Population Neuroscience and Public Health

May 03, 2023

Information

YCSC Grand Rounds April 18, 2023

Henning Tiemeier, MA, MD, PhD, Professor of Social and Behavioral Science and the Sumner and Esther Feldberg Chair of Maternal and Child Health - Harvard T.H. Chan School of Public Health

ID9893

To CiteDCA Citation Guide

00:05Good afternoon, everyone, and welcome
00:07to Grand Ryans and especially to
00:09everyone joining us on Zoom.
00:11And I'd like to remind you that for the Q&amp;A,
00:13please feel free to put on your video
00:15cameras and we'll project you here on
00:17our screens here in the Cohen Auditorium
00:20and we'll hope for a lively discussion.
00:22Now as usual, we just want to preview
00:23a couple of our presentations that are
00:25coming up over the next couple of weeks.
00:26And so next Tuesday,
00:28we will hear from Doctor Jessica Cardena.
00:30And this is a very special by
00:32Ola Barnard lecture series.
00:33And so Doctor Cardeno will be talking to
00:36us about what we can learn from Latino
00:38mothers and what Latino mothers can
00:40teach clinicians about trauma and recovery.
00:43And then a special date for
00:45your diary on Monday.
00:46And we have Doctor Tracy Bale
00:48coming to give a seminar in the
00:51Division of Reproductive Sciences.
00:53So that's in the department of OB GYN,
00:55my other home department.
00:56And so on Monday from 12:00 to 1:00,
00:59Doctor Bale will be coming to talk
01:01to us about extracellular vesicles
01:02as a novel form of communication
01:05between the mother and the fetus.
01:07And as you'll all know,
01:07Doctor Bale has done some seminal
01:09work trying to uncover the molecular
01:11mechanisms that underpin the
01:13intergenerational transmission of stress.
01:15And now to our speaker for today,
01:17it is my distinct pleasure to welcome
01:19Doctor Hennington Meyer to the Child
01:21Study Center for the very first time,
01:22I'm told.
01:23And we did a little bit of history of
01:25the Child Study Center earlier on and a.
01:26Tour So Doctor Timmeyer is joining us
01:29from the Harvard School of Public Health,
01:31where he is the Professor of Social
01:33and Behavioral Science and holds
01:35the Sumner and Esther Feldberg
01:36Chair of Maternal and Child Health,
01:38where he also directs the Maternal
01:40and Child Center for Excellence at
01:42Harvard School of Public Health.
01:44And, of course, Dr.
01:45Timmeyer also holds a professorship at
01:47the Erasmus University in Rotterdam,
01:50where, as many of you know,
01:51he set up the Generation Rotterdam cohort,
01:54the Gen. R cohort.
01:56Which has made a tremendous
01:58contribution to our understanding of
02:00how the environment shapes individual
02:02differences in child development.
02:03And I hope we'll hear a little
02:04bit about that today,
02:05as well as many of the other
02:07initiatives that Doctor Tiamar is
02:09involved in since moving to Harvard.
02:11And of course,
02:12he has published prolifically and is
02:14regarded as a ISI highly cited researcher.
02:16So please join me in giving
02:18a warm child study.
02:19Welcome to Doctor Tiameyer.
02:20Thank
02:24you. Thank you very much.
02:27Let me put on my mic and thank
02:29you very much for the kind,
02:31very kind and warm introduction
02:34and the invitation to come here.
02:36Indeed, I'm quite proud to talk here.
02:38I should say that because just teach
02:41currently again the the course child
02:44Psychiatric EPI at Harvard and on
02:47my third slide there I show the Yale
02:49Study Center and the work of gazelle,
02:52which I think shaped longitudinal studies.
02:56More than many others or anybody
02:59else was that introduction.
03:00For those that are also interested in
03:03more recent work I'm doing or more other
03:05work on the maternal child space space,
03:07I must disappoint you or focus on
03:10generation R still doing much of my work.
03:12What I did is I okay is I selected work
03:18from ongoing studies or older studies even
03:22because I do much population or imaging.
03:24I'll show you.
03:25And the theme I thought was answering
03:28an all discussion saying this work
03:30should do now that you're at the
03:31School of public health is not really
03:33relevant to public health at all.
03:35And after 20 years or 30 years of imaging
03:38research, it's still not relevant.
03:41And that doesn't insult me.
03:43I think it's a fair critique,
03:44but at least I have to live with
03:45it and address it.
03:46And that's what I'm trying to do
03:47with you today. Discuss it with you.
03:49Could it be relevant?
03:50It's not so obvious.
03:53So yes,
03:54they asked me to do learning objectives.
03:56So here you are a bit,
03:57it's a bit about the prenet exposures,
03:59which I'll Kieran is working on.
04:01So I'll focus on that.
04:02And the question really is,
04:04is it identified?
04:05I don't think that's the
04:07learning objective to be honest.
04:08It would be discussed with me how
04:10child imaging might possibly in
04:13theory a bit impact public health.
04:16What am I talking about?
04:18I see Euroscience population of science
04:20not as broad as somebody like Thomas Powells.
04:23Thomas Powers, I would see.
04:24It's really the intersection of,
04:26if you wish,
04:28population research or etymology
04:29and neuroscience.
04:30Essentially that's what happened in genetics,
04:32that genetics has been now 1520 years really
04:35infused with genetics as we just talked,
04:38epidemiology,
04:38but now also influences
04:40epidemiology with new methods.
04:43And then I'll focus on prenatal exposures,
04:45psychosocial or chemicals.
04:46I've got one more chemical
04:48exposure pull that up after I met.
04:51Somebody yesterday night,
04:53I thought that's a good one,
04:57how that impacts child development.
04:59I'll start with what I think
05:01is not public health relevant.
05:02So I thought I'll start with
05:04something where I think it's not
05:08what imaging research is not and I
05:09start with not other people's work.
05:11That's not very cool.
05:12I start with my own work,
05:14so I'll show you my.
05:16Were my best publication last
05:17year or one of my nicest,
05:19but I don't think it is any
05:21public health relevance.
05:22It's answering the question child psychiatry.
05:26Really it's giving you an example
05:27of that because much of my work
05:29or all of my work was funded under
05:32the premise that it will inform in
05:34the prediction and the causality
05:36of child psychiatric disorders.
05:37And now 20, not 15 years later,
05:41what have we delivered?
05:43It's this type of work.
05:45Can we really predict adolescent
05:47hallucinations with imaging?
05:49Does it add anything?
05:50So last year we published work on
05:53this question, Public Health Relevant.
05:55You ask yourself,
05:56can we predict adolescent hallucinations
05:58would be very, very important.
06:00We measured that in Generation
06:02R at 10 and 14 years.
06:05It's actually quite easy to measure.
06:07You can ask the adolescents themselves,
06:09You can ask them to hear voices.
06:12You have strange thoughts.
06:15I don't know if anybody here has an idea
06:17how prevalent that is at age 10, At 14,
06:20Any idea if it's a fringe thing happening
06:22at 2% of the population or 10 or 15%?
06:25But actually if you ask them,
06:29do you hear voices, it's 25% easily.
06:32And that is not just waking up
06:35after dreaming, it is really work of
06:39Keleha and Mary Cannon in Ireland.
06:42Has shown it's somewhat less frequent,
06:44so it goes down to 15% if you wish.
06:46If you really get them bothered by voices,
06:5029 is really what you get with these
06:52population assessments if you do it crudely.
06:54But trust me, it is easily 15% at
06:57age 10 and then it drops to 12%.
06:59And again, if you do it more carefully,
07:01it would probably be six, 7% at age 14.
07:03That hear voices,
07:05which is huge, don't forget.
07:07Don't confuse that with schizophrenia.
07:09That's nowhere near schizophrenia.
07:10Actually, if you know their work,
07:12it predicts depression, anxiety,
07:15borderline much more than schizophrenia.
07:18And we did,
07:19so that's the special thing.
07:20We did repeated imaging at age 10 and age
07:2414 so we can show does the brain change.
07:27We can also say, can we predict it?
07:30In the paper in Biological
07:32Psychiatry last year we showed
07:33something after all lots of studies,
07:37different approaches,
07:38different work with the brains.
07:40What we found really is if you hire voices,
07:44then the typical decline,
07:47this is sort of exaggerated.
07:48This is a bad curve.
07:49It should be much more than sort
07:50of trajectory curve.
07:51But forgive me for that that the
07:53decline in Gray matter which.
07:55Originates probably much earlier than age 10,
07:57probably age 6 onwards is a tiny
08:00bit far faster in those that have
08:09new onset hallucinations at age 14.
08:15I'm showing this.
08:17It is an association.
08:18It has a tiny effect size.
08:21You need a few thousands 2000s to
08:25find it as a tiny effect size.
08:28It is a specific,
08:30it is much of your Gray matter,
08:33and actually it also maps on
08:35other psychiatric problems,
08:36so it would not be that very
08:39specific for hallucinations.
08:40You can also zoom in and find other
08:42structures, of course we did that.
08:43And the hippocampus,
08:46that's what the small A says
08:48it's The effect is again small.
08:51It survives multiple testing,
08:52correction for other structures.
08:53It's a tiny effect.
08:54Again, it is unspecific.
08:56The conclusion here is useless.
08:59As a predictor,
09:00I have little doubt over and above any
09:03prediction model which we published.
09:06These brain imaging does nothing.
09:07You can better do predict
09:10with socioeconomic factors,
09:12better predict with clinical factors.
09:14You can better predict with well-being.
09:17It does not predict and this is
09:19the biggest imaging study so far.
09:21So it may be that one day all of
09:23you will search for more specific
09:25markers and we'll do resting
09:26state analysis and whatever.
09:28But we had this unique data
09:30set with repeated imaging and
09:32repeated hallucinations over
09:33the really relevant period.
09:35I would give this a one out of five
09:38in population public health relevance.
09:41It does not add to any child
09:43psychiatric clinicians.
09:47Addiction, therapeutic,
09:49understanding model.
09:50I would say we've done
09:53lots of this type of work.
09:54It's fascinating, it's fun.
09:55I think it's important to understand
09:58that the brain can predict,
10:00but it is not clinically useful.
10:02Let me go on with transition to the work.
10:05I'm going to show where I think we can
10:08discuss public health relevance And again,
10:12this is a crude analysis, I know that.
10:14Actually more fine grain didn't predict,
10:16more if you do multiple testing
10:18correction and the prediction was small.
10:21This is a paper I'm not going to discuss,
10:22I'm just going to recommend it for your read.
10:24From last year we said thought
10:26it harder that the population of
10:28science is the best paper of the year
10:30and so it got our prize for that,
10:33whoever cares.
10:35And what it does is it uses the biggest
10:40databases like the UK Biobank and
10:42the ABCD studies and others to show.
10:44That for a
10:49if you don't zoom in on our ones
10:52but you take a broader approach
10:55for resting state and for volumes
10:59that you need to find anything.
11:02They say it's three to 6000 people in
11:05the general population to find anything
11:08you can argue in your clinical sample
11:10it's different there's a letter or.
11:12An answer to nature arguing that very
11:15recently I actually fundamentally
11:16disagree with that letter.
11:18I think they have it right.
11:20It's my own experience too,
11:23and the only thing I'm not so sure,
11:26and that's the judgment that's out.
11:28This analysis is clearly only cross-sectional
11:33and actually I'm not so interested
11:36in cross-sectional prediction.
11:37So we would have to move to longitudinal
11:39and if you've got repeated brain measures,
11:41I would argue because you control
11:44for quite a bit and you have change
11:47that could be different, although we
11:49don't know what's the interval change.
11:51Secondly, they have very poor phenotypes.
11:54You could argue,
11:55I think they should have used multiple
11:57informant and other approaches, but anyway,
11:59it was all that critique I think.
12:02It's very humbling that all of a sudden
12:04after so many years where we had studies
12:06of 1520 people and found big effects,
12:09we now have people that say if we want
12:11to do it well, we need 3 to 5000.
12:15So I would argue in child psychiatry so far
12:18without very few exceptions you can think of,
12:21you know, but very rare syndromes,
12:24not so sure OCD, there's some debate that
12:26that's quite specific, but otherwise.
12:29I think it's poor discrimination,
12:31poor specificity, poor sensitivity.
12:33We've done machine learning.
12:35I'm not talking about that.
12:36To overcome that,
12:37what we find is a very small signal and
12:39actually something we already knew.
12:41It's quite broad changes in
12:44externalizing behaviors,
12:45nothing very specific either.
12:47So I think in sharp psychiatry,
12:49my research has not contributed
12:51that much for public health.
12:56That does not mean it's useless, of course.
12:58I would like to discuss prenatal exposures,
13:00some old work and then zoom in more
13:02recent work, ongoing work even.
13:05And we identify important introduction
13:07influences on the Turtle Shine house.
13:09And we've done a lot over the years.
13:11We sort of in generation are measured as
13:14much as we could and we were quite creative.
13:17We've got environmental toxins,
13:18we've got thyroid poverty.
13:20That's the recent thing that I added to
13:22the list because I was interested in that.
13:24Depression.
13:25So the bold ones are zooming on today.
13:27One or two of you will know earlier work,
13:30but poverty is very recent and the
13:33environmental stuff is just out last year.
13:37So that's cool.
13:38Discuss with me what you think is the
13:41role of imaging, which is so well funded.
13:43You know, if you take the European
13:45funding in the neuroscience,
13:47probably 1/2 goes to brain imaging,
13:50which is shoot.
13:52Not as much in the US
13:54interestingly relatively speaking,
13:55but a lot Okay just a bit about
13:59Generation R as a prospective cohort.
14:00It started in early fetal life but
14:02early the inclusion we promised
14:04and we had funding for 10,000,
14:06I don't know for whatever reason the
14:07end of the year came and we had to stop.
14:08So we didn't manage the 10,000
14:10but we got close,
14:11it's 10,000 if you know who's active.
14:13It's still more than 5000 are
14:15contributing participating 6000 which
14:17is very good if you start prenatally.
14:19I think it's much better in a way
14:21than ABCD because they have a 15%
14:23response rate at baseline or lower.
14:25So this is a 62% response rate
14:28and then the Dutch majority group,
14:30it's actually 70%.
14:31So it's more selective in minorities,
14:34it's urban and multiethnic and I do
14:36because I have a slide later on this
14:38ethnicity normally I sort of gloss over it.
14:40Note that if you're on Rotterdam,
14:42it's not much different than in many of
14:46the Americans cities that about half.
14:48Of the population is Dutch means
14:53that has Dutch ancestry origin
14:5610% would be other Europeans,
14:57So that's expats largely.
15:00And then you've got both migrant
15:02or guest worker I should say,
15:04which are the Turkish for example,
15:06and the Moroccans.
15:07And then you've got colonial history,
15:09people like tsunamis, Cape roses,
15:11also guest workers,
15:12but Dutch Antilles.
15:13Are ex colonies of the Netherlands where
15:16people could migrate easily into meaning.
15:19It's a very dangerous city.
15:20And yeah, that's important because we'll
15:24talk about poverty just about the measures.
15:28I have no pointer,
15:29but I have a cursor,
15:30I'm told,
15:31so I don't want to go through measures.
15:33Nothing is more boring than telling
15:34you what we measured in the study,
15:36but we measured a lot ultrasound
15:37in the beginning questionnaires,
15:39lots of motor development was exciting.
15:41We have IQ measures but also actually
15:43of the parents which are the mother.
15:44It's very important to control
15:46for baseline confounding.
15:48If you have intrauterine infectors,
15:49what is sort of genetic
15:52background and then the imaging,
15:54I'll focus much of my talk.
15:56On the imaging at age 9 to
15:5810 which is at 4000 people.
16:01I have one study later where we do a
16:03follow up of the imaging which we have.
16:05This is actually a typo.
16:06It shouldn't be 4050.
16:08This should be 3 thousands
16:10and 52 hundred 3200.
16:11Just copy pasted the wrong thing here.
16:14So we have now three wave
16:16completed and the 4th wave ongoing.
16:17In total it would be 6000 different
16:20individuals that have been scanned
16:22of 5500 and the overlap is not that
16:24big but it is there to do nice multi
16:27level analysis over three waves already.
16:30I would like to start with my
16:31classical one of my classical
16:33papers maternal depression.
16:34So I think there we can learn a
16:36bit about public health relevance
16:37and actually I'm saying that also
16:40because it informed a study or
16:42work that I'm doing currently.
16:43Maternal depression from fetal life forward.
16:46What I'm trying to show you is
16:48that we've measured maternal
16:50depressive symptoms at three time
16:53points during pregnancy.
16:54This would have been forgive me
16:56that the error is not quite good.
16:58It should be after birth at two months,
17:00after birth at three years.
17:03We didn't use that and we used
17:04it at 9 to 10 years.
17:06So 4 measures of maternal depression 1/2
17:09just after birth in the early childhood and.
17:13At 10,
17:14why is 10 interesting?
17:15That is interesting because that
17:16is cross-sectional if you wish.
17:17Was the brain imaging,
17:19focusing on the brain imaging at
17:2110 years when we measured 4000
17:23children and not all in study
17:24at the end there will always be
17:26only 2000 or 3000 in the study,
17:28but that's at that time a very big study.
17:31Certainly the biggest study was
17:34prenatal exposure assessment,
17:36prospective prenatal exposure assessment.
17:40And I always ask when I see the
17:42slides and those who have not seen it,
17:44what time is there a strongest relation
17:49of maternal depressive symptoms to
17:51the brain of a child measured at age 10?
17:57So we've got it at during pregnancy,
18:00just after birth, early childhood
18:04and cross-sectional with the MRI.
18:08And the question is when?
18:11Is there a relation between the
18:15maternal depressive symptoms
18:17and the volume and connectivity
18:21of the child brain at age 10?
18:24So is there a long term influence
18:27from prenatal life forward?
18:29Is there an influence of early
18:32after birth perinatal depression?
18:36Is an influence of childhood depression
18:39or an influence of cross-sectional just
18:44concurrent depression to the mother?
18:47Talk about structure of the
18:49brain of the child at age 10.
18:55So I'm not, as I sometimes do,
18:56pull somebody up and say what
18:57do you think I'll do it myself.
18:59You can think many people would think
19:03it's either prenatal depression.
19:05That has a big effect because that's
19:07when the child is in the womb.
19:09So you would think that the mother's
19:12depression influences her Physiology,
19:15and that impacts the child.
19:18You could argue for just after the birth,
19:21because that's a key period of attachment.
19:25You could even argue somewhat
19:26less for the childhood,
19:28but you could argue for that because it's a
19:30long period of childhood upbringing anyway.
19:33If you look at this,
19:35this is just very broad.
19:36Total measures, Total white measure.
19:38Total Gray measure.
19:39Because we start with
19:41hierarchical approaches,
19:41doing big parts of the brain
19:43and then zooming in on specific
19:45regions if we find something.
19:47You can look at these small effects.
19:49They're actually translatable in centimeters.
19:52Cubic.
19:55You can see nothing was
19:57the white matter much.
19:58And if you look at the Gray matter.
20:01There is a period,
20:02two months where there is an effect and
20:04again that survives multiple testing.
20:06So if you want an answer from this,
20:09it is not.
20:10And I've said that many times,
20:11for me, this is one of the
20:12big it's a few years old now,
20:14four years ago we published it.
20:16It's
20:19not the prenatal exposure that is most
20:22important and we see that in some of this,
20:24it is actually just after birth.
20:28Where we see an effect and
20:30that's actually very consistent.
20:32So there's two ways to look at the data.
20:36One is prenatal is not everything and
20:39sometimes the Doha people would tell you.
20:42Secondly, effects are small and if anything
20:47that is a small effect postnatal depression,
20:50which makes sense if you know
20:53the literature and attachment,
20:55maternal bonding and how important it is.
20:57To have and how that is impacted
20:59in clinically depressed mothers.
21:01If you look at and I'll show later some,
21:03I think some more DTI.
21:04This is a slide of how we look at DTI.
21:06We sort of don't integrate it all.
21:08We look at different tracts
21:10which we then sometimes sum.
21:11So this would be the connectivity
21:13in the white matter.
21:14You measure that with two measures FA or MD,
21:17but it essentially shows you
21:18the integrity of the in these.
21:21Different tracts.
21:22We measured that. Well,
21:23that was just the global brain measures.
21:26And I can tell you this effect is
21:28quite broad across parts of the brain.
21:31So it's not just in the temporal
21:32lobe or the frontal lobe.
21:33We find it a global effect.
21:35And then we also looked at the
21:37DTI and what is interesting,
21:39that's not so surprising.
21:40Well, there was nothing in the white matter.
21:42We saw that the tracts, the general tracts,
21:46the integrity of the tracts.
21:48Again, depression at 2:00.
21:51Months Postnatally there was
21:54less integrity of these tracts
21:56together and trust me there's not
21:58a single track that does it.
21:59These global integrity of tracts is less,
22:03is less clear is there's less
22:06integrity in these tracts.
22:08And then in fact was the depression at
22:11two months on the child brain of 10 years.
22:14So it's different exposure times was
22:16one outcome time always at 10 years.
22:18So you see the effect and there's
22:19nothing Again,
22:20there's prenatal
22:25If we discuss public health relevance,
22:29you will not want me to say we now
22:33found that maternal depression is
22:35important because there's fifty years
22:37or 100 years of research to show that.
22:39You might want to say, wow, he has a
22:44way of finding sensitive periods.
22:48And that's why I would think perhaps,
22:51but really, honestly, I don't think so.
22:54And I'll tell you why I tell you that.
22:58And I know people in everywhere think
23:01differently that you can with Social
23:03adversity study sensitive periods.
23:07I actually have tried to do that
23:10now with measure of homelessness
23:11and other work in my group.
23:13We feel that is largely flawed.
23:17Because of the following thing,
23:19depression in mothers does not
23:21occur in isolate meaning over time.
23:24What I mean is that is a mother
23:26that is depressed at two months
23:28after birth has likely some elevated
23:31symptoms already during pregnancy.
23:33Not only likely, very likely,
23:36meaning that all these poverty, abuse,
23:41depression, all these risk factors,
23:43all these social adversities are studying.
23:46Have a high carry over and we cannot
23:50validly or have seen very little studies
23:54to validly study the period specific
23:56exposure because then you would have
23:58to have people that have it only in
24:00this period and not in the others.
24:01And if you see how carefully they
24:04account for the other periods,
24:06I can tell you in most models I've
24:08seen that is flawed, including my own.
24:11So I'll show you why it's flawed
24:13and this is the trajectories.
24:16It's flawed because the mothers who
24:17have that peak of depressive symptoms
24:19at two months were actually those
24:21that were on average as a group.
24:23If we just do these trajectories
24:25and we classify them in groups
24:26and we forget about that,
24:27this is of course a continuum, this,
24:29this series of continuum on that level.
24:31But if we do them in four groups,
24:32we see this group that actually I
24:34can tell you carries the results,
24:36has high levels here,
24:37super high levels here and then keeps
24:39on in the all these ten years after.
24:44Assessments to be reasonably high because
24:46this is 0.7 is exactly where the clinical
24:49line of clinical severity would have been,
24:51meaning that there is a group that has
24:53clinical symptoms but they're high all over.
24:55And of course there are
24:56some that have only high.
24:57When the children get older, only in
24:59sort of childhood life they develop it.
25:01It's a small group actually,
25:03but the important thing is that.
25:07These are so tied,
25:08So to say that this is the unique effect
25:11of this episode when they're far be above
25:14clinical levels and others makes no sense.
25:17It is because it's not like an infection.
25:19It's not like a COVID infection where
25:21you can say that during pregnancy
25:23because you don't have continuous COVID,
25:25well, not the infection probably
25:27over 10 years is different.
25:29I think it doesn't work.
25:30We've done it with homelessness and then we
25:32have a set where people experience only.
25:34Short time and then find housing again
25:36and if you have very detailed data I
25:38think you can do that with poverty.
25:40But people who are really below the
25:43poverty line will have been mostly
25:45in a tough spot a year or two later
25:49or a year or two before.
25:51So indeed that was consistent.
25:53So there is this carry over effects,
25:55there is these.
25:57Perhaps there's a biological
25:59rapid development post natally
26:00there are sensitive peers,
26:02there's good ideas.
26:02I think we might be able to do that.
26:04I'll show you later something with
26:06the thyroid hormones where we managed
26:07to do that with sensitive peers.
26:08I think with social adversities
26:11we cannot do that.
26:12So if you judge this study
26:15against public health relevance,
26:17give me a two out of five because I
26:20think the carry sort of the sensitive
26:23period effects which I marketed as.
26:25Don't convince me myself,
26:28and I hope I don't convince you.
26:31It's interesting,
26:32but I don't think it should guide.
26:35It did for a while,
26:36influenced me that I thought, you know,
26:37I have to put more of my research
26:40time into very early depression.
26:41I think that's still valid,
26:43but I'm not so sure that we need
26:46imaging research to show that.
26:48I'll show you because it's very
26:49popular now to do imaging and poverty.
26:52I'll show you a bit of that result
26:54and then an angle I tried to take
26:56and I'm trying to hear your thoughts
26:58or at least look at you whether it
27:00might convince you what we did there.
27:02So household income has been
27:05associated with brain morphology.
27:07We had this data prospectively from it.
27:10Life again that's a sort of marketing trick.
27:13So we show you that we did that and.
27:16I was interested in two things, the timing.
27:18So is it different if it's prenatal or later?
27:21And I was also interested if it's
27:23different in minority majority
27:24and I'll come to that why I'm so
27:27interested in that in a minute.
27:28So if we have 2000 children against
27:30imaging at 10 years poverty defined as
27:33national low income threshold in the
27:35Netherlands, that's nicely defined.
27:36So you get different analyses, you can do it.
27:40Never low income and ever low income.
27:42Note that we have repeatedly assessed income,
27:44So what people? You can just simply do it.
27:46Have you ever in any period been
27:49poor and we can do that chronic
27:51or for example in pregnancy only.
27:53And what you see is just the
27:57distribution which made it for me,
27:59made this a very complicated distribution
28:03because in the Netherlands and you
28:07see a very similar pattern in the US,
28:09it's just not immigrant.
28:12Or non western,
28:14it's just classified as white and non white.
28:17You would see a very similar pattern
28:20that those that are poor are very often
28:24from here from a non western background.
28:26So there is a racial ethnic patterning
28:29of poverty in the Netherlands.
28:31There's a racial ethnic pattern
28:34of poverty in the in America.
28:38So you see that of the four hundreds,
28:40quite a few that were poor,
28:41so 20% were poor at one time a
28:45majority would have been from
28:47long Western and then we have,
28:49you can see the numbers 100 people
28:50that were poor in pregnancy,
28:52100 and 200 that were poor at any one time.
28:54So you can see the breakdown
28:56of these numbers.
28:59Here is so how it looks at truth.
29:02You can see all the different
29:03results that you know.
29:04If you analyze, you get,
29:05even if you take this broad approach of
29:07total brain volume and Gray volume and
29:09then the typical hippocampus, amygdala.
29:11If you do this mix of global and
29:14to specific areas, researchers,
29:16regions of interest, you see with
29:19these many poverty categorizations,
29:21you see all these patterns and then you
29:24can look where there's significance.
29:27And honestly, you could find,
29:29that's why I had it in red,
29:31some association between the
29:32amygdala volume and poverty.
29:34And if you look at it carefully,
29:35this is the reference group.
29:37Never. Then you see what's this?
29:38This is the low income childhood only.
29:40There seems to be no effect.
29:41But if you're chronically poor,
29:43if you're chronic poor, or if you're.
29:47So ever low income is cost,
29:48a combination,
29:48but it's really by low income
29:50and pregnancy or chronic force.
29:51So it really seems to be, if anything,
29:53the pregnancy that might drive it,
29:56the amygdala.
29:57But I can truthfully tell you that this
29:59does not survive multiple testing.
30:01So there would be,
30:03if anything overall in the total group,
30:05no real.
30:11Convincing or strong consistent effect?
30:13Not on the global measures for sure
30:15and on these regions of interest.
30:16Well, if you find it somewhere
30:18just borderline significant,
30:20you should probably discount it.
30:23However, we had very good data
30:25from Child IQ that certainly the
30:28pregnancy was very different in
30:30minority groups and majority groups,
30:33so we had reason from that
30:35paper to stratify a sample in.
30:38Let me call it Western or Nonwestern.
30:40That's the Dutch, Dutch language in America.
30:42Western on Western is not
30:43really cool anymore,
30:44So I'd rather should say
30:48it's not immigrants,
30:49it's people whose ancestors were born
30:52in probably not high income countries
30:55and came to the Netherlands for colonial
30:57history reasons or work reasons,
30:59and the Dutch and the Dutch and European.
31:04Community on the other hand,
31:07and why do I think that's a very
31:09important difference in poverty?
31:11Not only did we have some prior results,
31:13but also we know that if you're
31:15financially strained and you
31:16have a network in the country,
31:18that's a different thing if your
31:20family lives there than if you come
31:22as an immigrant from the Cape Verin
31:24Islands to work in the harbour.
31:25If you're then out of job then you're really,
31:28it really is tough.
31:30So that's why we stratified
31:32for these groups and then.
31:34We see if we do that and
31:36this is only the Dutch,
31:37we actually all of a sudden saw
31:39very broad effects on cerebral and
31:42other broad parameters of the brain.
31:44So taking out this big group of
31:48non Dutch ancestry participants,
31:50let's call it that way, not Dutch ancestry,
31:54but taking them out shoulders all of
31:56a sudden we had a strong effect in.
32:00The overall brain volume.
32:02But what was perhaps more interesting
32:04in the numb Dutch.
32:05So this was the Dutch.
32:06This is the numb Dutch.
32:08We didn't see any global parameters,
32:09but we see very consistently
32:14the effects of in pregnancy or chronic,
32:17which also means in pregnancy
32:18and later on as well.
32:20If you pull that group to sort of
32:22ever in pregnancy, we get a very.
32:24Very significant effect on consistent because
32:27it's significant in both of the subgroupings.
32:29If you pull it, it gets very
32:31significant in effect on the amygdala.
32:33So we get a very different pattern.
32:35So we get a much more stress related
32:38grain poverty pattern in the non Dutch
32:41ancestry group and a very global effect.
32:44It's very hard to think what that means.
32:47Does that valid?
32:47I can tell you I immediately
32:49looked at the ABCD data.
32:50Does it also fall apart in similar
32:52patterning and again of course that would be.
32:54Would be none, white and white,
32:56probably what you could do.
32:58And it's interesting,
32:59we saw the same similar different complicated
33:01pattern for the behavioral outcomes,
33:02but not so much for the brain.
33:04So there is some reason to think that if
33:07poverty comes with different stresses,
33:10it could have a different
33:11meaning for the brain.
33:12We see that very clearly for
33:14the behaviour also in ABCD,
33:15but not for the brain.
33:17And I haven't looked at the amygdala.
33:19As they make that and actually in the
33:21Dutch this really predicted school
33:22performance so it was meaningful.
33:24So if you summarize early in life poverty
33:27and pre adolescent brain morphology,
33:30there is an association but they really
33:32differ from majority and minority groups.
33:34And was all the caveats that you
33:36hate this subtyping of majority and
33:37minority that's up to you to dislike it.
33:40I think that is some evidence that we do it.
33:42In America,
33:43I would say we should do it to
33:44some extent because poverty and
33:46discrimination go together,
33:47which makes a very different terrible mix.
33:50In the Netherlands it is also discrimination
33:52and stress of surviving financial strength.
33:54So there is some reason to do that
33:56and this I think what that reflects,
33:59I'll be very,
33:59very careful to speculate about that.
34:01I think it could also be
34:03genetically associated,
34:04we don't know,
34:06but in the in the minority groups
34:09as I call them here or.
34:11If you want the real nice terminology,
34:13I think the exact terminology the
34:16non Dutch and ancestry group.
34:18I think it is likely stressed
34:19by discrimination and because we
34:21have that variable in the model I
34:23can tell you pull it in and it's
34:26substantially weakened the association.
34:28So it's not a real mediation analysis,
34:30but there is.
34:31About 30% of the association and
34:33that was a very crude measure of
34:35discrimination disappeared once
34:36we put that in the model.
34:37So I think there's real reason to
34:39think that could be different and we
34:41have to think more carefully about
34:43our neurodevelopmental measures.
34:44I'll do the thyroid and then another,
34:47I'll do that quickly.
34:49I've presented that for many times.
34:51So what would that be in if you give me.
34:56My sort of scale,
34:57my own scales is rating your
34:59own work. But let's do it critical.
35:01I think we're still only at a three out
35:03of five of public health because to
35:06think that poverty measures mentioned
35:09poverty matters for the brain,
35:11I don't think we need too much brain imaging.
35:14But, you know, to carefully dissect the
35:17effects of minority groups again, Well,
35:20really, do we need the imaging for that?
35:24Although I think that it has
35:26lasting effects on child brains,
35:27that it what is it affects is
35:30associated with child brains may be
35:32very different How it associates
35:34with the brain where you come from,
35:37It's at least makes us think so give
35:39me a three out of five perhaps we want
35:42to go to four out of five, don't we?
35:44So here's thyroid where I think
35:46we can manage thyroid.
35:48Old work of mine and it was one recent
35:51update which is I think quite spectacular.
35:53Thyroid of the brain,
35:55So note that the maternal thyroid brain.
35:57So maternal thyroid hormones are very
35:59important for the brain development.
36:01Animal work has shown convincingly
36:03that actually it's fascinating
36:04that the neurogenesis and it's
36:07particularly the neuro neuromigration
36:08which actually comes from around,
36:10you know the central,
36:12the ventricles and then the neurons
36:15migrate out to your cortex.
36:17Obviously that's where they are
36:19in our brains that is guided by.
36:22Thyroid hormones that happens in
36:26early embryonic life when the embryo
36:30is reliant on the maternal thyroid.
36:35So much of the neurodevelopment.
36:37So the nature is seem sort of
36:39very pasimonious.
36:40It has only 10 mechanisms
36:41and what does you know,
36:43vitamin D does something and
36:45serotonin do something very different
36:46in the fetal life they're much
36:49more neurodevelopmental than in
36:50us where they have.
36:51Very new endocrine function,
36:52but they have very new
36:54developmental functions.
36:55All these systems and in particular
36:57thyroid in pregnancy and only in
37:00week 14 does then the fetus produce
37:03its own thyroid and only by week 20,
37:06so sometime later does it produce
37:08somewhat sufficient levels and takes over.
37:11So in that time the mother
37:13supplies the thyroid.
37:15In that time many women who have a low
37:17thyroid function actually become a bit
37:19hyperthyroid because they need more.
37:21There's very good graphs.
37:22I haven't got them with me because
37:24I do a short version of this talk.
37:25But trust me, there's very good work,
37:27mostly animal work,
37:28very consistent that we need the
37:30thyroid levels for a brain development.
37:33And what we showed in the very early
37:35publications, nearly ten years ago now,
37:37is that if you take the total sample,
37:39and this is the measure of s s.
37:44SRS,
37:44the social responsiveness of good
37:46population trait measure of autism,
37:48you see that the people with
37:49good levels of the mothers,
37:51the offspring of mothers with normal
37:53levels of thyroid hormone have much
37:56lower levels than those that have subautimal.
37:59And this is subclinical,
38:00we're not talking about a clinical,
38:01this is untreated hypothics,
38:04thyroxinemia, you can do severe,
38:05you can do less severe,
38:06but it's all subclinical,
38:08so it's just low levels.
38:10Of thyroid hormone in the mother
38:12and you saw that association which
38:14we showed and then we move on
38:16to more recent work,
38:19a first Lancet endocrinology paper where
38:22we showed that if we take the levels
38:26continuous now FT-4 that's the thyroid.
38:29So this means more thyroid,
38:30this means less thyroid.
38:32We showed actually a a
38:34curvilinear association with IQ.
38:36It's most robust in the
38:38low thyroid levels here.
38:40And then this is a quite
38:41a wide confident role,
38:41but you see some significant down decline.
38:44So there is a tightly regulated level
38:46and that's where most mothers are.
38:48If you see the distribution of hormones,
38:50it would be just most people are in
38:52this space, some are in the low,
38:55some are in the high.
38:56And we saw a very robust relation with IQ.
39:00And later we've replicated this in 2-3
39:02other cohorts where I must be honest,
39:04this ups this low levels of and the
39:08relation to low IQ is extremely robust.
39:10This in other cars looks more like this,
39:12going sort of much more flat.
39:13There's not such a decline,
39:15but there is a very robust association
39:19between prenatal thyroid hormones and I Q.
39:23And then we move to another
39:26hormone thyroid parameter.
39:28So be careful.
39:28This is now thyroid,
39:29thyroid stimulating hormone.
39:31This means that now you'd beware that
39:33higher levels of the stimulating hormones
39:36means lower levels of thyroid hormone.
39:38It's flipped.
39:39I think you have to be a doctor or
39:41an endocrinologist or physiologist
39:42who immediately get it.
39:44But trust me whereas we had easy more
39:47hormone is we thought better but this
39:49is not the case because it gets worse here.
39:51But this is more hormones.
39:53This is less hormones and
39:55less hormones means lower IQ.
39:56Here it is two different things modeled.
39:59It's not IQ, it's Gray matter.
40:01So it's not a brain parameter.
40:02And you see this is essentially flipped.
40:06So this means.
40:07Less hormones.
40:08This would mean more hormones,
40:10but I'm presenting at the stimulating
40:12axis hormone and what you see is
40:15the same similar inverted U-shaped
40:18curve tightly regulated on all
40:21levels of the thyroid between the
40:24brain and the between the brain.
40:27And the thyroid hormone and
40:29it's highly significant.
40:31So it's 2000 children at age 10,
40:33it's their prenatal,
40:35their mothers in the early
40:37mostly around week 10 to 14,
40:39it's their thyroid hormone levels.
40:41And this has led to some guidelines
40:43and discussion and guidelines.
40:44Should we measure more thyroid
40:46hormones in women that have no
40:48symptoms and no history of and there
40:51have been trials based on this work
40:52which it have to have been largely
40:54negative or very small effects.
40:55So they're sort of.
40:57Equivocal trials have been done,
40:59so we don't know,
41:00but there is some evidence that it is
41:02a very important parameter to regulate.
41:04And now comes the recent work.
41:07I don't know,
41:08I don't have a date when that
41:09was published 2 years or so ago,
41:11which is very fascinating.
41:13We did that.
41:14We just realized this data because
41:16we had the idea what actually we
41:19included the women at different ages.
41:21So we can model always in about 200 women.
41:25The curve essentially continuously
41:28moving the curve with a time interaction
41:32variable across the inclusion period.
41:34So the first women came to
41:37generation out to be included and
41:39we took the blood at week seven.
41:41The latest that we included
41:43were week eighteen.
41:44Note these are not the same women.
41:46This is the first blood assessment we
41:49had where we did the thyroid hormones.
41:52So what we modeled it as a
41:55sort of continuous model,
41:56but then cut it for the doing
41:59essentially the intercept for
42:01the different week 7 to 18.
42:03And what we see is that this curvy linear
42:08pattern which is very remarked up to age,
42:11then sort of disappears at the
42:13end of this inclusion period.
42:15And this was still 200 women on average.
42:20Time period per week and what this
42:22shows you I think is convincingly
42:25a sensitive period because it
42:27is in the same study measured at
42:31different time points specifically.
42:33And why is that so credible?
42:34Because the reviews they sort
42:36of were extremely excited.
42:37I've never got anything in that sort
42:39of Lancet like paper that easily
42:41because as in chronologist said,
42:43I've done animal work and I showed
42:45by week 15 the child produces
42:47on thyroid and thus the mother.
42:49Thyroid is just not informative anymore.
42:53So while I marketed as a final we
42:56got their sensitive period study,
43:00the reviewer toned it down to saying
43:02it's really showing that the measure
43:04is not informative At age 15.
43:06It may still influence the brain,
43:08but you're measuring the wrong parameter.
43:10So this is getting closer to the
43:12sensitive period. Holy Grail.
43:14That's all these Doha epinologists
43:15want to get to, but even there,
43:17a very careful reviewer can tell.
43:19Tell you you're not there.
43:21It just means that from week 14 onwards
43:23you're measuring the wrong person,
43:25essentially like having the wrong informant.
43:27But what does it tell you?
43:30It tells you what.
43:32I think that this is valid because how
43:36could it's if you then have the right
43:39measure and you find what you expected,
43:41perhaps that's sort of a circumvential say.
43:44I don't think it proves causality.
43:46But it's getting better that
43:48this is quite credible.
43:49So I do think in all honesty there
43:51is a true curvilinear relationship
43:54between thyroid hormone and the brain.
43:57I do think given the biology
43:59it is likely to be causal.
44:00Whether that's amenable for
44:02intervention is another study.
44:04I've got the wrong slides.
44:06I was going to ask you so
44:08transition to new results.
44:10I missed my transition slide because I
44:12pulled it up yesterday night after the.
44:15Chemical exposure for the colleague.
44:17I hope she's there on the zoom.
44:21Does anybody know what trans fatty acids are?
44:25Take a sip of coffee while you tell me.
44:31Is that forgotten? You're not bisphosals and
44:38organophosphates?
44:39Which are about what?
44:40Does anybody still know
44:41what trans fatty acids are?
44:47I'll tell you, trans fatty assets
44:51in the Netherlands were a big scandal,
44:55a public health scandal of big proportions.
44:57Why? Because in the 1990s eighties,
45:01I don't know, to that time,
45:03your grandparents wouldn't have eaten butter.
45:06Which they would have
45:06lived in these countries.
45:07And then there comes the introduction of
45:09margarines which is better for public health?
45:11Okay, it's better for your fat because
45:14it's unsaturated and saturated.
45:16Fatty Acids in butter versus margarine.
45:18And these are people that eat,
45:19you know bread butter and I don't know cheese
45:24on twice a day.
45:27And so the problem was that these
45:30margarines where fatty acids,
45:32but they also had trans fatty acids.
45:35Meaning these are industrial fatty
45:37acids which come with the production
45:39of fat and essentially if you produce
45:41fat and if you have sort of a if you
45:44fry your French fries and you have very
45:46poor fat in one of these, I don't know,
45:49I don't want to point at any cart here,
45:51but if you have very poor,
45:52you get trans fatty acids in them.
45:54And those were really,
45:56it turned out to be terrible
45:58for cardiovascular health,
46:00actually so bad so that the whole
46:02benefit of eating margarine was offset
46:05by the effect of trans fatty assets.
46:08It was a real scandal in the 1990s.
46:11OK, It's sort of forgotten.
46:13And I don't know,
46:14and I don't know anything much about America.
46:15My work is mostly from Europe.
46:17So there what happened is
46:19there were countries.
46:20It's already interesting to see what
46:21happens in countries once that's,
46:23you know.
46:23Detected that Hans very as a mouse
46:26models and humans and observational and
46:28is really bad and sort of kills you
46:31the there's countries that forbid it.
46:33Okay Denmark said gone two years
46:35and we phase it out of production.
46:38It's easy you can just make
46:39a bit more expensive oils.
46:42The Dutch you might not know them
46:45are sort of compromising country
46:47so they say to the industry
46:49you know it would be good.
46:52If you reduced it in your
46:53products in the next five years,
46:55we do that on a voluntary basis and we
46:57will also do a bit of shaming and naming.
46:59So there is some pressure.
47:01That's the Dutch approach.
47:02Now you would laugh about the Dutch,
47:04but they do get it done so slowly, by slowly.
47:07Uni Lever,
47:08whom you know from the Dove products,
47:10is a real big you know you Lever
47:12is the modern maker in that time.
47:15I don't know if they still do it and
47:17they phased it out, which leaves.
47:19Other products like um,
47:22cheap bakery products where it's still
47:24used because they couldn't care less.
47:26You know, that's the fringe market
47:28and they couldn't care.
47:28It's cheap to do that.
47:30So what we found, um, so here is.
47:34If you want to know trans and sis
47:36fatty acids, so this is the big difference.
47:38Industrial fatty acids like trans would have
47:41the hydrogen here instead of like this.
47:43Wow, you think that's the difference?
47:45That's it.
47:45Yes, that's the difference.
47:47That's it.
47:47And where are they?
47:48They're found in fried foods,
47:50commercial bakers and processions.
47:51But what happens in the statcha pros?
47:53There was not a law to stop them,
47:55but really,
47:56in the early 2000s in this country, this.
47:59And the changes in the Netherlands went down
48:02in vegetable oils and fat in those years.
48:05The production went down dramatically.
48:08So without,
48:09I'm not saying that's the best approach,
48:10but both in the Netherlands
48:12and Denmark and other countries
48:14in Europe, they reduce these fatty acids.
48:17And why is that? Why am I telling
48:19you all this in an imaging study?
48:21I'll tell you why.
48:23It's really fascinating.
48:24I saw this once and I thought these are.
48:28The inclusion years of the Generation R study
48:32we included from 2003 to actually to 2007,
48:40we included in exactly the years when
48:43trans fatty acids disappeared in the in
48:49the food industry in the Netherlands.
48:52That means we can look at.
48:55The blood levels of women who came
48:58at different times in those years,
49:00and we did. And you know,
49:01you'd think you'd see the same exact curve,
49:03But you know,
49:04we saw a 10/10 a quarter of decline,
49:09which for anything in
49:10biology is quite dramatic.
49:12So in 2000,
49:14the people who included in 2005 had only 3/4
49:17of the levels of those included in 2002.
49:21So it indeed related to a.
49:25Reduction in the blood of a women and
49:29I don't know if they knew and they
49:30didn't change their eating behavior,
49:32they just ate the same bread
49:34and French fries as before,
49:36but they got less of this.
49:38Which means if we can relate
49:41this to a child outcome,
49:43we have something which we call
49:46instrumental barrel approach
49:47because it is a policy change.
49:49That is related to biology
49:51in the blood of people.
49:53And so we published that last
49:55year after sort of after I had
49:56what I don't know how I came.
49:57I come from a Baker's family to be honest,
49:59I was, I don't know,
50:00reading this in the sort of Baker thing
50:02digest and I was quite fascinated
50:04and I thought, yes, you can do that.
50:06And so we did the.
50:08Trans fatty acids in the blood and
50:10then we showed and you know you
50:12this is a very bad slides taken
50:14directly from the publication.
50:16But you can see a highly,
50:17highly significant association of
50:20trans fatty acids with fetal head growth.
50:24And this is true head growth.
50:25This is the change from fetal head
50:27size from one point to the other.
50:29It's not just growth and you say birth
50:31weight is a measure of fetal growth.
50:33This is really fetal growth
50:34as it is a change from.
50:362nd to 3rd trimester,
50:38there was no effect when the head is very,
50:41very small, but when it expands,
50:42when it gets big,
50:43that's where all the growth is.
50:44And that second to third end of
50:47trimester and 6000 children.
50:48So that's a good inclusion.
50:50We see super significant associations and
50:52then we can actually do the same trick.
50:54We can only do the, the,
50:56the, the TFA measures.
50:58We can do that with the high
51:00and see very clear patterns,
51:01but we can see that this calendar time.
51:05There is an association of
51:08calendar time with fetal growth,
51:11meaning that in the course of
51:15that studies the the heads of
51:18the children became a tiny bit.
51:21I must admit it's a tiny bit,
51:22but fetal measures in 6000 are very precise,
51:25bigger and we think and we can show that
51:28was an instrumental viral approach,
51:30which is a different sort
51:32of statistical technique.
51:33We can show that the association is
51:37driven by the reduction and the policy
51:41change and that is something I've been
51:44working 20 years towards and never got done.
51:46So that we show that policy translates
51:50into biology and sad thing is we didn't
51:54get it to behaviour, so bigger heads.
51:55And I'm not really much related to behavior
51:57and then it becomes very messy and noisy.
51:59But you know,
52:00the journal loved it that it was.
52:02And why does it have
52:04clinical health relevance?
52:06Well, first of all, it does.
52:07This is causality,
52:08not only policy,
52:10it is quite a causal approach,
52:11but really interesting.
52:12If you look at the production
52:14of East Europe and South Asia,
52:15that's the Indian region and
52:17the East European region where
52:19there's nobody cares about this,
52:21The levels are still shockingly high.
52:24So I think it's still relevant,
52:25although for us it's a historic
52:28study to be honest.
52:30And do I do one more or should I do
52:32for questions. This is a good ending.
52:34So I could do a physical activity
52:36in the brain,
52:37but
52:41good, then I'll wrap up.
52:42So I'll leave away that
52:44there is an association with.
52:46Brain change that I should
52:47I just do one slide?
52:49No, I don't do one slide.
52:50It doesn't work.
52:50I do the, I do the IT doesn't work.
52:53I just tell you it is we show that
52:56would have been the last one.
52:57I sort of did too much
52:58fatty acids carried away.
53:02I was going to show you that
53:04we can show that physical
53:06activity is related not just to
53:08brain size and brain volume,
53:10but it is related to the volume
53:13change over adolescence.
53:15Which is quite a bit more and that's
53:17essentially an answer to the,
53:18you know we need bigger studies
53:20or we need studies of change.
53:22We've now got the first studies of change.
53:24If you want to show,
53:24just show the result,
53:28it's total physical activity really,
53:30not just the the also quite a
53:33bit of the hippocampus grows or.
53:37Grows a bit faster if you do physical
53:39activity and it's interesting
53:41because it's consistent across
53:43parent and child reported physical
53:45activity reports Okay I'll wrap up.
53:47So the dominant force in research is
53:51the is the know you're imaging a lot
53:54in autism and a DHDI would challenge
53:57or like to discuss with people who say
54:00it's made a change in our clinical
54:02treatment or in our public health
54:04understanding of autism and brain I think.
54:06It did a lot for understanding the disease.
54:09I'm not so sure it did a lot for how we
54:12treat disease, which is a big difference.
54:15I would say The effect sizes are often
54:17small and often correlational and not causal.
54:20There's a real problem which I didn't show
54:22you, but we've struggled with that a lot.
54:24Can we reproduce imaging results?
54:28Anybody who might talk today was talking
54:30about the heterogeneity of populations.
54:31That's the same.
54:32And I showed you that was the
54:35minority majority is one example.
54:37I think we have to, and that was my
54:40first talk this morning was Kerim.
54:41I think he's there in the back row.
54:43We should really go to developmental
54:45approaches and longitudinal trajectories.
54:46I think that's the only way forward.
54:49I fell short of showing you
54:50that was the physical activity,
54:52but I think that's what matters.
54:55I'd like to wrap up,
54:56it's not a diagnostic or prognostic tool.
54:58It does have some public health relevance,
55:01but I would say.
55:03Occasionally,
55:03and sometimes even sort of coincidentally,
55:07but it does as many other fancy techniques.
55:11These are the students that helped
55:13Ryan Mitzler.
55:14I want to mention him because he does
55:15much of my imaging in the Netherlands
55:17and students who did these papers,
55:19and of course the participants.
55:20Thank you very much.
55:27Thank you so much, honey.
55:28I will just say that we do have
55:30time after the presentation.
55:31So if anyone would like to stay in the
55:33room and continue the conversation,
55:35we're free until 2:30.
55:36And but any burning
55:38questions for Doctor Tamar
55:49that was that was pretty interesting to me.
55:51And I just wonder your thoughts
55:53about how far do you go on
55:56restrictive public policy to?
56:00Get the good for for young children
56:04who can't protect themselves.
56:06So for instance you know it's good to
56:10keep lead away from babies and and young
56:13children but when you start talking
56:15you know dietary and cultural things,
56:18just your thoughts.
56:19How how far do you go with this?
56:20Do you do you, you know say that's it,
56:23fruits and vegetables and
56:25Mediterranean diet for everyone or.
56:28That's an interesting one.
56:29So that's sort of the whole public health
56:32school of Harvard debates that every day
56:36and seriously does if it's good to to
56:38zoom in on an example because otherwise
56:40I'm going to give a sort of overreaching,
56:43I would be struggling.
56:44That's a little evening thing debate.
56:46If you take the dietary example,
56:48I am in favor. Of restricting
56:52soda and sweet drinks in schools.
56:55We have seen now that
56:57that is really so obesity,
57:00making so much diabetes down the road.
57:02I think we should go there.
57:05Many of the others like no sweets,
57:07which are also, you know,
57:08sugar is bad, but I would be very,
57:10very hesitant.
57:12I think the best way to do it
57:15is to think carefully with the
57:17schools should sell them but.
57:19To forbid them, perhaps a sugar tax.
57:21But other than that I think
57:23very little is evidence based.
57:24So much of these things are not causal.
57:26We changes.
57:27You know look at the history of
57:29the Harvard schools of department
57:31of petition advice for diet.
57:32You know that's a funny changing thing.
57:34You know nuts and that and oils
57:36and meat and alcohol.
57:38Just look at the alcohol.
57:39You know 20 years ago I was
57:41taught in Rotterdam alcohol is
57:42better than any concentration.
57:44You come to Harvard and they say no,
57:45but of alcohol is very good.
57:47Now they have to sort of change that,
57:49but it took them 15 years to
57:52really come to a conclusion there.
57:54So that encouraging of your
57:57glass of red wine
57:58is now gone. You know, you have to
58:00have it was a bad conscience tonight,
58:03but I think still think,
58:04still think you should.
58:05So I'd be very, very restrictive,
58:07very, very cautious,
58:09but I wouldn't shy away from a few measures,
58:12very, very cautious.
58:13But sometimes I think soda, we got it.
58:16Sugars, we've got it.
58:17So restrict the sugars in a
58:19creative way and for not forbid
58:21but tax it and don't have it. Yeah,
58:25just really quickly. Perhaps relatedly,
58:28you know when you talk about your
58:30trans fatty acid decline over time,
58:31I was thinking about PER and
58:33polyfluoroloco substances.
58:34You know, these forever chemicals and and.
58:36You know, what we've seen in with
58:38some pilot data there is that
58:40there's a patterning by income level,
58:41a patterning by income level per country.
58:44I'm just wondering the decline in
58:46trans fatty acids that you described,
58:48was there a patterning by SES?
58:50Did you observe A steeper? No.
58:52We see much less of that patterning
58:53in the US than in the US.
58:55In the US, every environmental exposure
58:58is highly socially patterned to an
59:00extent that sometimes escapes me.
59:02I don't quite know, you know,
59:03why are they having so much more?
59:05And then I hear they have different
59:06hair products and this and that.
59:08It's very hard for me to understand.
59:10In the Netherlands, for example,
59:11I'll tell you, organo phosphates,
59:14which is pesticides,
59:15were higher in the high SES
59:18because they ate more fruit.
59:19So in in the US we looked at the
59:22same thing and lo and behold,
59:24organo phosphates are lower in high SES.
59:29I don't understand the US
59:30enough to understand why that is
59:32such a ubiquitous pattern.
59:34In the Netherlands,
59:35it's much less so people live.
59:39I don't know as many reasons.
59:42I don't quite understand that.
59:43So in the Netherlands?
59:44No, not always,
59:45although some of some of the chemicals,
59:48yes, very much so,
59:49but not as not as dramatic as here.
59:53I think you're trans fatty policy
59:55example is one of the most profound
59:58statements in support of integrating
01:00:00the research and policy says
01:00:01Thank you so much for sharing.
01:00:03I definitely want to find out more
01:00:04about that and track that
01:00:06and try to replicate that.
01:00:07My question for you is building
01:00:09on all that you've done,
01:00:10especially in the area of policy,
01:00:12what do you see next?
01:00:13What do you see is the next area that
01:00:16you could be pursuing building out?
01:00:19What does policy mean?
01:00:20Because when I when I looked at the data,
01:00:22I thought back to let.
01:00:24Because in the United States there's
01:00:26definitely an association with with
01:00:28low income and and lead in your
01:00:30pipes and in your drinking water.
01:00:32So what is on your horizon
01:00:34next in the space of of poverty
01:00:38and research and policy?
01:00:42Yeah there's there's in my school and in
01:00:45my world thinking too 2 lines of research.
01:00:48One is always which we have.
01:00:50You know can you dissect. Why poverty?
01:00:56What makes poverty relate to behavioral
01:00:58and new developmental cognitive school
01:01:00achievement problems or do you just
01:01:02think you know it's money. That's it.
01:01:08You know I am it's I think those
01:01:12two are are totally separate.
01:01:14I think we should fight LED
01:01:17and environmental things really
01:01:18more better and full force.
01:01:22Lead is just I can't yeah we've
01:01:23discussed that I don't need to say that
01:01:24no none of us can believe that it's
01:01:26still around as a public house hazard.
01:01:28It should be gone.
01:01:29It's just not acceptable.
01:01:31At the same time I think make very
01:01:34clear that as long as we have these
01:01:37substantial poverty gradients that
01:01:41that is a policy taxing and that's you
01:01:44know beyond me to to do much about the
01:01:46but it's clearly something that has
01:01:48to be addressed because I think with.
01:01:50Addressing LED, you will not
01:01:53substantially address the poverty
01:01:54inequality in this country as
01:01:56much as I think it's important,
01:01:58but it's completely different thing.
01:02:00And you see that in
01:02:04we all know that, you know homelessness.
01:02:08I'm yeah, the the the extent of homelessness
01:02:11in Boston and other areas is just so
01:02:14dramatic and such a health hazard.
01:02:17I don't know why that's not addressed.
01:02:18I really fail to see that could
01:02:20easily be addressed. Wonderful.
01:02:22Well, just in the interest
01:02:23of everyone's time,
01:02:23if anyone would like to stay on,
01:02:25please do wait in the room.
01:02:26We can continue this conversation.
01:02:28And but just please do join
01:02:29me again in thanking Dr.
01:02:31Kmar for his presentation.
01:02:34Yeah. Sorry to talk so
01:02:36long and see you later.