatomical tracks of interest. And then we can also look at things like
sort of the [inaudible] or chemical composition
of different regions. And there are many more as well, but
they’re sort of the most commonly used. So, what can it tell us
about mental health? Well, if you look at the top row of brains
there, these are neurological conditions. And I don’t think you need too much
neuroanatomy to be able to see that there’s something severely wrong with
those brains. However, when you look at the brains down
the bottom, and these are sort of behaviorally defined psychiatric
disorders, there’s nothing that you can pinpoint looking at a particular brain
that would assist in a diagnosis or be able to identify that individual having
that disorder. But what we can do is have a look at the
brain-behavior relationship. So, this is a figure I like to show which
kind of represents where I think neuroimaging fits in to trying
understanding from sort of that molecular level through to the behavioral symptom
level that we see at the end. So, at the top there we have ADHD,
but this could be any disorder. And those disorders are a collection of
symptoms. Those symptoms arise based on sort of
deficits within
particular brain processes. Take inhibition, for example,
those basic brain processes, you know, are attributed to particular brain
structures or networks of structures that can have their own different sort of
neural mechanisms. And in this example, I’ve got some
neurotransmitters, but there
are other mechanisms. And those, you know, mechanisms are
driven by different genes as well. And then, of course,
environmental factors can influence sort of a range of levels
along that. So, neuroimaging sort of allows us to look
at the marker that’s sort of halfway in between what’s happening at the
molecular level and the symptom outcomes. So, I mean, ultimately the goal would be
to try and use this to aid clinical diagnosis and help
reduce misdiagnosis. It’s not quite there at the moment,
but this is sort of ultimately the goal, or potentially, you know,
parcel out different clinical subtypes or even identify precursors
before onset happens. Yeah. At a research level,
it’s kind of on a group level that’s being used to try and predict whether
individuals will respond to treatment. But that’s not something that’s being used
clinically at the moment. But what we can do
is use it to understand the neurobiology
underlying these disorders. So, is it a
particularly distinct entity or whether it’s just a dimension
across a normality? Can we identify particular structures that
are affected or networks of regions? Is it lying on the gray matter? Is that the white matter? Is it neurochemical? And what are
the sort of…what contributes to the functional deficits
that they’re seeing? So, now I’ll introduce you to the
Children’s Attention Project, which is the community-based cohort,
longitudinal cohort that we’ve been following
of these children. Essentially started before it was an
imaging study, and we screened over 6,000 children across
43 different Melbourne schools. And the aim of this study was to really
assess the long-term functional outcome of
children with ADHD. So, not just whether they remit it or not,
but looking at their mental health outcomes, the broader
co-morbidities, the trajectories of academic, social,
and cognitive function. The really important thing about this
study is that it’s multi informant too. We have information from the child as well
as a direct assessment with the child, information from the parents as well as
their teachers. And being a community-based cohort,
we actually have quite a representation of the different ADHD subtypes as well as
both males and females, which can be both the sexes and the
subtypes can be missing from a lot of detailed clinical studies that try and
maintain a really tight homogenous group. We also get measures of the parent’s
mental health and the family functioning. And really the idea of this study was to
try and see if we can identify a particular risk or even protective factors
that are associated with an individual having a good
or a poor outcome. So, I’ll just go through some of the
details. So, this was the original screening of
children. They had to screen positive by both the
parent and the teacher. And if they screened positive by both and
they consented to the longitudinal study, they had a full diagnostic interview,
and we ended up with a group of 179 children with ADHD and 211 controls
which were matched for gender and school. Quite of note though,
I think it’s worth…these children were around the age of
six or seven when they were recruited in the first
year of school in grade one. Only 17% of those in the ADHD group had
actually had a prior diagnosis, clinical
diagnosis of ADHD. And then along the way,
aligning with the third wave of data collection,
we got another NHMRC grant to do MRIs on a subset
of these children. And so, that we’ve now collected three
additional waves of data collection where we’re scanning
these children in each wave. These waves are 18 months apart. So, the scanning had
started from around the age of 9 to 11 through to sort of 12
to 14 years of age. Those stars on the side indicate the
points at which a full diagnostic assessment
was run so that those diagnostic assessments were repeated
every three years. Now we’ve collected so much,
very detailed, phenotypic data and I’m not going to list all of the measures we’ve
collected, but just sort of give a brief overview of the sort of domains that we’ve
been collecting. So, from the parent surveys,
we’re getting information about the child’s behavior, quality of life,
their social skills and autistic symptoms because there’s a lot of overlap in
symptom domains with autism and ADHD, treatment history as well as
sociodemographic factors. From the parents, we also get information
about the family quality of life and the parents’ mental health, family stress,
different parenting styles. Where possible from the teachers,
we try and get similar measures. So, once again, items on the child’s
behavior and their social skills and
autistic symptoms. But we also got some unique measures like
the parent-teacher relationship as well as whether the child’s using any school
services. Then with the child,
we get measures of their visio-perceptual and motor coordination,
verbal and nonverbal cognitive functioning, working memory,
language, academic functioning as well as response inhibition
and attention. And we also get some physical assessments
like height and weight. And at the later stages,
assessment of puberty. And we’ve also got some data linkage to
NAPLAN results, which is the standardized
academic testing. So, onto the Neuroimaging of the
Children’s Attention Project, or NICAP. Following this really
nicely phenotyped cohort, we started collecting
multimodal imaging. So, different types of sequences of MRI on
a subset of these children, 180, at three time points
with 18-month intervals. The children would come into the Royal
Children’s Hospital for roughly a three and a half hour assessment,
where we did a cognitive assessment, of self-reports
from the child. Parent was there doing questionnaires,
teachers were sent their questionnaires. Children were trained
or have a practice in a mock scan,
which I’ll talk about later. And then had an MRI scan where they’re in
the scanner for roughly 45 minutes. The aim of this main study was to look at
the trajectories of how their brain structure and function change across late
childhood into adolescence. But more importantly,
I think the same too is how these different trajectories in their brain
reflect different outcomes, not just the persistence or remission of
ADHD, but the difference in functional outcomes across those academic,
social and cognitive domains. This image just highlights some of the key
multimodal images that we’re acquiring. So, we’re getting
different images that look at structure A and B there,
a T1 and a T2. See there is a QSM, is quantitative
susceptibility mapping, and this is quite a
novel imaging sequence which looks at iron content
within the brain. The reason we’re looking at this is that
iron is highly concentrated within some of the basal ganglia structures deep
within the brain that are really key for cognitive
and motor function. And it is also related to or in quite
involved with dopamine system, dopamine being sort of a driving
hypothesis in ADHD and the main neurotransmitter system
that medication impacts. Then we also have a range of diffusion
values which look that white matter tracks and they’re optimized for different
measures we’re trying to pull out. And then we also have a resting site
functional. So, looking at when that brain is at rest,
what parts of the brain are
communicating together. This figure just shows you the overall. We’ve just finished earlier this year our
final wave of data collection. So, in total, we have 471 scans,
which is for an imaging study, this is really
quite a large cohort. This figure shows along the y-axis each
individual participant with each dot indicating the scan with all
connected by a line for each individual. So, you can see, there is some drop out as
we go along in time. But really quite a comprehensive
longitudinal cohort. So, this is…, and we’ve just completed
the longitudinal analysis. I’m not going to be really talking
about any of those analyses yet, but it’s still
in the pipeworks. But I will present to you one of the key
papers that’s come out from that baseline data
the first time point. But just before I do,
I’ll just highlight some of the issues and challenges working and trying
to scan not just ADHD children, but
children in general. You know, the MRI is quite a medically
intimidating environment for children, and it’s, you know, quite a challenge to
get this sort of behavioral cooperation. You know, we don’t want the child feeling
anxious, we don’t want it affecting them. We also don’t want that anxiety to mean
that they give up and don’t go in for the scan, and that ends up wasting
a lot of research funding as well as the time of the staff
and the participants. But also emotion is a really,
really important point. And I’ll talk again on the next slide
about this. And so, we need the children’s cape as
still as possible in the scanner. Children in general move more than adults,
but obviously, children with ADHD stereotypically
move a lot. So, this is a challenging population to
work with in the scanner. And so, having good preparation is really
key for good quality scans. And we try and start this right from the
recruitment. You can see up there in the top right-hand
corner, this was our participant information
and consent book. So, this was sort of targeted towards even
the children and the adolescents themselves to really engage them from the
get-go rather than having this sort of big black and white legal-looking document
that the parents sign and the children have no real say over,
we really wanted to engage the children and this worked
really, really well. We ended up having the kids pestering the
parents to call us to say, “Yes, we’re going to
participate.” We also have a mock MRI scanner at the
children’s hospital. So, this is a full shell of an old scanner
without the magnet inside. And this allows us to have a little
session before the scan with the child, get them used to and familiar with that
scanner environment, be able to play the noises, you know,
there’s a lot of loud noises that they hear so they can get familiar with it and
get comfortable and reiterate the importance
of keeping still. And so, from the 180 we had scanned at
baseline, only six then failed to go through to the main scan
from this sort of preparation, which I think’s a
really good outcome. And then also we give the children,
a sort of certificate at the end with a picture of
their own brain. Just to highlight this issue of data
quality. The top row there is good quality data. We’ve got a structural image on the left
side and a functional image on the right, and I’ll just play the functional image
for you. You can see
very little movement. You see a little bit of pulsing as blood
gets pumped up through the middle of the brain,
but very still brain. The brain on the bottom here has moved a
lot during the structural scan, and you
can see it’s very blurry, just the same as if someone
was running through a photograph. And this makes it very difficult to sort
of be able to identify the boundaries between
different structures. And if I show you the functional scan,
you can see that bouncing around, it looks more like they’re headbanging
at a concert than lying still. And this sort of data rarely becomes
unusable. And once again
that wastes dollars and time. But what’s becoming more noticed these
days is that even in good usable quality data,
motion is still having a big impact on some of the
metrics that we get out. So, it can affect the cortical thickness
measures or the long range sort of functional
connectivity measures. And where this becomes really problematic
is when you’ve got, you’re comparing two groups, and there’s more motion in one
group than the other because that can then artificially give
differences in the brain metrics. This is just a slide from some work. One of my Ph.D.
students was looking at… Well, she’s been looking at the head
motion in an MRI scanner. And obviously, children with ADHD have
greater head motion, but they also have poorer sustained
attention when we look at just the dimensionally
those children with poor sustained attention also have
higher head motion. So, what this is sort of showing is that
rather than just being sort of an artifact of ADHD, having head motion as an
artifact, it’s really interwound within the phenotype
and the presentation of ADHD. So, simply getting rid of data with lots
of motion might not necessarily be the right answer, or regressing out the
influence of motion might be actually regressing out the
phenotype of interest. So, this is a particular challenge across
the field at the moment. Other issues when looking at children,
adolescents, obviously, braces. And children with braces can go into the
scanner, doesn’t cause them any harm, but it causes just big distortions to the
magnetic field. So, these MRI images down the bottom,
that individual doesn’t have a gaping hole in the front of their head,
that’s a distortion due to the magnetic field
based off the braces. And this significantly impairs functional
images, much worse than structural images. All right. So, that was just some of the few
challenges of working with children. But what I’m going to talk to you now is a
paper that came out earlier this year which was like the first key baseline
paper from our data. I’m trying to overcome some of these
limitations. So, when we look at some of the existing
MRI data in ADHD, there’s, once again, like the clinical phenotype,
there’s a lot of heterogeneity in the findings about what
brain regions are involved. A lot of the early studies, you know,
based on a priori theories really just focused on deep brain striata,
basal ganglia structures, volumes as well as the prefrontal cortex
like those in the bottom
right-hand corner. But as imaging techniques and technology
has advanced and the resolution has got better, we’ve been able to look at
focusing on whether the cortical thickness and the surface area as well as some of
the white matter microstructure. And then in the bottom left-hand corner
there, that’s really the only other longitudinal cohort in ADHD and this is
almost 10 years ago now but has sort of suggested there might be this delay in
maturation where by late adolescence, ADHD brains catch up with
typical controls. But there’s been a sort of a lot of
inconsistencies. For example, in the top right-hand corner
there, this is actually a large, what
they call mega-analysis. So, multiple sites from around the world,
this consortium have all come together. They use the same processing pipeline to
pull out the same bits of data, and then
they pull them together. And, you know, this is just one brain
region, and it does over show that the chordates are overall smaller
in children with ADHD. But you can see a lot of variability
between sites there and even as four or five sites that show
greater volume in ADHD. And so, most of the recent meta-analysis,
whether it’s in instructional diffusion or resting state data
are really finding very little if any consistency in
brain regions across ADHD. This is a real problem. The one thing about most of that data,
though is they typically will be looking at one
aspect of the brain. They will look at the volume,
or the thickness, or the surface area, or
white matter properties. But we know that these structures don’t
operate in isolation and particularly when we’re looking over development,
they don’t develop in isolation. These figures just shows you just some
sort of the different trajectories as the brain develops in different brain regions
and how there’s different tempos and timings of that development in whether
it’s cortical thickness or the white matter volume
or the different microstructural
properties of white matter. So, there’s been some advances in
statistical analysis that now allow us to be able to model sort of patterns of
covariation across imaging modalities. So, just to introduce the sort of concept
of what we’re trying to do here, the technique we use is called a linked
independent component analysis, which just stems from just independent
component analysis, which has been used for a long time
and not just in imaging. This kind of comes
from any sort of signal detection. Another sort of scenario to think about,
you know, our brains are very good at doing this,
is think about being at a party. You’ve got this single sound source coming
into your ears, but your brain can take spacial and temporal components of that to
pull out different components so that you can dissociate different people speaking,
or the band playing, or trumpeter. So, your brain’s very good at doing this,
and it’s been now applied to neuroimaging techniques for quite a long time to try
and identify, say, artifacts within the data or more recently
for that resting state. So, to be able to identify parts of the
brain that have a similar time course across the sequence so that even though an
individual’s at rest, we can still pull out a motion network,
or a visual network, or
an attention network. So, Linked ICA’s just taking that a step
further to look at not just what areas that co-vary within an imaging modality,
but also what varies across different
imaging modalities. So, that’s what we did with our data. We took all our brain volume data,
cortical thickness, surface area as well as two white matter microstructural
measures, which is fractionalized atropy or FA
and main diffusivity, or MD. And we extracted 25
specially independent components or brain patterns
that co-varied together. And you can see those
25 components in the bottom right-hand corner
with how they were made up. Some of them were heavily driven by one
particular metric, say, for example, surface area, but others there
have fairly mixed contribution from
different imaging modalities. So, these are done in a completely
data-driven way. Then on the phenotypic side,
we took a heap of different phenotypic variables,
loosely categorized into individual factors such as subject’s age, weight,
puberty stage, clinical factors. So, I’ve mentioned this, you know,
moving away from a categorical ADHD or not to looking
at dimensions. So, we’ve taken a symptom dimensions as
well as whether they’ve got comorbidities. We took a bunch of cognitive factors like
their IQ, their academic functioning, working memory,
sustained attention. We have some retrospective perinatal
factors such as their birth weight and whether they spent time in a neonatal
intensive care unit or whether their mothers smoked or drank alcohol during
pregnancy. And then a range of family factors as
well, such as parent education, parenting styles, family quality of life,
stressful life events. So, what we want to do is to try and
account for a lot of this heterogeneity we’re finding both in the imaging
literature as well as across the heterogeneity that’s seen within the
populations. We took these whole lot of phenotypic
factors that vary. We took these brain factors,
brain patterns in a data-driven way, and we looked at multivariate
associations. So, can we reveal
associations between the subjects’ phenotype and
their brain patterns? This was done using a canonical
correlation analysis, and it pulled out four significant
independent relationships. And I’ll step through those over the next
few slides. So, the first one, I’ll firstly just
orientate you to the figure cause this is the format that presented
in the next few slides. You’ve got five rows there. That’s the brain volume at the top
followed by the diffusion factors, FA and MD, cortical thickness
and then cortical area. Warmer colors increases,
and cooler colors are decreases. So, what this is showing you is whole
brain across different imaging modalities. This pattern of increases and decreases
across the brain that were associated with the
phenotypic variables. Now, this first one’s not particularly
exciting or interesting. It was mainly driven by their overall head
size, but as we might expect it was, you know, head size has been associated
previously with being male. You know, males have larger head sizes. Larger head sizes
typically have better cognitive performance,
comes from better socioeconomic status, and had a higher birth weight
to start off with. But we can then
remove that variance of head size before moving
to the next factor. And we’re referring to this as our marker
of development. So, independent of the head size,
this pattern of increases and decreases across that whole
brain was heavily driven by the individual’s weight,
puberty stage, and age. On the clinical side,
it was on the inverse, it was less associated with medication and
also hyperactivity. That figure only shows the ones that were
associated with less than 0.001, but there are other significant factors
there which including hyperactivity. And if we flip that around the other way,
what that says is that if an individual has a brain that’s less developmentally
mature than their peers, they were more likely to have
hyperactivity and be medicated. However, the third marker was really what
we’re defining as our ADHD phenotype. So, this particular pattern of brain
increases and decreases across those imaging modalities was heavily associated
with hyperactivity. Also, it was associated with inattention,
but hyperactivity explained sort of twice the
variants than inattention. So, these individuals were also more
likely to be male, older but less pubertally developed,
more likely to have comorbidities of autistic symptoms,
more likely to be medicated, more likely to have stressful life events,
have less consistent parenting, poor quality of life
are more likely to have spent time in a neonatal
intensive care unit. Our last marker
which was not quite as robust as the others but was a
marker of cognition. So, this wasn’t
representative of overall IQ, but it was a particular pattern
of cognitive deficits. So, these children were more likely to
have poor academic achievement, poor language skills,
poor visual-spatial reasoning, poorer cognitive flexibility,
their mothers were more likely to have smoked
during pregnancy. On a clinical aspect,
it was associated actually less with ADHD and more with irritability,
which is very interesting because irritability often has a high comorbidity
with ADHD. And this really has pulled it apart as
separate components, which is
really quite nice. We were kind of really excited that they,
in a data-driven way, essentially pulled out these separate
components that have all been in some way related to ADHD, you know,
overall brain size, development, obviously the ADHD phenotype,
and also cognitive profiles. So, even though we didn’t want to start
out looking at a categorical approach, we wanted to look at this dimensionally,
given that there was kind of a heavy ADHD phenotype
brain pattern. We looked to see whether this could be
differentiated categorically, and we could, that brain pattern did
significantly differentiate between ADHD
in controls. But the effect size is not huge. You can see there was a lot of overlap
there, but importantly, other variables and other markers of
development cognition didn’t significantly
differentiate ADHD. So, we thought this
was really nice in our data, but we wanted to think,
“Well, you know, can our marker of ADHD actually predict
ADHD in an independent cohort.” So, there are some
other cohorts available that are publicly available, this one,
the ADHD 200 cohort. Now, they didn’t have the full multimodal
imaging that we acquired, but just simply taking their structural
volume data and applying our brain marker to that, we were able to predict from our
marker of head size predicted overall head size in that independent cohort,
our marker of development, predicted age in that cohort,
but importantly that marker of development didn’t predict ADHD or IQ whereas our
marker of ADHD did predict ADHD symptoms
in that cohort. And interestingly, it predicted that
stronger in hyperactivity than inattention in the same way
that it did in our data. And then our cognitive marker didn’t
predict anything in that cohort, but there wasn’t cognitive measures other
than IQ if you remember, IQ wasn’t associated
with that component. So, just to… In summary of this paper,
what we did is we took this data-driven analysis with multimodal imaging and
multi-informant phenotypic data. We identified this sort of novel set of
brain images that was able to account for variation in the children’s
development, clinical factors as well
as cognitive factors. And all these factors have been to some
extent, associated with ADHD. So, you know, it might be that ADHD is a
summation at different levels of these factors which have distinct anatomical
foundations, which may then explain why we have such heterogeneity
in the findings. And then importantly that we were able to
predict this in an independent cohort. All right, so that’s just
a key baseline paper. We’ve got lots of other students and
things working on different projects at the moment, but sort of what’s coming
next with this cohort and this dataset. As I mentioned, we’ve just recently
finished the longitudinal imaging. So, I’m really excited about getting stuck
into looking at how these brains change over time and how they change for those
that have persistently poor outcome versus
a better outcome. Other things that we’re doing with this
cohort is, I mentioned this large
mega-analysis before. This is an international consortium known
as ENIGMA, and they have a range of
different working groups. So, we’ve been
contributing some of our data towards this consortium for
the ADHD working group. We’ve also been collecting genetic samples
from these kids, and they’re currently being genotyped, and this data is also
contributing to a larger international consortium,
the Psychiatric Genetics Consortium. And through some other funding,
we’re currently doing some
epigenetic testing. If you’re not familiar with epigenetics,
at DNA, our genetics is stable. It’s the same in all cells,
but our epigenetics is kind of the switches that turn on and off gene
expression, and that happens naturally across our life span
and our development but it can also be influenced
by environmental factors. So, can we see sort of different
epigenetic factors that are associated with the sort of functional
outcomes with ADHD? These are some of the cool new prospects
to just stay tuned to. And, you know, I haven’t done all this
work alone. There’s been a big group of staff and
students as well as the families that have been involved in coming in
over multiple years. So, just want to acknowledge their
contribution and I’ll open the floor for any questions
you might have. Thank you. – Thank you so much, Tim. That was great. So, listeners, feel free to type in your
questions in the question box, and I’ll read them
out to Tim for him to answer. So, we’ve got a few questions that came in
throughout the presentation. One from Brett, “Looking at the variables,
have you looked at phenotypical data against metric measures
like the DSM using autism? Very interesting to see if ADHD-specific
patterns map to imaging.” – So, is the question… I think assuming the question’s sort of
looking at that overlap between
ADHD and autism? Yes. That’s certainly something that we are
looking at within our data set. A number of students that have been
looking at that overlap one particular that looking
at the imaging data. She’s currently looking at the white
matter tracks involved in sort of emotion regulation and
looking at that overlap. So, yes. That’s certainly questions that we’re
looking at, and I think that’s sort of a big area of interest at the moment,
particularly with the changes now to the DSM that allow that kind of morbid
diagnosis. Thanks for the question. – [inaudible] Glen [SP] has
asked, “Has anyone looked at possible correlations with
psychometric data?” – Can you just clarify
like on what you mean by
psychometric data there? Oh, personality. It’s not a literature that I’m across,
but we actually have been, I think, in maybe our most recent wave,
have collected some sort of
temperament measures. So, that’s something that we have
potentially got in the data to look at. Yeah. Interesting. Thanks. – Tim, as you can tell, Alison [SP] is
more about how they can find out more information
on your research. – Yeah. Certainly. So, we have a Children’s Attention Project
website that’s up, right actually through the
Murdoch Children’s Research Institute. But if you come directly to the School of
Psychology at Deakin University, the Brain and Cognitive Development Lab,
there’s some links there. Feel free to email me,
get in contact with me if anyone’s interested in
any aspect of our study. Yeah. – Yeah. It sounds good. And we’ve got a question from Ben. He says, “You mentioned that the study
investigated factors that predicted better and poorer outcomes
for these individuals with ADHD. Is that work that is still in progress or
where this amount says?” – Yup. So, yes. We’ve… So, in terms of that baseline non-imaging
data, we’re currently working out a three-year
longitudinal paper at the moment, so I won’t talk about those
findings just yet. But in terms of the imaging trajectories,
we haven’t got to yet. We’ve only just finished that third wave
of data collection earlier this year. And so, that’s things we are currently
working on. But please stay tuned. – That’s great. Thank you so much, Tim. We might wrap it up there,
but if anyone has any follow-up questions, feel free to
email Tim directly. And thank you so much for listening in
today and participating in today’sDeakinAlumni and
School of Psychology Webinar
. – Thank you.

Leave a Reply

Your email address will not be published. Required fields are marked *