YC-SCAN2 May 2025 Webinar

Transcript

• 00:01Hi. Good afternoon, everyone. We've
• 00:03had a bit of technical
• 00:04difficulties, but
• 00:06I'm, pleased to
• 00:08to introduce,
• 00:09Matt Alberg who is,
• 00:12based at the University of
• 00:13Vermont. Matt, we are really
• 00:15excited to hear from you.
• 00:17Matt has a background as
• 00:18a clinical psychologist and
• 00:20and is a neuroscientist
• 00:22where he did his training
• 00:24both,
• 00:25both his neuroscience and and
• 00:27psychology training at
• 00:29University of Vermont,
• 00:31and we are delighted to
• 00:32have him.
• 00:33So please welcome me in
• 00:36oh, wait. Well, please join
• 00:37me in welcoming Matt.
• 00:39And, hopefully, Matt, you can
• 00:41you have all everything working.
• 00:44Yeah. Yeah. I'll, I'll I'll
• 00:46see if I can start
• 00:47sharing,
• 00:48my Slack set here.
• 00:51Let's see.
• 00:58Alright.
• 01:00Great. Yeah. We can see
• 01:01your screen.
• 01:04Great.
• 01:06Perfect. Perfect.
• 01:07And you can hear me
• 01:08okay?
• 01:12Yes. Great. Yeah. We had
• 01:14some, construction going on earlier,
• 01:16so I'm I'm hoping that
• 01:18they
• 01:19that they hold off.
• 01:21Well,
• 01:22thank you so much. I
• 01:23really appreciate,
• 01:25the warm welcome and and
• 01:27and the invitation,
• 01:29doctor D'Souza. This is
• 01:31exciting to
• 01:32to tell tell you, in
• 01:34in your center a little
• 01:35bit about some of the
• 01:36work, that we've been
• 01:38up to, at the University
• 01:40of Vermont.
• 01:43So I I guess I
• 01:44will just, jump right into
• 01:46it if that's alright. Yeah.
• 01:48Great. So, I I don't
• 01:50I don't have any,
• 01:52conflicts to disclose.
• 01:57And just kind of a
• 01:58general
• 01:59outline for today.
• 02:01So, you know, we'll we'll
• 02:02maybe start just very briefly
• 02:05talking a little bit about
• 02:06the evolving cannabis landscape.
• 02:09I'm sure it's something we're
• 02:10all very aware of, but
• 02:11just providing a little context,
• 02:14for some of the work
• 02:15we've done. And then we'll
• 02:17jump right into some of
• 02:18the,
• 02:18longitudinal imaging findings,
• 02:21that that we've been involved
• 02:22in, here at UVM.
• 02:25And and these analyses have
• 02:27primarily involved,
• 02:28the imaging cohort that I'll
• 02:30touch on a little bit
• 02:31later in the presentation.
• 02:33And then, hopefully, we'll have
• 02:34time. I can I can
• 02:36share a little bit, some
• 02:37of the ongoing projects,
• 02:39that are going on in
• 02:40the lab? So here here
• 02:42is, my disclosure slide then.
• 02:46And
• 02:47sorry about that. I'm not
• 02:49sure what happened.
• 02:51And
• 02:52here is the outline.
• 02:55So, as I was saying,
• 02:56if we have time, hopefully,
• 02:58we will. We'll touch on
• 02:59some of the work, that's
• 03:00that's ongoing in the lab
• 03:02involving a, b, c, d
• 03:03data and imaging.
• 03:05And then, we'll wrap up
• 03:07with some,
• 03:09concluding remarks and and, some
• 03:11some future directions.
• 03:14But before we before we
• 03:15get into it, you know,
• 03:17I I still feel like
• 03:18I'm I'm relatively
• 03:19new,
• 03:21in this in this space.
• 03:24And, you know, going back
• 03:25a few years ago when
• 03:26we first started this line
• 03:28of research,
• 03:30I I think I was
• 03:31very naive,
• 03:32with
• 03:34just how this,
• 03:35would be covered and kind
• 03:36of what the reaction would
• 03:38be.
• 03:40Some of our early work,
• 03:41ended up in a number
• 03:42of,
• 03:43you know, very very social
• 03:45media streams.
• 03:47I made the horrible mistake
• 03:48of going down the wormhole
• 03:50of,
• 03:52like,
• 03:52looking at some of the
• 03:53comments,
• 03:55and, you know, this is
• 03:57just a small sampling
• 04:00of some of the comments
• 04:01out of hundreds and hundreds.
• 04:04But everything from why didn't
• 04:06we look at alcohol
• 04:08and tobacco,
• 04:10products,
• 04:12were we sponsored by the
• 04:14alcohol industry or or the
• 04:16tobacco industry?
• 04:19We have people claiming to
• 04:20have IQs of a hundred
• 04:21and fifty four, and they've
• 04:22been using cannabis their whole
• 04:23life. Clearly, this is, you
• 04:25know, this research is flawed.
• 04:27And then also some some
• 04:28actually, you know, fairly
• 04:30accurate,
• 04:31comments, even though we tried
• 04:33our best in some of
• 04:34our, you know,
• 04:35initial papers to make clear
• 04:37the limitations.
• 04:39You know,
• 04:40we're working with inherently observational
• 04:42data.
• 04:44But anyway,
• 04:46all of this to say,
• 04:48just kind of an additional,
• 04:51disclaimer here that,
• 04:55just because, you know, we're
• 04:57we're we're looking at
• 04:59potential effects of cannabis use,
• 05:02especially during adolescence and early
• 05:04adulthood,
• 05:05we're certainly
• 05:07not suggesting
• 05:08that,
• 05:09you know, other substances
• 05:12are safe.
• 05:14There seems to be, you
• 05:15know, quite a literature with
• 05:16respect to alcohol
• 05:18and cigarette smoking.
• 05:21And, you know, we were
• 05:24kind of reacting to what
• 05:25looked like a relatively sparse
• 05:27literature,
• 05:29certainly with respect to
• 05:30longitudinal imaging.
• 05:33So anyway yeah. Just because
• 05:35we're we're, you know, we're
• 05:36focused on this cannabis question,
• 05:39it doesn't mean we are,
• 05:41you know, condoning
• 05:44other forms of
• 05:46adolescent substance use.
• 05:50And
• 05:50I've already touched on this
• 05:52point. You know, we're we're
• 05:53working with inherently observational
• 05:55data. It's clearly not ethical
• 05:56to do some sort of
• 05:57randomized
• 05:59study,
• 06:00in in in
• 06:02humans.
• 06:03So
• 06:04that being said, you know,
• 06:05we we are in the
• 06:07process of
• 06:08of leveraging causal inference methods
• 06:10to to perhaps get a
• 06:11little bit closer
• 06:14to that causal question.
• 06:16But we are,
• 06:18you know, very aware of
• 06:20just the limitations around observational
• 06:22data.
• 06:23And then also, you know,
• 06:24we have no political agenda
• 06:26here.
• 06:28Really, I saw I'm a
• 06:29clinical psychologist,
• 06:31and I do a lot
• 06:32with respect to mood and
• 06:33anxiety symptomatology,
• 06:36and in Vermont, cannabis use,
• 06:39comes up frequently,
• 06:40with clients.
• 06:42And so the idea here
• 06:43is really just to provide
• 06:45individuals with with data. Right?
• 06:47So that they can make
• 06:49their own informed decisions
• 06:51and to kind of empower
• 06:52them,
• 06:54to to make, you know,
• 06:56the best health related decisions
• 06:57they can.
• 06:59So anyway,
• 07:01I I always feel compelled
• 07:03to kind of,
• 07:04put that put that out
• 07:06there
• 07:07as an additional kind of
• 07:09disclaimer.
• 07:12So
• 07:13back to our outline here.
• 07:14So just a very quick
• 07:16look at at, kind of
• 07:19what has happened with respect
• 07:20to the cannabis landscape in
• 07:22the last couple of decades.
• 07:23I'm sure we're all very
• 07:24aware
• 07:26of these factors.
• 07:28So this is a map
• 07:30now of,
• 07:32where,
• 07:33recreational cannabis use
• 07:36is now legal.
• 07:38So,
• 07:39right now, I believe as
• 07:40of twenty twenty five, there
• 07:41are twenty four states, so
• 07:43almost half,
• 07:44the US
• 07:45where recreational
• 07:47cannabis is legal.
• 07:51And this slide here just
• 07:52gives you a feel for
• 07:53the pace at which this
• 07:55has happened over the last
• 07:57twelve, thirteen years.
• 07:59So, you know, very quickly.
• 08:02And,
• 08:03so here we see past
• 08:04year use,
• 08:06for number of substances.
• 08:08And you can see,
• 08:11that
• 08:13really,
• 08:14relative to,
• 08:16other substances, there has been
• 08:18this kind of steady uptick
• 08:20in
• 08:21past year use,
• 08:23in individuals
• 08:24age twelve and older.
• 08:28This is just a more
• 08:29granular breakdown of cannabis use
• 08:32by age group. And really
• 08:33just, you know, with the
• 08:35exception
• 08:36of,
• 08:37the twelve to seventeen group
• 08:38here shown in green,
• 08:40there really has been significant
• 08:43increases in cannabis use in
• 08:45the last couple decades,
• 08:47you know, very significantly for
• 08:50for, young adults.
• 08:53And then
• 08:54also,
• 08:57coupled with these factors,
• 08:59these these changes in in
• 09:01perception of risk.
• 09:03So currently, over fifty percent
• 09:04of individuals
• 09:05believe
• 09:06using cannabis one to two
• 09:08times a week
• 09:09poses little to no health
• 09:11risk.
• 09:12And you can see, again,
• 09:13relative to other substances here,
• 09:16there has been a market
• 09:17shift with respect to cannabis
• 09:20in in the last decade
• 09:22or so.
• 09:25And then another slide, I'm
• 09:26sure everybody here is is
• 09:28all too familiar with,
• 09:30but just the increased THC
• 09:32concentration,
• 09:34in in seed cannabis,
• 09:36over the last several decades
• 09:39and really strong evidence indicating
• 09:41that high potency cannabis, particularly
• 09:44when used
• 09:46daily or near daily,
• 09:49presents another whole magnitude of
• 09:51risk.
• 09:52In a recent Lancet Psychiatry
• 09:56paper, high potency cannabis in
• 09:58particular,
• 09:59when used on a daily
• 10:01basis was
• 10:02related to a nine fold
• 10:03increase
• 10:04in psychotic disorders.
• 10:06And someone worked this out
• 10:07to be, you know, roughly
• 10:08the equivalent,
• 10:10to the risk of lung
• 10:12cancer, in individuals smoking thirty
• 10:14cigarettes a day. So that
• 10:16helps you just kind of,
• 10:18put that in context.
• 10:20And and another kind of
• 10:22sobering,
• 10:23detail here that that that
• 10:25study that I just mentioned,
• 10:26the, De Forte et al,
• 10:30high potency was defined as
• 10:31as, you know,
• 10:33THC content of of greater
• 10:35than ten percent.
• 10:37And I think,
• 10:38certainly within,
• 10:39the Burlington area, any of
• 10:41the local dispensaries, I I
• 10:43think you'd probably be hard
• 10:44pressed
• 10:46to find
• 10:49any cannabis,
• 10:51you know, with THC content
• 10:53as low as, you know,
• 10:56ten percent low teens.
• 10:59So
• 11:00even that has evolved significantly
• 11:03in the last several years.
• 11:07So
• 11:08all this to say, you
• 11:09know, again, these are really
• 11:11kind of, I think, some
• 11:12of the motivating factors that
• 11:14went into some of our
• 11:16initial work here.
• 11:17So we have kind of
• 11:18this these changes in in
• 11:21risk perception.
• 11:22We have this kind of
• 11:23rapidly changing legal status,
• 11:26in many states,
• 11:27as as far as cannabis
• 11:29use is concerned.
• 11:30We see what looks like
• 11:32this kind of widespread increase
• 11:33in cannabis use across most
• 11:35age groups.
• 11:37And then lastly, this,
• 11:39this kind of
• 11:40increase in THC concentration in
• 11:43cannabis and in cannabis products.
• 11:47So that was a little
• 11:49bit of a whirlwind there,
• 11:49but that was, you know,
• 11:50again, just kind of the
• 11:51the the motivating context,
• 11:53for for some of this
• 11:54initial work here.
• 11:56So now I'll dive right
• 11:57into,
• 11:59our imaging work. I always
• 12:01have this slide in here.
• 12:02So so,
• 12:03work coming out of, doctor
• 12:05Yasmin Hurd's lab really,
• 12:07I think was the primary
• 12:10scientific motivation for us.
• 12:14Our lab had looked at,
• 12:15the Miller et al paper,
• 12:18and there was very solid
• 12:20experimental
• 12:21evidence
• 12:22for adolescent
• 12:24THC exposure
• 12:25resulting in accelerated dendritic
• 12:28pruning,
• 12:29primarily in pyramidal prefrontal neurons.
• 12:33And there are some very
• 12:34compelling graphics,
• 12:36where, you know, you look
• 12:37at the exposed
• 12:39animals
• 12:40versus the controls,
• 12:42especially a few weeks out,
• 12:43and you can visually see,
• 12:46just how,
• 12:47the dendritic arbors are sparser
• 12:50in in the exposed animals
• 12:52relative to the controls.
• 12:55So all of this suggesting
• 12:57that THC exposure during adolescence,
• 13:00resulting in accelerated pruning and
• 13:02possibly removing connections that may
• 13:04otherwise have persisted into
• 13:06adulthood.
• 13:08And this certainly seems to
• 13:10fit with,
• 13:11just what we think we
• 13:13know about the endocannabinoid
• 13:15system,
• 13:16and its involvement in adolescence
• 13:19and early adulthood in synaptic
• 13:21pruning.
• 13:24And so
• 13:25at the time, just a
• 13:27few years ago,
• 13:28you know, after looking through
• 13:30this paper
• 13:31and also realizing
• 13:33that, just
• 13:34looking at what was out
• 13:36there in the literature,
• 13:37there seem to be
• 13:39very, very few,
• 13:41longitudinal
• 13:42imaging studies looking at cannabis
• 13:44use,
• 13:45and,
• 13:46longitudinal brain change,
• 13:50and
• 13:51especially,
• 13:53studies where, you know, there
• 13:54was a sizable
• 13:56sample.
• 13:57What was out there,
• 13:59the the few studies that
• 14:00were out there,
• 14:02they were relatively small
• 14:04samples, particularly
• 14:05in kind of the Merrick
• 14:06et al
• 14:08era,
• 14:10where, you know, there is
• 14:12a
• 14:13a, certainly a premium for
• 14:14for larger samples.
• 14:17And then this other piece
• 14:18too that, so some of
• 14:20the work I had been
• 14:21involved in, even going back
• 14:23to, my time in graduate
• 14:24school,
• 14:26it was really looking at
• 14:27longitudinal
• 14:28cortical,
• 14:29change,
• 14:31looking at trajectories of age
• 14:33related,
• 14:34cerebral cortical development.
• 14:37And this idea that part
• 14:39of what drives,
• 14:41MRI assessed
• 14:43age related cortical thinning
• 14:47is dendritic and synaptic pruning.
• 14:49So,
• 14:50this idea that
• 14:52this particular method might be
• 14:55might be sensitive to some
• 14:57of the the neurobiological
• 14:58changes
• 15:00that, you know, that that
• 15:02may be associated with cannabis
• 15:03use,
• 15:04in in humans during adolescence.
• 15:08So
• 15:10this might be
• 15:12a review for for most
• 15:13of you, but just just
• 15:14quickly,
• 15:16so we'll be talking a
• 15:17lot about,
• 15:18the measure of cortical thickness.
• 15:21So there are surface based
• 15:22imaging pipelines,
• 15:24that we, run t one
• 15:26weighted data through.
• 15:27When we're talking about cortical
• 15:29thickness, we're talking about the
• 15:30distance between the, essentially, the
• 15:32white matter surface and what's
• 15:35commonly referred to as the
• 15:36pial surface,
• 15:38are just kind of, you
• 15:39know, the,
• 15:41outer exterior
• 15:42of the cortical ribbon.
• 15:45And,
• 15:47there is,
• 15:49compelling evidence that, you know,
• 15:51there there are significant age
• 15:53related changes,
• 15:55when we look at cortical
• 15:56thickness over time.
• 15:58And, again,
• 16:00one of the drivers behind
• 16:01this MRI assessed thinning is
• 16:03believed to be
• 16:05dendritic and synaptic pruning,
• 16:08but also,
• 16:10you know, myelination particularly of
• 16:12of
• 16:13lower cortical,
• 16:14layers.
• 16:16And just to note too,
• 16:18it's maybe getting a little
• 16:19bit off,
• 16:21into the weeds here. But,
• 16:22you know, early on
• 16:24trajectories of cortical,
• 16:26thinning,
• 16:27across childhood and adolescence suggested
• 16:29kind of this inverted quadratic
• 16:32shape. Over the years,
• 16:34and with the advent
• 16:36of more rigorous QC techniques,
• 16:39and larger samples,
• 16:41it really looks like as
• 16:42of age
• 16:44five or so, there is
• 16:45this,
• 16:46kind of first order,
• 16:48linear thinning that's taking place
• 16:50throughout most
• 16:51of the the cortex.
• 16:56So,
• 16:58at about the time I
• 16:59was starting my my my
• 17:00k award,
• 17:02I had been working with
• 17:04the image and dataset.
• 17:06Hugh Garavan, who has just
• 17:08been, an incredible
• 17:10mentor and and part of
• 17:11the the mentorship team on
• 17:13my k,
• 17:15you know, he had a
• 17:17long standing interest
• 17:19in cannabis use and brain
• 17:20development.
• 17:22And so, again, because we
• 17:24had
• 17:25access
• 17:26to the imaging dataset,
• 17:29Hugh is one of the,
• 17:31or was one of the
• 17:32site PIs,
• 17:33on the imaging study.
• 17:35And this is a study
• 17:37of over two thousand two
• 17:38hundred youths.
• 17:40They were recruited at fourteen
• 17:41years of age through schools.
• 17:44There,
• 17:45there's a whole host of
• 17:47substance use and mental health
• 17:48measures, cognitive measures, as well
• 17:50as genetics,
• 17:52that, you know, that, that
• 17:53were collected.
• 17:55And then also MRI scans
• 17:58at fourteen,
• 18:00nineteen, and twenty two.
• 18:02So, just a very
• 18:04impressive dataset
• 18:06and actually
• 18:08kind of
• 18:09perfect
• 18:10for this this,
• 18:12this cannabis use cortical development
• 18:15question.
• 18:16And I at the time,
• 18:17I had just kind of
• 18:18finished running
• 18:20all of the image and
• 18:21data,
• 18:22so the fourteen age fourteen
• 18:24and age nineteen data
• 18:26through Civette, which is, for
• 18:28those of you who haven't
• 18:29heard of Civette, it's it's
• 18:31it's,
• 18:31it's analogous to to free
• 18:33surfer.
• 18:34It was developed up at
• 18:36McGill by, Alan Evans and
• 18:38his colleagues.
• 18:40But I had run all
• 18:42the t one weighted data,
• 18:44through Sivette.
• 18:46And we'd already started to
• 18:47look at,
• 18:48age related change,
• 18:51in these data.
• 18:56And so to address the
• 18:58this cannabis question,
• 19:00we,
• 19:02we looked at individuals
• 19:05who, first of all, just
• 19:06had
• 19:07QC'd,
• 19:09structural data,
• 19:10at at baseline and follow-up.
• 19:14But then importantly,
• 19:15we limited
• 19:17participants to those
• 19:18who,
• 19:20who who had reported being
• 19:22cannabis naive,
• 19:24at study baseline. So, again,
• 19:25at fourteen years of age.
• 19:28And then, again, as long
• 19:29as they had,
• 19:31follow-up imaging data,
• 19:33they got included into the
• 19:34study. So, overall, we had
• 19:35almost eight hundred subjects,
• 19:38four hundred and fifty females,
• 19:39three hundred and forty nine
• 19:41males.
• 19:43Again, the data were run
• 19:44were run through Civette,
• 19:46and I used a,
• 19:49MATLAB toolbox developed by Keith
• 19:52Worsley,
• 19:53called Servstat
• 19:54to analyze,
• 19:56the data.
• 19:57And in particular, we we
• 19:59ran vertex
• 20:00level linear mixed effects models.
• 20:03We for these analyses, we
• 20:05ran a whole host of
• 20:07sensitivity
• 20:08analyses. But kind of to
• 20:09start, we were controlling
• 20:11for,
• 20:12total brain volume,
• 20:14sex at birth,
• 20:16handedness,
• 20:17the actual scanner,
• 20:19and also alcohol consumption.
• 20:21We ended up, you know,
• 20:23in in subsequent,
• 20:25analyses controlling for tobacco use
• 20:27and some other things
• 20:28and findings largely held.
• 20:31But, again, that was kind
• 20:32of the, those were the
• 20:34initial
• 20:35variables here. This shows you
• 20:36a distribution
• 20:38of cannabis use,
• 20:40at five year follow-up. So,
• 20:42again, when they were roughly
• 20:43nineteen years of age. And
• 20:44just a reminder, everybody,
• 20:47at baseline
• 20:48was reporting
• 20:50to be cannabis naive.
• 20:52So we can kind of
• 20:53think of this
• 20:54distribution here
• 20:56as the change,
• 20:58in in kind of,
• 20:59lifetime use total use,
• 21:02between fourteen and nineteen years
• 21:04of age.
• 21:05So
• 21:06little over half, can, you
• 21:08know, remain cannabis naive at
• 21:09follow-up, but,
• 21:11roughly the other half went
• 21:13on
• 21:13to, some degree of use.
• 21:19And so these are just
• 21:20cross sectional results. So this
• 21:22is just looking at,
• 21:24essentially that change in use,
• 21:28relating it to,
• 21:30age nineteen cortical thickness.
• 21:33And we were seeing these
• 21:34significant
• 21:36negative associations.
• 21:38So thinner cortices
• 21:41in these highlighted regions
• 21:45relating to
• 21:47more cannabis use between fourteen
• 21:49and nineteen years of age.
• 21:52And this was really the
• 21:53the primary analysis. So this
• 21:54was a linear mixed effects
• 21:56model,
• 21:58so leveraging all the available
• 22:00data.
• 22:01And so really what we
• 22:02saw here was
• 22:05the the greater degree of
• 22:07use between fourteen and nineteen
• 22:09years of age,
• 22:11the,
• 22:12the the greater the rate
• 22:13of cortical thinning in these
• 22:15regions.
• 22:18And because everybody had complete
• 22:21data kind of by design,
• 22:22we were able to,
• 22:24actually look and see
• 22:26the extent to which
• 22:28baseline
• 22:29brain structure or, you know,
• 22:31in this case, specifically,
• 22:33cortical thickness baseline cortical thickness,
• 22:36the extent to which that
• 22:37predicted,
• 22:39subsequent cannabis use.
• 22:41And, really, what we found
• 22:42here, even when we were
• 22:43very lax with our, kind
• 22:45of uncorrected significance threshold,
• 22:49was not much.
• 22:50And the few kind of
• 22:52very weak trends,
• 22:54they were in areas that
• 22:56were not overlapping
• 22:57with where we saw this
• 22:59time by cannabis interaction that
• 23:01I just showed.
• 23:03So really here just suggesting
• 23:05that
• 23:06did not seem to be,
• 23:08kind of preexisting
• 23:10differences
• 23:11in brain structure,
• 23:13driving
• 23:15subsequent use.
• 23:17And when we looked at
• 23:18so these are just the
• 23:19the,
• 23:20predicted,
• 23:22values from the models themselves.
• 23:24A very similar story here,
• 23:26where you see, what looks
• 23:29like,
• 23:32you know,
• 23:33kind of
• 23:35across levels of
• 23:37of cannabis use between fourteen
• 23:39and nineteen years of age.
• 23:41Really not much in the
• 23:42way of differences at baseline,
• 23:44but then what's looking like
• 23:45this, you know, dose dependent
• 23:48association
• 23:49at follow-up.
• 23:51It's just another peak region
• 23:52here I'm showing.
• 23:54These are so
• 23:57this measure here is symmetrized
• 23:59percent change. So these are
• 24:00unadjusted values. So these are
• 24:02aren't aren't predicted values. These
• 24:04are just,
• 24:05raw values here,
• 24:07and I I think there's
• 24:09a very similar pattern here.
• 24:11I like, you have to
• 24:12squint too hard to see
• 24:14something that looks like a
• 24:15fairly dose dependent association. So
• 24:18so, again, the the more,
• 24:21cannabis used between fourteen and
• 24:23nineteen years of age, the
• 24:24greater the rate of thinning
• 24:26in these largely prefrontal areas.
• 24:31And,
• 24:31we also found some evidence
• 24:34for
• 24:37thinning in these regions. So
• 24:39this cannabis related thinning
• 24:41mediating the association
• 24:43between
• 24:44cannabis use from fourteen to
• 24:46nineteen years of age and,
• 24:49attentional
• 24:49impulsivity
• 24:51at at the follow-up, at
• 24:52five year follow-up.
• 24:55So more cannabis use relating
• 24:57to a greater degree of
• 24:58thinning in turn relating
• 25:03to higher degrees of attentional
• 25:05impulsivity
• 25:06at five year follow-up.
• 25:09This is kind of the
• 25:11this is a busy slide.
• 25:12I'm sorry.
• 25:15So this is really kind
• 25:16of the the the take
• 25:17home,
• 25:18slide
• 25:19from our,
• 25:21initial work here.
• 25:22So,
• 25:23in the top left in
• 25:24green,
• 25:25that is just the the,
• 25:28the age effect.
• 25:30So in this particular sample,
• 25:33this is,
• 25:35these are the areas that
• 25:36are showing significant age related
• 25:39thinning.
• 25:40And the top in blue
• 25:42is the,
• 25:44the time by cannabis interaction
• 25:46that I just described.
• 25:48I've
• 25:48relaxed the threshold here just
• 25:50for visualization
• 25:51purposes, but that's what's shown
• 25:53in blue.
• 25:55And then thank you to,
• 25:56Doctor. D'Souza
• 25:59and colleagues.
• 26:00We were able to obtain,
• 26:03a pet map, albeit, you
• 26:04know, a pet map derived,
• 26:07from a
• 26:08a separate sample,
• 26:10of adults. Clearly would not
• 26:11be ethical to do pet
• 26:13scanning,
• 26:14on on
• 26:16developing adolescents here.
• 26:18But, we have a PET
• 26:19map of,
• 26:21CB one receptor availability.
• 26:23And so really just kind
• 26:25of what we're showing here
• 26:26is that these maps are
• 26:27correlated,
• 26:30and kind of the two
• 26:31bottom figures here. So on
• 26:32the bottom left,
• 26:34where you kind of see
• 26:35the green and
• 26:37and blue,
• 26:38overlaid,
• 26:40it's suggesting that, you know,
• 26:42where we're seeing this cannabis
• 26:44time interaction,
• 26:46these are largely
• 26:48areas
• 26:49that are already undergoing the
• 26:51greatest degree of age related
• 26:52change.
• 26:53So,
• 26:54you know, we might think
• 26:55of these areas as perhaps
• 26:57being more plastic, right,
• 26:59in in the context of
• 27:01development.
• 27:03And then,
• 27:04the bottom
• 27:05right where we kind of
• 27:06see
• 27:08the blue, red, and magenta,
• 27:10this is
• 27:11showing
• 27:12that we're, again, we're we're
• 27:14seeing this time by cannabis
• 27:16interaction.
• 27:17On average,
• 27:19these areas tend to be
• 27:20higher in CB one receptor
• 27:22availability.
• 27:24So, just kind of in
• 27:26again,
• 27:27all circumstantial evidence,
• 27:29but, you know,
• 27:31a potentially
• 27:32kind of compelling
• 27:37message coalescing out of this.
• 27:42And so
• 27:43at at at at UVM,
• 27:47I'm incredible
• 27:48I mean, incredibly fortunate to
• 27:50have, so there's so there's
• 27:51a t thirty two program
• 27:52here in complex systems.
• 27:54I I have the wonderful,
• 27:59fortune of
• 28:00being able to to interact
• 28:02with incredibly
• 28:04smart people.
• 28:06At the time, there was
• 28:07a
• 28:08postdoc, Max Owens,
• 28:10and there, is,
• 28:12there's a faculty member, Nick
• 28:13Allgier, who
• 28:14actually, was, yeah, running the
• 28:16t thirty two.
• 28:19They looked at this initial
• 28:21work,
• 28:22and
• 28:23they applied Bayesian,
• 28:25causal network modeling.
• 28:27And and the basic idea
• 28:29being here that, you know,
• 28:30you can,
• 28:31you can
• 28:33essentially have your variables represent
• 28:35nodes,
• 28:36and you can have directed
• 28:38edges that
• 28:39are informed by conditional probabilities,
• 28:43and then the actual structure
• 28:44of the network itself,
• 28:47informed by conditional dependencies in
• 28:49the variables.
• 28:51And so this is kind
• 28:52of a data driven approach,
• 28:54that at least conceptually,
• 28:57might bring us a little
• 28:58bit closer to,
• 29:00some
• 29:01some some causal evidence.
• 29:04And
• 29:06so,
• 29:07what what they did, they
• 29:09started off with just kind
• 29:10of a
• 29:12a limited set of variables
• 29:14here.
• 29:15They ran
• 29:16ten thousand
• 29:17bootstrap
• 29:18resamplings,
• 29:21And what they found was,
• 29:22I think, over ninety six
• 29:24percent
• 29:25of
• 29:27the models
• 29:28were indicating that cannabis use
• 29:31was having a causal impact
• 29:33on
• 29:35the the thinning between fourteen
• 29:37and nineteen years of age.
• 29:39So,
• 29:41again,
• 29:42nothing that is,
• 29:44definitive,
• 29:45but perhaps,
• 29:46another
• 29:47kind of piece of of
• 29:49circumstantial evidence here. And and
• 29:51what's interesting, even when,
• 29:53we provided these algorithms
• 29:55with other potential
• 29:57confounders,
• 29:58like measures of childhood trauma,
• 30:02like SES,
• 30:03various SES measures, ADHD
• 30:06symptomatology,
• 30:09in this version,
• 30:11co occurring substance use was
• 30:13also put in there. So
• 30:14tobacco use, alcohol use. We
• 30:16kind of threw even more
• 30:18stuff at it that could
• 30:19potentially be related. I didn't
• 30:20highlight it, but we also
• 30:21had a few, polygenic risk
• 30:24scores.
• 30:26So for, you know, polygenic
• 30:28risk for lifetime cannabis use,
• 30:30polygenic risk for,
• 30:32cannabis use disorder.
• 30:35And, again, overwhelmingly,
• 30:38the the,
• 30:39algorithm indicated cannabis use
• 30:42impacting,
• 30:45prefrontal thickness.
• 30:48So,
• 30:50few years ago,
• 30:53age twenty two data became,
• 30:55newly available,
• 30:57in Imogen.
• 30:59And so we immediately had
• 31:00a, you know, a a
• 31:01couple questions.
• 31:04So one of the questions
• 31:06was, you know, can we
• 31:07still see,
• 31:10any
• 31:11markers of adolescent cannabis use
• 31:14in early adulthood? So, again,
• 31:15around twenty two years of
• 31:16age,
• 31:18or
• 31:19is that is that gone?
• 31:22And then another question too
• 31:24was,
• 31:25you know, we figured that
• 31:26we would also be able
• 31:28to look at individuals
• 31:29who had initiated cannabis use,
• 31:33between nineteen and twenty two
• 31:34years of age.
• 31:36And we could look to
• 31:37see,
• 31:38what the what the
• 31:41pattern of longitudinal
• 31:42brain change was,
• 31:45with early adult initiation.
• 31:48Did it look similar to
• 31:49what we were seeing,
• 31:50with our fourteen to nineteen
• 31:52findings?
• 31:53So starting off with kind
• 31:55of that that first question,
• 31:57we identified,
• 31:59people,
• 32:00with,
• 32:01in the age twenty two
• 32:03data,
• 32:05where we there, was QC
• 32:07imaging data for them.
• 32:10And we were able to
• 32:11look and see
• 32:12the extent to which adolescent
• 32:14cannabis use was relating to
• 32:16cortical thickness at age twenty
• 32:18two. And what's
• 32:19this is maybe a a
• 32:21nerdy point, but
• 32:23I'm impressed with it
• 32:24anyways.
• 32:26These data also,
• 32:27when we got h twenty
• 32:29two,
• 32:30data, from Imogen,
• 32:33because there was a newer
• 32:34version of Civette that was
• 32:36available and we were recommended
• 32:38that we should run everything
• 32:39through the newest version.
• 32:42All of the data got
• 32:43reprocessed
• 32:45and re q c'd.
• 32:47We had I I'm not
• 32:48showing it here, but we
• 32:49had nearly identical
• 32:51findings
• 32:53from what I just showed.
• 32:56And then also, you know,
• 32:57interestingly here, you know, we're
• 32:59seeing areas that showed up
• 33:01in that time by cannabis
• 33:03interaction.
• 33:04So what we're what we're
• 33:06actually seeing here is that
• 33:07the more adolescent
• 33:09cannabis use reported, so between
• 33:11fourteen and nineteen years of
• 33:12age, the thinner
• 33:14these,
• 33:15these cortices
• 33:16shown in blue were at
• 33:18age twenty two.
• 33:19And what's interesting is that
• 33:21even when we controlled
• 33:24for past year use,
• 33:26at age twenty two,
• 33:28it had no effect.
• 33:29There,
• 33:30so it really was
• 33:32suggest suggesting that there was
• 33:34something about adolescent
• 33:36use
• 33:37that seemed to have this
• 33:39more enduring mark,
• 33:41on cerebral cortical structure. And,
• 33:44again, in areas where we
• 33:45saw that time,
• 33:47by cannabis interaction from fourteen
• 33:49to nineteen.
• 33:53And then yeah. So so
• 33:55then looking so, you know,
• 33:57looking to address the question
• 33:58of what, you know, what
• 34:00brain changes were associated with
• 34:03young adult cannabis initiation.
• 34:06And so we had our
• 34:08groups here. So,
• 34:10roughly
• 34:10seven
• 34:12so seven hundred four participants,
• 34:14almost two thousand MRIs.
• 34:16So we had
• 34:19adolescent initiators,
• 34:21who started using between fourteen
• 34:23and nineteen.
• 34:24And as just an aside
• 34:26here,
• 34:27and we were hoping that
• 34:28there would be enough
• 34:31adolescent initiators
• 34:33that had used
• 34:35between fourteen and nineteen and
• 34:37then not used,
• 34:39from nineteen to twenty two.
• 34:41And, unfortunately,
• 34:43I mean, there I forget
• 34:44the exact number, but I
• 34:45think it was on the
• 34:46it was literally it was
• 34:47less than twenty
• 34:49where,
• 34:51we saw that pattern.
• 34:52So, you know,
• 34:53the adolescent
• 34:54initiators
• 34:56were also kind of continued
• 34:57users. So they they kind
• 34:59of continued right through age
• 35:01twenty two.
• 35:03The young adult initiators
• 35:05were cannabis naive at baseline,
• 35:07but then,
• 35:09when in end,
• 35:10at at age nineteen, but
• 35:12went on to initiate between
• 35:14nineteen and twenty two. And
• 35:15then,
• 35:16the cannabis naive,
• 35:18group, those individuals that remained
• 35:21cannabis naive throughout
• 35:23the the entire study.
• 35:26And so this is just
• 35:27kind of a demographics table.
• 35:29So, really, we only saw
• 35:31kind of significant differences with
• 35:33respect to,
• 35:35age at the nine year
• 35:37follow-up. So when,
• 35:38participants
• 35:39were
• 35:41roughly twenty two years of
• 35:42age, and it was, you
• 35:43know, a relatively small difference,
• 35:46and and, you know, we
• 35:47were aware of the,
• 35:49just kind of the the
• 35:50significant
• 35:51differences with respect to sex.
• 35:54We did our we still
• 35:56controlled for these,
• 35:57covariates.
• 35:58But,
• 35:59for the other measures,
• 36:01we did not find significant
• 36:03differences.
• 36:05And just kind of maybe
• 36:06a little added face validity
• 36:08here,
• 36:09we had a, an amazing
• 36:11postdoc,
• 36:11Renata Cupertino,
• 36:13here at UVM,
• 36:14and she was able to
• 36:16relate these groups
• 36:17to genetic liability,
• 36:19for cannabis use disorder. She
• 36:21did something similar again using
• 36:23the polygenic risk score for
• 36:25lifetime
• 36:26cannabis use.
• 36:28We kind of saw a
• 36:29similar pattern here.
• 36:30Again, what almost looks like
• 36:32kind of this
• 36:34stepwise
• 36:35relationship.
• 36:36So
• 36:38the the adolescent initiators
• 36:40having a greater genetic liability
• 36:42for cannabis use disorder and
• 36:44lifetime cannabis use relative to
• 36:46the cannabis naive group.
• 36:51So leveraging all of the
• 36:53data, again, almost, I think,
• 36:54two thousand MRIs.
• 36:58This is what we saw
• 36:59for a a group by
• 37:00time interaction.
• 37:03We and this was kind
• 37:04of the pattern throughout,
• 37:06the significant regions.
• 37:08So,
• 37:09really, what we saw in
• 37:11it it it this interaction,
• 37:14seemed to be driven by
• 37:15this kind of accelerated
• 37:17thinning
• 37:18in the adolescent,
• 37:20initiators.
• 37:22And really, in these group
• 37:23level analyses,
• 37:25we were not really able
• 37:27to
• 37:28distinguish
• 37:29between
• 37:30the young adult initiators and
• 37:32the cannabis naive group. So
• 37:33kind of statistically
• 37:35across these regions,
• 37:36these groups were looking the
• 37:38same. It was really the
• 37:39adolescent
• 37:40initiators,
• 37:42that were were differing. And
• 37:43in particular, it was kind
• 37:45of this fourteen to nineteen
• 37:47window where,
• 37:48when they were initiating, they
• 37:50were showing this,
• 37:52accelerated
• 37:53thinning in largely prefrontal
• 37:55areas.
• 37:57And, you know, because we
• 37:59had observed these
• 38:01differences
• 38:02with respect to genetic liability,
• 38:05we we also looked to
• 38:06see if if genetic risk
• 38:08was
• 38:09qualifying this pattern.
• 38:11And what we saw was
• 38:13that it it it was
• 38:14not.
• 38:15And we had tried this,
• 38:17with a number of
• 38:18of polygenic risk scores,
• 38:21and kind of
• 38:22across analysis.
• 38:25There was no evidence
• 38:27of genetic liability
• 38:29qualifying
• 38:30the the pattern I just
• 38:32described.
• 38:35And interestingly,
• 38:37it it may be also
• 38:38kind of dovetailing a little
• 38:39bit with,
• 38:40the the previous mediation model
• 38:42I showed,
• 38:44suggesting that some of this,
• 38:47thinning was relating to increased
• 38:49attentional impulsivity.
• 38:51What we saw here was
• 38:55change in cannabis use from
• 38:56fourteen to nineteen
• 38:58relating to thinning,
• 39:00in the highlighted areas
• 39:02and,
• 39:04that relating to
• 39:06past month use,
• 39:09past month cannabis use at
• 39:10age twenty two.
• 39:12So we found evidence for
• 39:14partial mediation here. And interestingly,
• 39:16we saw
• 39:18this for a couple other
• 39:19substances,
• 39:21ecstasy
• 39:22and cocaine.
• 39:24So
• 39:25again, maybe,
• 39:26you know, lending some support
• 39:28to this idea that,
• 39:32that this accelerated thinning might
• 39:34be relating to,
• 39:36aspects of impulsivity.
• 39:39And I think also there
• 39:40was a fairly recent paper,
• 39:43from
• 39:45Doctor. Heard's lab,
• 39:47which combined
• 39:49rodent in human data. And
• 39:51I think that there was
• 39:52a, you know, a similar
• 39:54sort
• 39:55of finding here,
• 39:56so some potential convergence
• 39:59across labs.
• 40:01So at this point, you
• 40:02know, with the follow-up analyses,
• 40:04we really had not seen
• 40:06much evidence
• 40:08of the cannabis naive group
• 40:11looking any different,
• 40:14relatives to the young adult
• 40:15initiators.
• 40:17So, we decided
• 40:19to kind of replicate
• 40:22the analysis we did,
• 40:25from, you know, age fourteen
• 40:27to nineteen,
• 40:28the the kind of analysis
• 40:30and results that I started
• 40:31off with.
• 40:33So, you know, we can
• 40:33think of this as maybe
• 40:34so this is no longer
• 40:36kind of a a a
• 40:37group based analysis. Now we're
• 40:38we're kind of treating,
• 40:41cannabis use
• 40:43or that change in cannabis
• 40:44use as a continuous
• 40:46quantitative measure,
• 40:47and we're looking to see
• 40:49if that relates to brain
• 40:51change,
• 40:52from
• 40:53nineteen to twenty two.
• 40:55This is, the histogram
• 40:57of change in in cannabis
• 40:59use from nineteen to twenty
• 41:01two.
• 41:03And you can see a
• 41:04little more than happy group
• 41:06remains cannabis
• 41:08naive at age twenty two,
• 41:10but, a little less than
• 41:12half go on to,
• 41:13initiate cannabis use.
• 41:16And,
• 41:17and similar to what we
• 41:18saw
• 41:19with the fourteen to nineteen
• 41:21analyses,
• 41:23there was really no evidence
• 41:25of of
• 41:26age nineteen cortical thickness,
• 41:31predicting
• 41:32subsequent
• 41:33use or subsequent initiation.
• 41:35So we didn't see any
• 41:36significant associations
• 41:38there. But when we ran
• 41:39it as a linear mixed
• 41:40effects model,
• 41:42we saw that,
• 41:44age related change,
• 41:46was was qualified
• 41:48by
• 41:50the degree of cannabis use
• 41:52from nineteen to twenty two
• 41:53years of age. But interestingly,
• 41:55right, this is looking like
• 41:57a very different constellation of
• 41:59brain regions.
• 42:01In fact, you know, there's
• 42:02really no overlap
• 42:03relative to what we saw
• 42:05from fourteen to nineteen.
• 42:08And this
• 42:09so then we, you know,
• 42:10we one interesting kind of,
• 42:13observation here is that
• 42:17this so this,
• 42:19figure,
• 42:20we're only limiting it to
• 42:22regions,
• 42:23where there is significant
• 42:25age related change
• 42:27in their respective
• 42:29developmental windows.
• 42:31So all this to say,
• 42:34there seems to be this
• 42:35this,
• 42:36you know, at least from
• 42:37our our our early work
• 42:39here that,
• 42:41cannabis
• 42:42seems you know, cannabis use
• 42:44seems to be associated with
• 42:45brain change in areas that
• 42:47are already kind of undergoing,
• 42:49the greatest degree of age
• 42:50related change
• 42:52during a particular developmental window,
• 42:54which is which is interesting.
• 42:57We all in this follow-up
• 42:58work, we also found some
• 43:00evidence,
• 43:00for some mediation effects.
• 43:03Some of the
• 43:05patterns of change we saw
• 43:07in in lateral temporal areas.
• 43:10Those changes
• 43:11mediated
• 43:12the association between cannabis use
• 43:15and,
• 43:16psychotic like experiences
• 43:19and in particular kind of
• 43:21positive
• 43:23symptoms.
• 43:24So, you know, more
• 43:26hallucination,
• 43:27delusion
• 43:29oriented
• 43:30experiences.
• 43:32So just kind of a
• 43:34a broad,
• 43:35set of conclusions here.
• 43:39I think I've I think
• 43:40I've probably already hit these,
• 43:42but,
• 43:43this, you know, this idea
• 43:45that we can still see
• 43:47what looks like this enduring
• 43:48mark of adolescent cannabis use,
• 43:51in the
• 43:52young adult brain.
• 43:54We kind of see these,
• 43:56different patterns of longitudinal
• 43:59brain change relating
• 44:01to, adolescent
• 44:02initiation
• 44:03versus young adult initiation.
• 44:06And,
• 44:07and,
• 44:08and also
• 44:10these two different patterns that
• 44:12we're seeing, you know,
• 44:15relating to adolescent and young
• 44:16adult initiation,
• 44:18they're also relating differently to
• 44:20behavior.
• 44:24So, yeah,
• 44:25real quickly, just some ongoing
• 44:27work with, a, b, c,
• 44:29d, and imaging.
• 44:31So there, there's also longitudinal,
• 44:34diffusion imaging data available,
• 44:36in imaging that has recently
• 44:38been reprocessed
• 44:40in QC'd.
• 44:44Just a reminder of kind
• 44:45of this this map where
• 44:46we see these adolescent effects.
• 44:51And another person on my
• 44:52k mentorship
• 44:53team, Nikos Makris at the
• 44:55Martino Center,
• 44:57we've worked closely together.
• 45:00But
• 45:01in kind of looking at
• 45:02this map,
• 45:04there is kind of just
• 45:05this,
• 45:07qualitative piece where
• 45:10it starts to look a
• 45:11little bit like the connectional
• 45:12topography of the cingulum bundle.
• 45:15And so,
• 45:16what we did was we
• 45:18actually
• 45:19pulled in some of the
• 45:20peak regions from our cortical
• 45:21analyses. We use these as,
• 45:24seeds
• 45:25in exploratory
• 45:27tractography
• 45:28analyses,
• 45:29and we could very, you
• 45:31know, very reliably,
• 45:33get get the cingulum from
• 45:34these,
• 45:35from from these
• 45:38peak areas.
• 45:40And so,
• 45:41we so then leveraging
• 45:43kind of all of the
• 45:44QC longitudinal
• 45:45diffusion data,
• 45:47And this just shows the,
• 45:48you know, the age distribution,
• 45:51of subjects,
• 45:52including the the, repeated scans
• 45:54and image. And,
• 45:56sorry, the the font's a
• 45:58little small.
• 46:02But interestingly,
• 46:03we find that there is,
• 46:06you know, it's, it's, it's
• 46:07kind of almost mirroring the,
• 46:10the,
• 46:11the thickness findings, the cortical
• 46:13thickness findings
• 46:14where we so, you know,
• 46:16typically during this,
• 46:17developmental window, we're seeing kind
• 46:19of significant age related,
• 46:22FA increases,
• 46:23across fiber tracts.
• 46:26Well,
• 46:27in in in the cingulum
• 46:28bundle, what we're seeing is,
• 46:30for those in, you know,
• 46:32essentially at higher levels,
• 46:34of adolescent cannabis use, we're
• 46:36seeing more attenuated
• 46:37age related increases,
• 46:40in fractional anisotropy,
• 46:42particularly in the cingulum.
• 46:44And,
• 46:45and what's even more interesting,
• 46:46right, is, you know, so
• 46:47we've become interested
• 46:49in
• 46:50the extent to which
• 46:52genetic liability
• 46:54may serve to qualify
• 46:56some of these cannabis
• 47:00related
• 47:01findings.
• 47:02And so it's really interesting.
• 47:03We're starting to see evidence
• 47:07across
• 47:09a couple different polygenic
• 47:11risk scores.
• 47:13We're starting to see evidence
• 47:14of kind of almost this
• 47:15three way interaction. It gets
• 47:17a little complicated because these
• 47:18are tensor products
• 47:20in the context,
• 47:21of,
• 47:22GAM models.
• 47:24But we're starting to see
• 47:26evidence of what looks like,
• 47:28you know, essentially three way
• 47:29interactions
• 47:30where,
• 47:32we kind of see
• 47:33the greatest degree
• 47:36of,
• 47:37of,
• 47:39this kind of attenuated
• 47:40age related increase
• 47:43at the at higher levels
• 47:46of genetic risk for,
• 47:48for conditions like depression,
• 47:50and
• 47:51schizophrenia.
• 47:53Schizophrenia actually has a there's
• 47:54a very similar pattern here
• 47:55that we see. So,
• 47:58still kind of we're still
• 47:59piecing some of this this
• 48:01work together, but, we think
• 48:02that there's some very kind
• 48:03of interesting signal here,
• 48:06that we're that we're still
• 48:07kind of working through.
• 48:09And then very I'll try
• 48:10to make this super quick.
• 48:11A, B, C, D,
• 48:13we have a a brilliant,
• 48:15graduate
• 48:16student here,
• 48:18again, through this t thirty
• 48:19two program,
• 48:22Tony Barrows.
• 48:24And,
• 48:25he has really championed,
• 48:28this propensity score matching approach.
• 48:31And and really what we've
• 48:32done here is so in
• 48:34a b c d, we
• 48:36have looked or we've identified,
• 48:38cannabis
• 48:39initiators.
• 48:40So, you know, we've defined
• 48:42a particular
• 48:43window of initiation, and some
• 48:44of this is governed a
• 48:45bit by the numbers,
• 48:47just, you know
• 48:48so there's a little bit
• 48:49of a balancing act here,
• 48:50but we've defined
• 48:52a a kind of window
• 48:53of initiation.
• 48:54So here kind of year
• 48:55two to year six, we've
• 48:57also imposed kind of a
• 48:59threshold of what constitutes,
• 49:01you know, in our mind,
• 49:02maybe considerable
• 49:04cannabis use.
• 49:06Again, partially governed by the
• 49:08numbers here.
• 49:10And then,
• 49:11you know, the idea being
• 49:12that oops.
• 49:14The idea being that with,
• 49:16propensity score matching,
• 49:18we can really leverage the
• 49:20size of the a b
• 49:21c d study
• 49:22sample,
• 49:24and and we can
• 49:26find individuals that have similar,
• 49:29you know, propensity scores
• 49:31relative to the initiators.
• 49:33So kind of, you know,
• 49:34from a statistical standpoint,
• 49:37individuals
• 49:38that,
• 49:39statistically
• 49:41are,
• 49:42just as likely to go
• 49:43on to initiate
• 49:45based on a whole constellation
• 49:49of variables. So we can
• 49:51match kind of in this
• 49:52pre initiation phase,
• 49:55on psychopathology
• 49:56measures, demographics,
• 49:58substance use, you know, obviously
• 50:01not cannabis, but other aspects
• 50:03of cannabis use,
• 50:04BRAIN measures.
• 50:06And then during the actual
• 50:07window of window of initiation,
• 50:09we can still match on
• 50:11things like like co occurring,
• 50:13alcohol use, tobacco use.
• 50:16And then after we have
• 50:18done this this kind of
• 50:19elaborate matching,
• 50:21and I I won't bore
• 50:22you with the details. I
• 50:23you know, there are ways
• 50:24to ensure that you have
• 50:26adequate,
• 50:27matches,
• 50:29and you do end up
• 50:30losing a little bit in
• 50:31the way of numbers because
• 50:32sometimes
• 50:33even with a massive
• 50:38good match for a particular
• 50:39cannabis initiator.
• 50:41So, unfortunately, sometimes you're having
• 50:43to drop,
• 50:45cannabis initiators.
• 50:46But essentially, then you're able
• 50:49to to, run a
• 50:51a,
• 50:52linear mixed effects model. You
• 50:54can test for a group
• 50:55by time interaction.
• 50:57This is, you know, the,
• 50:59this is
• 51:00kind of the matching scheme
• 51:01here.
• 51:04I won't work through that
• 51:05right now, but it's not
• 51:06it's not shown here, but
• 51:07we also have matched you
• 51:08know, this is really, you
• 51:10know, throwing everything at this.
• 51:11We've also matched for pre
• 51:13initiation
• 51:14BRAIN structure.
• 51:16So even, you know, on
• 51:17an ROI level basis on
• 51:19global,
• 51:20measures like like total brain
• 51:22volume.
• 51:24So we've really done everything
• 51:25we can, right, to, like,
• 51:27ensure
• 51:28that,
• 51:29you know, these
• 51:30are,
• 51:31equivalent
• 51:33groups.
• 51:35You can see,
• 51:36that
• 51:37so just, you know, again,
• 51:39to remind you, so baseline
• 51:40to year two, no one
• 51:42has initiated.
• 51:43Then after year two, that's
• 51:45when the initiation window starts.
• 51:48And, you know, so this
• 51:49is these are, CBCL
• 51:52scales.
• 51:53So what we see here
• 51:54behaviorally is is kind of
• 51:56this what looks like,
• 51:58after, the,
• 52:00initiators
• 52:01initiate, we see these rises,
• 52:04in attention problems,
• 52:06externalizing and internalizing problems, and
• 52:08total problems.
• 52:09This
• 52:10dovetails with what we've already
• 52:12seen in Imogen.
• 52:14I didn't go over this
• 52:15earlier, but, we we've seen
• 52:18similar patterns. We're looking at
• 52:19adolescent use from fourteen to
• 52:21nineteen,
• 52:22how that
• 52:23moderates,
• 52:25psychopathology
• 52:26trajectories,
• 52:28into early adulthood.
• 52:32We're also this is still
• 52:33this is pretty new, but
• 52:35then looking to see emerging
• 52:37thickness differences.
• 52:39So we don't have anything
• 52:40that's surviving whole brain correction,
• 52:42but we are starting to
• 52:43kind of see
• 52:45these,
• 52:46you know, again, these are
• 52:47small samples. I don't know
• 52:48if you caught that,
• 52:50but
• 52:52a hundred and fifty nine
• 52:53and a hundred and fifty
• 52:54nine,
• 52:55you know, these are small
• 52:56samples,
• 52:58but we're already starting to
• 52:59maybe see little glimmers here
• 53:00of reduced thickness
• 53:02in, some prefrontal areas in
• 53:05the initiators.
• 53:08I realize I have gone
• 53:10way over.
• 53:11I don't maybe,
• 53:13would it be worth trying
• 53:14to do a few minutes
• 53:15of,
• 53:16of,
• 53:19questions? Or
• 53:21I'm sorry. Sure. Right? No.
• 53:23That's that's fine. I,
• 53:24this was really, really interesting.
• 53:26So
• 53:27maybe we can move to
• 53:28questions. Do people have questions?
• 53:36So if no one does
• 53:37I I do have one
• 53:38question, Matt.
• 53:39Yeah.
• 53:41Sorry. Should I should I
• 53:42stop sharing? Or I I
• 53:43can't, I've
• 53:45I can't see you. Oh,
• 53:46okay. Is it?
• 53:48Sure. You could stop. Oh,
• 53:49oh, I'm sorry. You know
• 53:50what? I found the window
• 53:51here. I got it. I
• 53:52got it. Sorry.
• 53:53So, the question is,
• 53:57do you have any data
• 53:58that could speak to reversibility?
• 54:01As in in people who
• 54:02stop using in adolescents,
• 54:04who stop using cannabis? Is
• 54:06there any
• 54:07reversal of those changes that
• 54:09you observed?
• 54:10This will be an important,
• 54:13you know,
• 54:15important reason to convince young
• 54:17people about,
• 54:19it's not too late to
• 54:21to cut down or stop.
• 54:23Do you have any data
• 54:24to support that? So,
• 54:27it's what you know, this
• 54:29was a really,
• 54:30pressing question that we had,
• 54:33you know, when we were
• 54:34first getting that age twenty
• 54:35two data. That was actually
• 54:37one of the questions that
• 54:38we were really, really interested
• 54:40in. But I don't know.
• 54:42Not sure if I mentioned
• 54:43it or not. But,
• 54:46so in those adolescent initiators,
• 54:49there was I I I
• 54:50believe it was less than
• 54:51twenty.
• 54:53Less than twenty of them
• 54:54reported
• 54:56no use,
• 54:57in you know, afterwards.
• 54:59So the overwhelming majority
• 55:01of those adolescent initiators
• 55:04continued to use.
• 55:06So it was actually a
• 55:07real bummer. I mean, we,
• 55:09and I and maybe even
• 55:10a reviewer with that particular
• 55:12paper,
• 55:13had had asked this, but
• 55:14we, yeah, we were very
• 55:16bummed out by that. We
• 55:17were hoping that there would
• 55:18be more in the way,
• 55:20of, you know,
• 55:21people that had used early
• 55:23on and then stopped so
• 55:24that we could get at
• 55:25exactly that question.
• 55:27And I don't know. Maybe
• 55:28that speaks to just the
• 55:29phenomenology
• 55:30of this, you know, just
• 55:32overall. Like, it it's it's
• 55:33maybe,
• 55:36not that common where, you
• 55:38know, you,
• 55:38where you see this kind
• 55:40of very time limited use
• 55:42and and then nothing.
• 55:44So I guess we can
• 55:45see what what happens with
• 55:47a b c d.
• 55:49You know, there the the,
• 55:51use,
• 55:52is really picking up in
• 55:54in that sample.
• 55:55And so there there is
• 55:57the possibility that a b
• 55:58c d might be able
• 55:59to help with that. I
• 56:01do know there's also age
• 56:02twenty eight data that will
• 56:03become available,
• 56:05in imaging. So,
• 56:08I'm not sure when that
• 56:10is
• 56:11slated to be released, but
• 56:12that's another possibility.
• 56:15But we we're right there
• 56:16with you. We I mean,
• 56:17that was a very kind
• 56:18of pressing question for us
• 56:20too.
• 56:22The the the other related
• 56:23question I have is your,
• 56:25in
• 56:26in looking at dose response,
• 56:29the your measure of dose
• 56:30was frequency of use or
• 56:32number of times a person
• 56:33is used.
• 56:34Did you have any more
• 56:36finer grain,
• 56:38details about
• 56:39the potency of cannabis they
• 56:41used?
• 56:42You know, that is one
• 56:44of the limitations to all
• 56:45of the imaging work is
• 56:46that we really have focused
• 56:48heavily on this SBAD measure.
• 56:50And so it really does
• 56:52not get
• 56:53more granular than those categories.
• 56:57A, b, c, d, there
• 56:57is more there's a better
• 56:59characterization,
• 57:00including another avenue that we're
• 57:02wanting to get into,
• 57:04you know,
• 57:05different ways
• 57:07of using cannabis. So in
• 57:09particular, kind of,
• 57:11vaping versus
• 57:12edibles
• 57:13versus,
• 57:14you know, smoking,
• 57:16looking to you know, because
• 57:17there's this what another thing
• 57:19that really kind of bothers
• 57:20me is, are if there
• 57:21if this effect is real,
• 57:23is it really just the
• 57:24inhalation of, like, combustion byproducts?
• 57:27Is is this,
• 57:28you know so it would
• 57:29be really interesting, you know,
• 57:31and, again, if we have
• 57:32the numbers to do it,
• 57:33that's another avenue that we
• 57:35wanna get into, kind of
• 57:36looking at edibles
• 57:37versus,
• 57:39vaping versus smoking,
• 57:41to see the extent to
• 57:42which some of these findings
• 57:43may may hold.
• 57:45I I don't wanna hog
• 57:46the questions, but there are
• 57:47some in the chat.
• 57:51Oh, I didn't even see
• 57:52the I see that. I
• 57:53I can read them out
• 57:54to you, from one from
• 57:56Sushitra, Krishna, and Sarin. Thank
• 57:58you for your very nice
• 57:59presentation
• 57:59in the in imaging or
• 58:01a, b, c, d. Were
• 58:02you able to look at
• 58:03cognitive function
• 58:05correlates of the cortical thickness
• 58:07changes?
• 58:08Yeah. So in in imaging,
• 58:10I almost included some slides
• 58:12here.
• 58:13But yet we use the,
• 58:16the Cantab measure.
• 58:19And what we found again,
• 58:20this seems to kind of
• 58:21converge,
• 58:22with some of the reports
• 58:24out of Doctor. Heard's lab.
• 58:27But what we found what
• 58:29so it's the, the gambling
• 58:31task.
• 58:32And what we found was
• 58:33that the thinning was,
• 58:35in some of these,
• 58:37areas where we were seeing
• 58:38the adolescent effects.
• 58:40They were relating to,
• 58:43I think it's, like, overall
• 58:44proportion bet,
• 58:46and then also,
• 58:47whatever the measure of risk
• 58:49taking
• 58:50is. And,
• 58:52it was it was not
• 58:53whoppingly
• 58:53significant, but it was it
• 58:55was nominally significant. We were
• 58:57seeing associations,
• 58:58with those kind of,
• 59:02risk taking measures,
• 59:04on the cantab
• 59:06with,
• 59:08with the thinning. Some of
• 59:09the more conventional measures,
• 59:11we were not seeing anything.
• 59:13Like, you know, some of
• 59:14the
• 59:16IQ measures,
• 59:19to
• 59:19the best of my recollection,
• 59:21we we did not see
• 59:22anything with that. It was
• 59:23mostly And and measures of
• 59:25attention and memory? No. Nothing?
• 59:27Well, so we had that
• 59:28you know, so those attentional
• 59:29impulsivity findings
• 59:31in ABCD, we're kind of
• 59:33seeing what looks like little,
• 59:34you know, some signal with
• 59:35respect to CBCL attention problems.
• 59:41So, you know, it certainly
• 59:42looks like there might be,
• 59:43you know, this association
• 59:45between,
• 59:47again, what was termed attentional
• 59:49impulsivity, but basically just the
• 59:51ability to stay on tasks,
• 59:52stay focused,
• 59:54in this accelerated thinning. And
• 59:56it was that mediation effect
• 59:57that I kinda very briefly
• 59:59outlined where, you know, the
• 01:00:00association
• 01:00:01between adolescent cannabis use
• 01:00:04and, these measures of attentional
• 01:00:06impulsivity, there was this partial
• 01:00:08mediation through the accelerated thinning.
• 01:00:11There are two more questions.
• 01:00:12One is why were why
• 01:00:13were the assessments at fourteen,
• 01:00:15nineteen, and twenty two? And
• 01:00:17this and the last question
• 01:00:19is any,
• 01:00:20sex differences?
• 01:00:23Great question. So we'd have
• 01:00:24to ask Hugh for the
• 01:00:26why fourteen, nineteen, and twenty
• 01:00:28two? That was just those
• 01:00:29were the waves, the the
• 01:00:30data collection waves.
• 01:00:33So I don't know if
• 01:00:34there was a more,
• 01:00:36specific rationale for those
• 01:00:39ages in particular,
• 01:00:41but that's kinda just what
• 01:00:42we had to work with.
• 01:00:44And then,
• 01:00:45so with respect to sex,
• 01:00:47we interestingly
• 01:00:48have not.
• 01:00:49So we, you know, we,
• 01:00:52you know, again, we ran
• 01:00:53a whole host of sensitivity
• 01:00:55analyses,
• 01:00:58in in basically,
• 01:00:59so all the analyses I
• 01:01:00kind of overviewed.
• 01:01:02We did a lot in
• 01:01:03each of kind of their
• 01:01:03respective publications,
• 01:01:05but we did not find
• 01:01:07evidence of of
• 01:01:09sex
• 01:01:10qualifying,
• 01:01:11the patterns that I overview.
• 01:01:14So,
• 01:01:15but that's something that we
• 01:01:16definitely will continue to look
• 01:01:17at, in in a, b,
• 01:01:19c, d.
• 01:01:21So so, Matt, this is
• 01:01:22a great presentation.
• 01:01:25You certainly convinced,
• 01:01:26me with some,
• 01:01:28with with the with the
• 01:01:29data that you presented.
• 01:01:31Just one last question for
• 01:01:32you is,
• 01:01:34when you present this data
• 01:01:35to the public, the general
• 01:01:37public, do this remain skeptical?
• 01:01:40Are they still
• 01:01:41raising questions about causality
• 01:01:43or,
• 01:01:45do you have a sense
• 01:01:46of that?
• 01:01:47Yeah. Well, I mean, so,
• 01:01:49you know, in in the
• 01:01:50kind of the
• 01:01:52I would say some of
• 01:01:53the, you know, the presentations
• 01:01:54I've done, I haven't done
• 01:01:55tons to, like, the just
• 01:01:57the the public.
• 01:01:58But I find that in
• 01:02:00person, you know, there doesn't
• 01:02:01seem to be a whole
• 01:02:02lot
• 01:02:03of skepticism.
• 01:02:05But I still am haunted
• 01:02:07by those,
• 01:02:08those social media con I
• 01:02:10mean, like, I just was
• 01:02:11blasted. And I really wasn't,
• 01:02:13I mean, I really wasn't
• 01:02:14expecting that. I really and,
• 01:02:16again, maybe, you know,
• 01:02:17now after a little bit
• 01:02:19of time in this space,
• 01:02:20I'm not as naive, but
• 01:02:21I I,
• 01:02:23I really felt that,
• 01:02:26with some of the, you
• 01:02:26know, some of the coverage
• 01:02:27and and a lot of
• 01:02:28the the kind of comments
• 01:02:30on social media. And and
• 01:02:32so there seemed to be
• 01:02:33a very healthy dose
• 01:02:35of skepticism
• 01:02:36and criticism
• 01:02:39and people who just, I
• 01:02:40I think, no matter what
• 01:02:42you present,
• 01:02:43I think there are it's
• 01:02:44kind of almost,
• 01:02:45like,
• 01:02:46tribal
• 01:02:47where it's like,
• 01:02:49people are very committed to
• 01:02:51their side of the of
• 01:02:52the,
• 01:02:54the argument. But,
• 01:02:55yeah.
• 01:02:57Well, thank you very much,
• 01:02:58and, thanks everyone for attending.
• 01:03:00We will send out an
• 01:03:01announcement about the next,
• 01:03:03webinar for June.
• 01:03:06Thanks again, Matt. Really great
• 01:03:07presentation.
• 01:03:08Thanks so much for having
• 01:03:09me. It was it was
• 01:03:10a real pleasure. Thank you.