YC-SCAN2 February 2026 Webinar

Transcript

• 00:01All right. Well, good afternoon,
• 00:03everyone, and welcome to today's
• 00:05webinar through the Yale Center
• 00:06for the Science of Cannabis
• 00:07and Cannabinoids.
• 00:09It's my pleasure to introduce
• 00:10today's speaker, Doctor. Godfrey Purlson.
• 00:13Doctor. Purlson is a nationally
• 00:15and internationally
• 00:16recognized leader in psychiatric neuroscience
• 00:19and brain imaging research.
• 00:21He serves as the founding
• 00:22director of the Olin Neuropsychiatry
• 00:24Research Center at the Institute
• 00:26of Living and is a
• 00:27professor of psychiatry and neuroscience
• 00:29at Yale University School of
• 00:30Medicine.
• 00:31Over the course of his
• 00:32distinguished career Doctor. Carlson has
• 00:34made seminal contributions to our
• 00:36understanding of the neurobiological
• 00:38underpinnings
• 00:39of psychiatric disorders particularly schizophrenia
• 00:42and substance use disorders
• 00:44His work has helped bridge
• 00:45the gap between clinical psychiatry
• 00:48and advanced neuroimaging
• 00:49utilizing
• 00:50structural and functional MRI
• 00:52connectivity analyses, and translational approaches
• 00:55to better understand brain circuitry
• 00:57and its relationship to behavior
• 00:59and psychopathology.
• 01:01He's authored hundreds of peer
• 01:02reviewed publications and texts, including
• 01:04a book titled Weed Science,
• 01:06Cannabis Controversies and Challenges.
• 01:09Doctor. Perlson has also been
• 01:10continuously funded by the National
• 01:12Institutes of Health for decades,
• 01:14a testament to both the
• 01:15rigor and impact of his
• 01:16research.
• 01:18Beyond his scientific accomplishments, Doctor.
• 01:20Perlson is widely respected as
• 01:22a mentor,
• 01:23collaborator, and a true leader
• 01:24in the field of neuropsychiatry.
• 01:26And today Doctor. Pearlston will
• 01:28be sharing insights on cannabis
• 01:29and driving and we're fortunate
• 01:31to have him join us.
• 01:32As a reminder, Doctor. Pearlston
• 01:34has shared his preference to
• 01:35hold questions until the end,
• 01:36but please feel free to
• 01:37put your questions into the
• 01:38chat and we'll save a
• 01:39few minutes for q and
• 01:40a at the conclusion of
• 01:41the presentation.
• 01:43And with that, please join
• 01:44me in welcoming doctor Godfrey
• 01:45Carlson.
• 01:49Thanks, Ashley, for that lovely
• 01:51introduction.
• 01:52Normally,
• 01:54doctor D'Souza would join us,
• 01:55but he sends his regrets
• 01:56and says he's somewhere between
• 01:58New Delhi and Abu Dhabi
• 02:00in the air at the
• 02:01moment.
• 02:04So no conflicts of interest,
• 02:05and I'll try and get
• 02:06through a lot of complex
• 02:08material
• 02:09as quickly as possible so
• 02:10there's ample time for discussion
• 02:12at the end.
• 02:13So
• 02:14I do want to give
• 02:15a lot of background
• 02:16just to
• 02:18because the context is so
• 02:19rich, and
• 02:20the issues we're talking about
• 02:22really do need framing.
• 02:24So I'll try and do
• 02:25that.
• 02:26But the central theme is,
• 02:29does cannabis impair driving behaviors?
• 02:32Which ones for how long,
• 02:35after an acute dose?
• 02:37And to what extent? Are
• 02:38they very impaired or just
• 02:40a little bit?
• 02:41Is that behavioral,
• 02:43profile anything like that of
• 02:45alcohol?
• 02:46And
• 02:48important for medical legal reasons,
• 02:50is that impairment of cannabis
• 02:53related in any way temporally
• 02:55to the dose or to
• 02:56the blood levels of THC
• 02:58or its metabolites?
• 03:00And finally,
• 03:01can we detect stone drivers
• 03:03at the roadside?
• 03:04And because I'm a neuroimaging
• 03:06person,
• 03:07I'm gonna talk for just
• 03:09a couple of minutes about
• 03:10brain changes. I can't help
• 03:12myself, but the focus is
• 03:13on behavior and not on
• 03:15those changes.
• 03:16So this is obviously a
• 03:18topic with many public health
• 03:19implications.
• 03:21This is an old slide,
• 03:22but you can see that
• 03:23since alcohol,
• 03:26DWIs
• 03:27have been,
• 03:28laws have been more strictly
• 03:30enforced,
• 03:31the number of drug related
• 03:34fatal vehicle crashes has trended
• 03:36downwards,
• 03:37whereas the
• 03:39intoxicated
• 03:40from other substances,
• 03:42mainly cannabis
• 03:43road deaths, have trended upwards,
• 03:45and that's a trend that's
• 03:46continued.
• 03:47This is something that's preoccupied
• 03:49the National Highway and Traffic
• 03:51Safety Administration. So they put
• 03:53out multiple
• 03:54public
• 03:55service announcements about marijuana impaired
• 03:57driving.
• 03:59So some of the background
• 04:01trends are that more states
• 04:02have legalized
• 04:04recreational
• 04:05or medicinal cannabis or both,
• 04:09that the percentage
• 04:10of THC
• 04:12has steadily risen as a
• 04:14triumph of selective
• 04:15breeding,
• 04:17whereas the amount of cannabidiol
• 04:20has stayed the same. In
• 04:21fact, these things are related
• 04:23genetically. If you breed the
• 04:25cannabis plant to express more
• 04:27THC,
• 04:28you're gonna drive down the
• 04:29amount of cannabidiol.
• 04:32For the first time last
• 04:34year, the amount of daily
• 04:36use of
• 04:37cannabis
• 04:39actually
• 04:40exceeded that of alcohol, which
• 04:41is trending down very slightly.
• 04:44So more people are smoking
• 04:45weed than getting drunk.
• 04:47And
• 04:48unlike the perception of safety
• 04:50for other drugs, people are
• 04:51increasingly seeing
• 04:53cannabis as a safe substance
• 04:55and are therefore more inclined
• 04:56to use it.
• 04:58So in talking about cannabis
• 05:00related driving, there are four
• 05:02big sources of confusion,
• 05:04and
• 05:05I really wanna go into
• 05:06those just to try and
• 05:08disambiguate
• 05:09what we're looking at.
• 05:11So the
• 05:12pharmacokinetics
• 05:13of THC
• 05:15differs radically from that of
• 05:17alcohol, so we'll talk a
• 05:18little bit about that.
• 05:20The cannabis plant contains multiple
• 05:22compounds,
• 05:23many of which are intoxicating,
• 05:25but some of which have
• 05:26opposing effects.
• 05:28So that's not like alcohol.
• 05:30That cannabis is used in
• 05:32multiple nonequivalent dosage formats.
• 05:35So taking a bong hit
• 05:37is not the same as
• 05:38eating an edible
• 05:39in terms of the
• 05:41time of onset,
• 05:42the metabolites produced,
• 05:45or
• 05:45the,
• 05:47ultimate blood levels.
• 05:49And unlike alcohol, there's no
• 05:51standard dose equivalent for cannabis.
• 05:54So for blood alcohol concentrations
• 05:57or breath alcohol concentrations,
• 05:59which are basically the same
• 06:01thing,
• 06:01the relationships with alcohol are
• 06:03very simple.
• 06:05So the more you drink,
• 06:07the more intoxicated you get.
• 06:09The more you drink, the
• 06:10higher your blood or breath
• 06:11alcohol concentrations.
• 06:14And the more you drink,
• 06:15the higher
• 06:16your BAC and the higher
• 06:18your
• 06:19odds of being involved in
• 06:21a a motor vehicle crash.
• 06:23So those things are all
• 06:24very simple.
• 06:25And, unfortunately,
• 06:26people have
• 06:28said,
• 06:29since that's the case for
• 06:30alcohol, it's also the case
• 06:31for cannabis, which is not
• 06:32the case.
• 06:33So what follows from alcohol's
• 06:35water solubility? We can make
• 06:37cocktails. Woo hoo.
• 06:40Concentrations
• 06:41everywhere in the body, including
• 06:42the brain and the breath
• 06:44and
• 06:47blood and saliva
• 06:48are all related and all
• 06:50pretty much the same.
• 06:52The BAC levels reflect plasma
• 06:54levels reflect brain levels.
• 06:56They are, as we said,
• 06:58proportional to dose decline predictably.
• 07:01And as breath is moist,
• 07:02there's plenty of alcohol available
• 07:04to sample at the roadside.
• 07:07With plasma THC, the relationships
• 07:09are way more complicated
• 07:11because
• 07:11THC is fat soluble, not
• 07:13at all water soluble.
• 07:15So if you take
• 07:17a puff from a joint
• 07:18or a bong hit,
• 07:19you get a quick rise
• 07:22in plasma THC that rapidly
• 07:24within five minutes or so
• 07:26returns to a basal
• 07:28low level.
• 07:29And for some metabolites, it
• 07:31stays up a bit longer
• 07:32like the seventeen hydroxy.
• 07:35But the tissue compartment levels
• 07:37are completely different based on
• 07:39how much fat is in
• 07:40there. So blood
• 07:42is
• 07:43moderate,
• 07:44brain is high, but adipose
• 07:46tissue is very high. So
• 07:48ultimately, what happens is you
• 07:49get this spike in blood,
• 07:50which disappears
• 07:52down to this low basal
• 07:53level. But
• 07:55hours later, the levels in
• 07:56fat are still rising.
• 07:58They peak,
• 08:00and then
• 08:01THC is very slowly released
• 08:03from those adipose tissues.
• 08:06So how long drugs stay
• 08:07in your blood
• 08:08for alcohol is somewhere
• 08:10here,
• 08:11which is relatively short, as
• 08:13we said. For cannabis, it's
• 08:15incredibly long.
• 08:17So
• 08:18you can still measure cannabis
• 08:19in the blood a couple
• 08:20of weeks if you're a
• 08:21regular smoker,
• 08:23after you last used.
• 08:25So that has profound medical
• 08:27legal consequences.
• 08:29And unlike alcohol, how high
• 08:31you feel at a particular
• 08:32time
• 08:33is not only related to
• 08:35your
• 08:36initial blood spike
• 08:39transiently,
• 08:39but, basically, you're still feeling
• 08:41high and getting high at
• 08:42a time when your blood
• 08:43levels are actually coming down.
• 08:46So that's very, very different.
• 08:48And finally, there are these
• 08:49hysteresis
• 08:50curves
• 08:51where higher doses
• 08:53don't necessarily produce more effects.
• 08:56So low doses of THC
• 08:57are anxiolytic,
• 08:59whereas higher doses provoke,
• 09:03anxiety.
• 09:05How you feel the drug
• 09:06changes as you take in
• 09:08more and more, and they're
• 09:09not linearly related.
• 09:13So what follows from THC's
• 09:14fat solubility is we need
• 09:16fat to make edibles,
• 09:19and the distribution through the
• 09:21body is uneven.
• 09:23And because
• 09:24breath is moist, there's almost
• 09:26no THC available to detect.
• 09:28We don't breathe out
• 09:30lipid soluble substances for obvious
• 09:32reasons.
• 09:33So there's
• 09:34almost
• 09:36zero THC in breath.
• 09:40So
• 09:41cannabis use does impair driving
• 09:43acutely,
• 09:44and
• 09:45I'll just put that
• 09:47that
• 09:48that statement out there and
• 09:49then try and show you
• 09:51why that's the case.
• 09:52And the impairment is primarily
• 09:54due to THC,
• 09:55which as we know is
• 09:56a c b one receptor
• 09:58partial agonist,
• 09:59and cannabidiol has very little
• 10:01relevance.
• 10:02In fact, although people commonly
• 10:04say CBD
• 10:05modulates THC,
• 10:07There's actually not much evidence
• 10:09for that. CBD has very
• 10:11little modulating
• 10:12influence on THC impairment.
• 10:15But the questions are, what
• 10:18sort of impairment, how long
• 10:19does it last, and with
• 10:20what consequences to driving?
• 10:23And because more and more
• 10:24people are using as legalization
• 10:27is happening in more and
• 10:28more states or decriminalization,
• 10:31the numerator is that more
• 10:32drivers are exposed.
• 10:34And because the dose is
• 10:35increasing, more people are exposed
• 10:37to higher doses and then
• 10:39driving. So this is potentially
• 10:41a public health problem.
• 10:43So what's the evidence for
• 10:45that? So epidemiologically,
• 10:48there are
• 10:50studies looking at post crash
• 10:52blood tests, so people involved
• 10:54in a major motor vehicle
• 10:55accident
• 10:56and
• 10:57epidemiologists
• 10:58taking samples
• 10:59or looking at results from,
• 11:01law enforcement.
• 11:03And looking at the increases
• 11:06in motor vehicle crashes
• 11:08following
• 11:09cannabis legalization in states that
• 11:12have legalized
• 11:13compares that to those that
• 11:15don't. So the best meta
• 11:17study is this Roggeberg and
• 11:19Elvick
• 11:20Scandinavian meta analysis,
• 11:23back ten years ago,
• 11:24looking at a quarter of
• 11:25a million individuals across multiple
• 11:28studies.
• 11:29And they recognized
• 11:30they calculated an odds ratio
• 11:33of being involved in a
• 11:34motor vehicle crash of about
• 11:36one point four.
• 11:38Okay?
• 11:39So compare that to alcohol
• 11:41where the odds ratios are
• 11:42about twenty,
• 11:44at a BAC
• 11:45of point one.
• 11:47So the these are very,
• 11:48very different beasts,
• 11:50and it's worth saying that
• 11:51in this in this comparison.
• 11:55Here are some typical studies,
• 11:58this one from, Cayman,
• 12:00six years ago,
• 12:02but there are multiple similar
• 12:04subsequent studies
• 12:05looking at states that,
• 12:07legalized
• 12:09cannabis
• 12:10comparing
• 12:11pre legalization study period,
• 12:13waiting a while
• 12:15from the law being passed
• 12:16until
• 12:17dispensaries were actually selling the
• 12:19substance,
• 12:20then looking at what happened
• 12:21to traffic fatality rates.
• 12:24And there's reasonable evidence
• 12:26that compared to these control
• 12:28states, you can see at
• 12:29the bottom,
• 12:30that most states that legalized
• 12:33had somewhat of an increase
• 12:34except for Washington,
• 12:36in traffic fatalities post legalization.
• 12:41So
• 12:44the epidemiologic studies are out
• 12:45there
• 12:47calculate that there's about a
• 12:48doubling in deaths
• 12:50per billion vehicle miles traveled
• 12:53due to cannabis.
• 12:55So
• 12:56because motor vehicle deaths are
• 12:57very uncommon overall,
• 12:59you really do need this
• 13:01billion
• 13:02vehicle miles traveled to be
• 13:03able to have a
• 13:06a genuine metric.
• 13:07So if that's accurate and
• 13:09it's disputed as we'll see
• 13:10in a moment, that translates
• 13:12into about a twenty percent
• 13:14or
• 13:15approximately seven thousand person
• 13:18increase in USA traffic fatalities
• 13:20that are about thirty seven
• 13:22thousand back in twenty eighteen.
• 13:24But what we don't know
• 13:26is whether alcohol consumption rates
• 13:28decrease
• 13:29as cannabis consumption rises, which
• 13:31would somewhat mitigate that figure.
• 13:34And what not all studies
• 13:36examine is what percentage of
• 13:38cannabis impaired drivers
• 13:39also had some alcohol,
• 13:42on board.
• 13:44So there are epidemiologic
• 13:46skeptics, most,
• 13:48notably,
• 13:49Michael White and Nick Burns
• 13:50in Australia
• 13:51who say we've revered reviewed
• 13:53the epidemiologic and psychological literature
• 13:56on the risk of THC
• 13:57positive driving.
• 13:59We concluded those risks are
• 14:00lower than most people would
• 14:02assume.
• 14:03And,
• 14:05their
• 14:06skepticism is based on two
• 14:08main factors. One is if
• 14:10someone
• 14:11is in a involved in
• 14:12a fatal crash,
• 14:14you take a blood sample
• 14:15and you find THC there.
• 14:17Is that THC because they
• 14:19are acutely intoxicated
• 14:21and their driving was impaired,
• 14:23or did they smoke a
• 14:24week ago
• 14:25and they're actually driving it
• 14:27baseline
• 14:28cap
• 14:29competence,
• 14:30but THC just happened to
• 14:32be there because it's released
• 14:33slowly from fat because of
• 14:35its p k profile.
• 14:37The other thing is that
• 14:38they think the,
• 14:39numbers are miscalculated,
• 14:41and it's closer to two
• 14:42thousand and seven thousand.
• 14:44So they're not denying that
• 14:45there are some increases.
• 14:47They just think they're exaggerated
• 14:49by
• 14:50epidemiologic miscalculations.
• 14:52Okay. So bear that in
• 14:53mind.
• 14:55So
• 14:56there are a few well
• 14:57controlled
• 14:59studies of actual cannabis impaired
• 15:01driving.
• 15:04And
• 15:04as I said, because of
• 15:06persistence, it's unclear to how
• 15:07to interpret
• 15:08body fluid sampling.
• 15:11Who presents which data
• 15:14has to be viewed with
• 15:15skepticism
• 15:17because there is
• 15:19agenda driven
• 15:21reporting of the statistics
• 15:23from both sides. Remember, there's
• 15:26there's big alcohol
• 15:28and big cannabis,
• 15:29and they have competing agendas.
• 15:31And some of the reports
• 15:33in states
• 15:34are,
• 15:35funded by one or the
• 15:37other of these,
• 15:38So you really have to
• 15:39take a step back and
• 15:41put on your skeptic glasses,
• 15:43to be able to look
• 15:44at reports
• 15:45carefully.
• 15:47And unlike alcohol, there's no
• 15:49agreed on roadside
• 15:50valid sobriety testing
• 15:52for cannabis impairment, which is
• 15:54something we'll be spending a
• 15:56lot of time looking at.
• 16:01Other things are to note
• 16:04that
• 16:05if we take you in
• 16:06the lab and give you
• 16:08cannabis
• 16:09and you were impaired on
• 16:10a cognitive task,
• 16:12that may be completely irrelevant
• 16:14if that cognitive task
• 16:16has no relevance to driving.
• 16:19So just because someone's impaired
• 16:21on a task doesn't mean
• 16:23there's anything wrong with their
• 16:24driving. So that's something that
• 16:27has to be shown empirically.
• 16:30Second thing is that we
• 16:32often, at the roadside, don't
• 16:33know,
• 16:34someone's baseline draping capability.
• 16:37So if you pull over
• 16:38someone and they have cannabis
• 16:40in their blood and they
• 16:41appear to be impaired,
• 16:43maybe they were impaired at
• 16:44baseline,
• 16:46and they're just a bad
• 16:47driver. And now they're a
• 16:48bad stone driver, but no
• 16:49worse off than they were
• 16:50before. So you have to
• 16:52know something about that.
• 16:54Another unanswered question is,
• 16:57if I take a couple
• 16:58of beers and take a
• 16:59couple of bong hits,
• 17:00is that a plus b
• 17:02or is it a times
• 17:04b, and they're synergistic
• 17:05rather than merely additive?
• 17:09There's a lack of understanding
• 17:10regarding
• 17:11the time course of THC
• 17:13related driving impairment,
• 17:15and
• 17:16we're pretty ignorant of THC's
• 17:18specificity
• 17:19in testing
• 17:21compared, say, to opioids or
• 17:22benzodiazepines
• 17:24as compared to alcohol. So
• 17:26bear those things in mind.
• 17:28So let's dig a little
• 17:29deeper into these issues.
• 17:32So multiple products with different
• 17:34pharmacokinetics
• 17:35and different dosage rates.
• 17:40So for alcohol, there are
• 17:42equivalences. So a standard drink
• 17:45is
• 17:46like
• 17:47sixteen ounces of beer, five
• 17:48ounces of wine, eight ounces
• 17:50of malt liquor, and so
• 17:51on. But
• 17:53taking
• 17:54a brownie
• 17:55with
• 17:56five milligrams of THC,
• 17:59smoking a dab with five
• 18:00milligrams of THC,
• 18:02and taking a bong hit
• 18:03with five milligrams of THC
• 18:05are not at all equivalent
• 18:07pharmacokinetically
• 18:08or in terms of intoxication,
• 18:11at its peak or in
• 18:13its duration.
• 18:14So there's no intuitive standard
• 18:16joint measure, and there's no
• 18:18getting around that.
• 18:20So driving.
• 18:23Can we detect cannabis impaired
• 18:25drivers reliably
• 18:26at the roadside?
• 18:28And as I said, we
• 18:29need validity. So if you
• 18:31have a measure, is that
• 18:32measure actually related
• 18:34to driving impairment
• 18:36rather than just a measure
• 18:37of being stoned? And that
• 18:39that's a very important distinction.
• 18:42So what is driving?
• 18:44I like to think of
• 18:45driving as a pyramid
• 18:47of increasing complexity.
• 18:49So if you look at
• 18:50bottom up, basically, can you
• 18:52see things?
• 18:54What's your visual reaction time?
• 18:57Can you integrate what you're
• 18:59seeing on the road
• 19:00with your motor patterns in
• 19:02terms of speeding up or
• 19:04braking?
• 19:05And at the abscissa of
• 19:07the pyramid at the very
• 19:08peak,
• 19:09you get driving as an
• 19:10emergent property, or you can
• 19:12look at it top down.
• 19:15So it's basically a complex
• 19:16behavior with multiple integrated cognitive
• 19:19and motor components
• 19:21that you can measure separately
• 19:22or integrate together.
• 19:25So cannabis certainly impairs individual
• 19:28psychomotor
• 19:29abilities related to vehicle driving.
• 19:32So that's well established. But
• 19:34as I said,
• 19:35they may be related to
• 19:37motor vehicle driving,
• 19:38but are they related to
• 19:40actual driving either
• 19:42virtually,
• 19:43in in virtual reality on
• 19:44a driving simulator
• 19:46or on an actual road?
• 19:49So
• 19:49some of them seem to
• 19:50be looking at sensitivity and
• 19:53specificity,
• 19:54particular tests,
• 19:56like a maze task, just
• 19:58to pick one example,
• 19:59related to planning
• 20:02and, frontal lobe function,
• 20:04gives you about seventy eight
• 20:06and eighty two percent sensitivity
• 20:08and specificity.
• 20:09And some individual tests perform
• 20:12a little better.
• 20:14But my contention, and this
• 20:15is something that our research
• 20:17has borne out, is that
• 20:18you really need a battery
• 20:19of tests
• 20:20and to get a behavioral
• 20:22fingerprint.
• 20:24So
• 20:25does cannabis impair on road
• 20:27or simulated driving abilities? And
• 20:29that's something we and others
• 20:31have looked at extensively.
• 20:33So
• 20:34the literature summary is often
• 20:36looked at single driving tasks
• 20:39and a particular task, which
• 20:41I'll get into in a
• 20:42second.
• 20:45So,
• 20:48driving impairment seems to be
• 20:50most noticeable from the literature
• 20:51in activities that are more
• 20:53automatic
• 20:54and simpler
• 20:55and require minimal attentional control
• 20:58and
• 21:00less
• 21:00impaired,
• 21:03on more complex frontal lobe
• 21:05planning abilities,
• 21:07which is interesting. Yeah. I
• 21:09would expect the opposite.
• 21:11And STLP,
• 21:14which if you come from
• 21:15Ireland is not the Social
• 21:16Democratic
• 21:17Labor Party, but,
• 21:19standard deviation of lane position
• 21:21or weaving
• 21:22is the most robustly affected
• 21:24in many tests.
• 21:25But,
• 21:27sometimes,
• 21:28the only metric that's been
• 21:30studied for driving, so hard
• 21:32to say.
• 21:33And
• 21:35there are suggestions of long
• 21:36lasting impairments,
• 21:38but
• 21:39the majority of studies have
• 21:40not looked more than a
• 21:41couple of hours out after
• 21:42an acute dose.
• 21:44And there's little support for
• 21:46impairment being related to THC
• 21:48blood or oral fluid levels,
• 21:51unlike alcohol. So we'll get
• 21:52into that.
• 21:55So probably one of the
• 21:56best tasks was done by
• 21:58Tom Marcotte at UCSD
• 22:00that he published four years
• 22:02ago.
• 22:03So So he looked at
• 22:04a couple of hundred
• 22:05regular cannabis users
• 22:07and randomized
• 22:08each person to either placebo,
• 22:11six percent THC, or thirteen
• 22:13percent THC
• 22:15in parallel groups. So it
• 22:16wasn't the same person
• 22:18challenged multiple times.
• 22:20So everyone smoked
• 22:22a point seven gram joint
• 22:24ad lib, so not everyone
• 22:26smoked the same amount even
• 22:28within a dose range, and
• 22:30then drove on a simulator
• 22:31for twenty five minutes.
• 22:33And
• 22:34no single driving metric was
• 22:36significant,
• 22:37so he summarized everything into
• 22:39a composite score
• 22:41and
• 22:42then found that composite score
• 22:45was
• 22:46impaired
• 22:46but unrelated to THC dose
• 22:49or use history
• 22:51or blood THC levels.
• 22:53Okay. So that's kind of
• 22:55the best study that we
• 22:56have out there at the
• 22:57moment.
• 22:58So how does this actually
• 23:00work in practice?
• 23:01So I'm driving along the
• 23:03road, and I'm pulled over
• 23:05by a law enforcement
• 23:07officer.
• 23:08And that could happen for
• 23:09probable cause, which is my
• 23:11driving is awful and weaving
• 23:13or driving excessively slow or
• 23:16fast,
• 23:17or I've got a busted
• 23:18taillight or something,
• 23:20or I encounter,
• 23:22a DWI
• 23:23checkpoint.
• 23:24So they're just pulling over
• 23:26everyone and testing them.
• 23:28And if my
• 23:30roadside impairment testing
• 23:33is abnormal, I'm staggering, I
• 23:35can't walk a straight line,
• 23:36I'm slurring my words,
• 23:38I appear to be intoxicated.
• 23:41I'll
• 23:42be tested with a breathalyzer.
• 23:45And
• 23:46if my breathalyzer is abnormal,
• 23:48that is I have more
• 23:50than the required amount of
• 23:51alcohol on board, then I'm
• 23:54automatically
• 23:55WI for alcohol.
• 23:57If my BAC is negative
• 23:59but I appear to be
• 24:00intoxicated,
• 24:02law enforcement may choose to
• 24:04have me go to a
• 24:05local testing facility
• 24:07and have my blood drawn
• 24:09for drugs including THC.
• 24:11And the average across the
• 24:13US for that is ninety
• 24:14minutes.
• 24:15And they may call in
• 24:16a drug recognition expert,
• 24:18who's someone who has been
• 24:20trained
• 24:21to recognize behavioral impairment in
• 24:23drivers
• 24:24due to specific substances.
• 24:27So,
• 24:29DREs are excellent for alcohol
• 24:31but prove to be extremely
• 24:33variable for cannabis at their
• 24:35best. Some can do it,
• 24:37most cannot.
• 24:39And the blood test test
• 24:41for THC are its metabolites,
• 24:43but that has a problem,
• 24:45because it takes ninety minutes
• 24:47to get me to a
• 24:48test site, and the blood
• 24:49that's drawn there may reflect
• 24:51prior
• 24:52use a long time ago
• 24:53that's leaching out of my
• 24:55fat stores.
• 24:57So
• 25:00the evaluation that's done at
• 25:01the roadside is basically
• 25:04based on
• 25:06intoxication patterns for drunk drivers
• 25:10and is,
• 25:12to put it bluntly, no
• 25:13pun intended,
• 25:15really not applicable to cannabis
• 25:17intoxication.
• 25:18Cannabis
• 25:20intoxicated drivers can perform perfectly
• 25:22okay on this
• 25:24in many cases.
• 25:26So what if THC is
• 25:28detected in my blood?
• 25:31A number of states have
• 25:32so called per se laws.
• 25:34Basically,
• 25:36if you have THC in
• 25:37your blood, they presume impairment.
• 25:40So eleven states prohibit driving
• 25:42with any amount of detectable
• 25:44blood THC.
• 25:45THC shows up. You're
• 25:47stoned, you're intoxicated while driving,
• 25:49you're convicted.
• 25:51A few states imply
• 25:53a legal THC cutoff level,
• 25:55but those levels are not
• 25:57the same in different states.
• 26:00And the remainder of states
• 26:01that don't specify levels
• 26:03have
• 26:04incapacitated
• 26:05by or under the influence
• 26:07of laws
• 26:09that are slightly different, but
• 26:10boiled down to a somewhat
• 26:11vague prohibition
• 26:13on driving while high,
• 26:14however they test it.
• 26:16So these just show you
• 26:18the
• 26:18per se laws in different
• 26:20states,
• 26:21and I'm not gonna go
• 26:22into that in detail. We
• 26:23can come back to it
• 26:24if people are interested.
• 26:26But the bottom line on
• 26:28per se laws
• 26:29is I'm driving along.
• 26:32I'm pulled over.
• 26:33Someone decides that I'm mildly
• 26:35impaired.
• 26:38I get my blood drawn.
• 26:39There's THC on board, so
• 26:41I get an automatic conviction,
• 26:42although I haven't smoked cannabis
• 26:44for two weeks.
• 26:45So these laws to us
• 26:47seem to be arbitrary and
• 26:48not science based.
• 26:50And
• 26:51they could be replaced by
• 26:53a THC breathalyzer,
• 26:55to detect recent use
• 26:58or by a specific field
• 26:59sobriety test for cannabis specific,
• 27:03behavioral driving
• 27:04impairment
• 27:06that actually reflects impaired driving
• 27:08and and not
• 27:10something random.
• 27:12So
• 27:13the Hound Labs
• 27:15for the last ten years
• 27:16has been trying to market,
• 27:19a cannabis breathalyzer,
• 27:21which is a very difficult
• 27:22task given the nanomolar
• 27:24equivalents
• 27:26of
• 27:27THC that are detectable in
• 27:28breath.
• 27:30And, ultimately, that venture failed,
• 27:32and they went out of
• 27:33business a year ago.
• 27:35And there's no replacement
• 27:36on on the horizon.
• 27:39So let's rethink these issues.
• 27:43What would be a valid
• 27:44and reliable measure of driving
• 27:46impairment?
• 27:47So we've chosen to validate,
• 27:51all of our measures
• 27:55using virtual driving and driving
• 27:56simulators,
• 27:58because we think that has
• 28:00sufficient construct criteria and validity
• 28:02since we can't ethically allow
• 28:04our subjects to drive stoned
• 28:06down I ninety five and
• 28:07avoid traffic.
• 28:10Some major considerations and choices
• 28:12are here.
• 28:15So I won't dwell on
• 28:17those, but I'll go through
• 28:18them very briefly.
• 28:20So simulated driving for us
• 28:23is the gold standard for
• 28:24relevant impairment, and we've done
• 28:26this with alcohol impaired driving
• 28:28over the last twenty years
• 28:30and other substances
• 28:32and more recently with cannabis.
• 28:34It's controlled. It's safe. You
• 28:36can have
• 28:38endlessly repeated variance
• 28:39so people don't learn that
• 28:41ten minutes into the task,
• 28:42a dog runs across the
• 28:44road
• 28:45or another vehicle pulls in
• 28:47front of you. It's different
• 28:48every time.
• 28:49And you can mimic scenarios
• 28:51that are impractical
• 28:52or unethical
• 28:54to test in real life,
• 28:55like a child suddenly runs
• 28:57into the road.
• 28:58And it's a very easily
• 29:00programmable,
• 29:01manipulated environment, so you can
• 29:03change the weather and make
• 29:04it skiddy or make it
• 29:06rain or foggy or anything
• 29:07you want.
• 29:08So the ideal
• 29:10is Iowa's
• 29:11National Advanced Drivers'
• 29:13Driving Simulator.
• 29:15Whoever thought of the acronym
• 29:16was clearly not having a
• 29:18thoughtful day.
• 29:19And this is basically an
• 29:21enormous,
• 29:23bubble
• 29:24on rails so that it
• 29:26actually accelerates and decelerates
• 29:28and feels like real driving.
• 29:30I've actually tried this. It's
• 29:32amazing.
• 29:33And you you can lower
• 29:34anything into that, an actual
• 29:35vehicle,
• 29:36an Abrams tank, or anything
• 29:38that you want.
• 29:41So, we've looked at this,
• 29:43and it wasn't practical for
• 29:44us to use in Iowa,
• 29:45but we've come up with
• 29:47something that we think is
• 29:48close.
• 29:49Now what we did is
• 29:51Virginia Tech has a so
• 29:53called smart road,
• 29:55And
• 29:56the
• 29:57smart road
• 29:58is
• 29:59a couple of kilometers of
• 30:01real
• 30:03interstate
• 30:04that connects with an interstate,
• 30:05but as a test bed
• 30:07for driving.
• 30:08They can make it rain
• 30:09or snow, and the entire
• 30:11tarmac is embedded with sensors.
• 30:14And all the cars that
• 30:15drive on it have a
• 30:16bunch of computer hardware in
• 30:18their trunks so that you
• 30:19can get an instant readout
• 30:21of how people are driving.
• 30:23We compared our driving simulator
• 30:25software performance
• 30:27under conditions of impairment
• 30:29to actual people driving
• 30:31one of these cars on
• 30:33Virginia Tech smart road, and
• 30:35they do correspond
• 30:36significantly.
• 30:37So we're confident that we
• 30:38have a valid set of
• 30:40parameters.
• 30:42We use people we don't
• 30:44use drug virgins. We use
• 30:45people who are have
• 30:48more or less experience
• 30:50with administering cannabis to themselves.
• 30:53And
• 30:54we try and make it
• 30:55as naturalistic as possible. So
• 30:57we have smoked, oral, or
• 30:59vaporized forms,
• 31:01versus intravenous
• 31:03THC,
• 31:04but we have not tested
• 31:06edibles
• 31:08as yet. We plan to
• 31:10in the future.
• 31:17And the amounts we've give,
• 31:19until recently, were limited by
• 31:22what NIDA supplies to investigators.
• 31:24We're not able to go
• 31:25into our local dispensary and
• 31:27buy what people are using.
• 31:29We have to use what
• 31:30NIDA supplies
• 31:31or authorizes.
• 31:33So until recently, we've given
• 31:35people half a gram of
• 31:36either seven percent
• 31:38or thirteen percent THC
• 31:40with no CBD
• 31:42added.
• 31:43The highest dose is equivalent
• 31:45to sixty five milligrams of
• 31:46THC,
• 31:47which our subjects tell us,
• 31:49eighty percent of our subjects
• 31:51tell us is equivalent to
• 31:52what they dose themselves when
• 31:54they're using.
• 31:55But, admittedly, that's a much
• 31:57lower percent
• 31:59THC than is available to
• 32:00most people at most modern
• 32:02dispensaries.
• 32:05And we do
• 32:06measurements of multiple measures following
• 32:08a single acute dose.
• 32:12We take repeated blood and
• 32:13saliva samples.
• 32:15We take
• 32:17we look at the Draeger
• 32:18for immediate THC in addition
• 32:20and look at multiple
• 32:23samples at all of these
• 32:24times following an acute dose.
• 32:30Nothing really relevant there.
• 32:34Several of the labs at
• 32:35Yale use intravenous THC, and
• 32:37the advantage of that is
• 32:38that everyone gets exactly the
• 32:40same dose at exactly the
• 32:42same time.
• 32:44UCSD
• 32:45uses a joint or a
• 32:46bong,
• 32:48so different people get very
• 32:50variable doses,
• 32:52and they smoke that on
• 32:53their own time ad lib,
• 32:55so you don't have to
• 32:56finish.
• 32:58We use paste inhalation from
• 33:00a vaporizer bag, and everyone
• 33:02has to finish the bag.
• 33:03So everyone gets the same
• 33:05dose, but not necessarily at
• 33:06quite the same rate.
• 33:08So there are pro and
• 33:09con arguments for those methods
• 33:11of administration,
• 33:12balancing
• 33:13knowing exactly what someone got
• 33:15versus
• 33:16natural
• 33:18how naturalistic that administration
• 33:20is.
• 33:22So
• 33:23what what we used is
• 33:24three different driving tasks over
• 33:27a thirty minute,
• 33:29driving
• 33:30scenario.
• 33:32We used a a steering
• 33:34task that measures operational aspects
• 33:36of driving,
• 33:37a car following task that
• 33:38looks at tactical aspects,
• 33:41and then overtaking task that
• 33:43looks at strategic aspects.
• 33:48I don't wanna dwell on
• 33:50that.
• 33:51So
• 33:52what we wanted were tests
• 33:54that were a battery of
• 33:55tests that were simple, quick,
• 33:56and robust,
• 33:58practical to give it the
• 34:00roadside.
• 34:01So these have to be
• 34:02things you could administer
• 34:04when it's raining at night,
• 34:07on a computer, when a
• 34:08bunch of traffic is driving
• 34:10by.
• 34:11So that constrains what you
• 34:13can do.
• 34:15We wanted things that accurately
• 34:17predicted
• 34:19and had strong criterion related
• 34:22validity. That is they actually
• 34:24did match and that were
• 34:26highly correlated,
• 34:28with
• 34:31virtual driving.
• 34:33And as I said earlier,
• 34:34we wanna know how long
• 34:35these lasted,
• 34:36how reliable they are across
• 34:38different populations,
• 34:39occasional versus frequent users,
• 34:42and are they specific to
• 34:44cannabis?
• 34:47So the first
• 34:50thing I'll tell you about
• 34:51used
• 34:52thirty eight subjects, twenty seven
• 34:54men,
• 34:54exposed to placebo,
• 34:56low and high doses on
• 34:58on three separate days at
• 35:00random,
• 35:01completely double blind.
• 35:03And
• 35:04we assess those three driving
• 35:06risk measures,
• 35:09and that this was mainly
• 35:10looking at driving
• 35:12rather than developing a test
• 35:13battery, which is the next
• 35:14set of data I'll give
• 35:16you.
• 35:17So this is our ideal
• 35:18study subject.
• 35:20In fact, these are the
• 35:22statistics
• 35:23of the people that we
• 35:24looked at,
• 35:26so I won't dwell on
• 35:27that.
• 35:28And this is the overall
• 35:29study design. So
• 35:31we got predose measures, including
• 35:33blood tests.
• 35:34Person had an IV catheter.
• 35:36We got blood and then
• 35:38saliva at all of these
• 35:39test points.
• 35:41And,
• 35:42they went in the scanner
• 35:44and had the driving paradigm
• 35:46in the scanner so we
• 35:47could look at their brain.
• 35:48And that these other three
• 35:50remaining test points,
• 35:53over five plus hours,
• 35:55they got a battery of
• 35:58little tests,
• 36:00plus either were randomized to
• 36:03you a a desktop driving
• 36:05simulator
• 36:06or driving in the scanner.
• 36:07They got one or the
• 36:08other.
• 36:11So here's the person practicing
• 36:13flat on their back so
• 36:15that they can drive flat
• 36:16on their back, which they
• 36:17do in the scanner.
• 36:19This is the desktop driving
• 36:21simulator.
• 36:22This is someone using the
• 36:23driving simulator.
• 36:25This
• 36:26is like a Burger King
• 36:28extractor fan. So when subjects
• 36:30are getting high,
• 36:31they breathe out and the
• 36:33cannabis vapor disappears
• 36:35so that our staff don't
• 36:36become intoxicated.
• 36:37This is the driving software
• 36:40that's,
• 36:41basically fiber optics, so it's
• 36:43completely compatible with the scanner.
• 36:45So this just is our
• 36:47old scanner, but shows you
• 36:49the basic design,
• 36:51steering wheel,
• 36:52adjustable
• 36:53gas and
• 36:55brake pedals.
• 36:58And
• 36:59we do infrared eye tracking.
• 37:02So if someone crashes into
• 37:03another vehicle,
• 37:04the question is, what were
• 37:06they looking at? Were they
• 37:07looking at the car in
• 37:08front, and were just completely
• 37:10oblivious to the fact that
• 37:11it wasn't car in front?
• 37:13Or were they so stoned
• 37:14they were looking up at
• 37:15the virtual sky and saying,
• 37:17oh, what pretty virtual clouds?
• 37:19We wanna know what they
• 37:20were looking at.
• 37:23And the driving paradigm, as
• 37:24I said a little earlier,
• 37:25just to elaborate on that,
• 37:27lane keeping, you're driving on
• 37:28a straight highway. There are
• 37:30wind gusts. You've got to
• 37:31stay in lane.
• 37:33Car following, you're following your
• 37:35lead vehicle.
• 37:36It speeds up and slows
• 37:37down.
• 37:38You have to
• 37:40stay a safe distance from
• 37:41it, a fixed distance.
• 37:43And overtaking,
• 37:44you're driving on a two
• 37:45lane highway. The car in
• 37:47front of you breaks down.
• 37:48There's oncoming traffic you've gotta
• 37:50drive past in safely.
• 37:52So these are increasing
• 37:53complexity.
• 37:57People
• 37:58rate themselves
• 37:59this is their highness over
• 38:00time, so
• 38:02higher at the high dose,
• 38:04highest a few minutes after
• 38:06smoking,
• 38:07decreases thereafter.
• 38:10This these are
• 38:12blood THC and metabolite,
• 38:14values that also spike early
• 38:17and decrease
• 38:18coming down,
• 38:20more slowly for the eleven
• 38:22hydroxy
• 38:23eleven carboxy,
• 38:25measure.
• 38:26The impairments on multiple aspects
• 38:29of driving,
• 38:31these are our nineteen
• 38:32submeasures,
• 38:33and many of them were
• 38:35significant.
• 38:39And the only thing I'd
• 38:40emphasize here is that the
• 38:41high dose, some of those
• 38:43like lane keeping impairments,
• 38:45were still obvious at five
• 38:46hours after dosing,
• 38:48less so after the low
• 38:49dose. So these things are
• 38:51remarkably persistent,
• 38:53for many of the measures.
• 38:58Okay. I don't wanna dwell
• 38:59on that.
• 39:01The peak blood and saliva
• 39:03levels were dose related,
• 39:05but only at peak.
• 39:07The behavioral worsening was completely
• 39:09unrelated to blood or saliva
• 39:11metabolite or THC
• 39:13complicate
• 39:14concentrations,
• 39:16so big implications for per
• 39:18se laws.
• 39:19And if you want me
• 39:20to
• 39:21give you a one sentence
• 39:22summary of what all those
• 39:23impairments were, there was a
• 39:25slight generalized slowing in behavioral
• 39:27response. People are just a
• 39:29bit a little bit slower
• 39:30to respond.
• 39:33And although,
• 39:35people were impaired,
• 39:36two thirds of them reported
• 39:38willingness to drive across the
• 39:40entire time
• 39:41despite their awareness of being
• 39:43subjectively impaired
• 39:45and objectively worse driving performance.
• 39:47So that's concerning.
• 39:49People know they're driving worse.
• 39:50They are driving worse,
• 39:52but still in all, they
• 39:54are willing to drive.
• 39:56What's happening in the brain?
• 39:59A bunch of regions
• 40:01over the entire thirty minutes
• 40:04do change from normal.
• 40:07Many of those regions are
• 40:08correlated with altered driving, but
• 40:11you can see there are
• 40:12little itty bitty regions.
• 40:14So what we did is
• 40:16use something more sophisticated.
• 40:18We used graph theory metrics.
• 40:20So if you think about
• 40:22people
• 40:24if you think about airports,
• 40:26airports
• 40:27are in hubs. Those hubs
• 40:29hubs connect to many other
• 40:31routes.
• 40:32Some of the hubs connect
• 40:33to a few routes. Some
• 40:35of the hubs,
• 40:37connect to many routes. And
• 40:38you can look at brain
• 40:40connections,
• 40:41during functional MRI in exactly
• 40:42the same way. And when
• 40:44we looked at that, many
• 40:45of these graph theory metrics
• 40:47were altered
• 40:48compared to controls.
• 40:50And if you look at
• 40:51the hub that was most
• 40:52affected,
• 40:53that was in the hippocampus,
• 40:55and the hippocampus is full
• 40:57of CB one receptors.
• 40:59So that's not entirely surprising,
• 41:01but
• 41:02that gave us much more
• 41:03sophisticated
• 41:04and much more significant results
• 41:06than other ways of looking
• 41:07at this. So that's all
• 41:08I wanna say about driving.
• 41:12Last thing I wanna talk
• 41:13about is our roadside sobriety
• 41:16testing battery,
• 41:18that came,
• 41:19as part of research we're
• 41:20doing for NHTSA.
• 41:22And NHTSA has embargoed embargoed
• 41:25the final data, so I
• 41:26can only tell you in
• 41:27general terms what we found.
• 41:29But this week, we are
• 41:31sending them our
• 41:32findings for a publication.
• 41:35So those data should be
• 41:36available
• 41:37publicly
• 41:38reasonably soon. Stay tuned.
• 41:41So there, we had subjects
• 41:43live on a CRC
• 41:45for five days
• 41:46to control our access to
• 41:48all other substances, including their
• 41:50personal supply.
• 41:52And every morning, they came
• 41:53to us,
• 41:54and they got either THC
• 41:56or placebo in the paradigm
• 41:58I told you about. A
• 41:59single acute dose
• 42:01from a desktop simulator,
• 42:04high or low dose or
• 42:05placebo.
• 42:06Two days high, two days
• 42:08low, one day placebo, random
• 42:09assignment, double blind,
• 42:11multiple tasks, multiple blood levels,
• 42:14including,
• 42:16virtual driving.
• 42:17So this is the design.
• 42:21No scanning,
• 42:23battery of tests given
• 42:25at these various times,
• 42:28with regular blood
• 42:30and oral fluid sampling exactly
• 42:32as in the NIDA study.
• 42:33So that's not shown here.
• 42:35So, again, twenty one subjects
• 42:38completed
• 42:39out of twenty seven. So
• 42:41we had twenty seven participants
• 42:43worth of data, but twenty
• 42:44one complete.
• 42:47And this is the test
• 42:49battery. I don't wanna say
• 42:50much more about that
• 42:51other than they got the
• 42:52alert meter, which is a
• 42:54comp commercial device,
• 42:56a bunch of tests from
• 42:57the COG state battery,
• 42:59a bunch of tests that
• 43:00we designed
• 43:02looking at time reproduction
• 43:04and time estimation.
• 43:06We used the Anam Pursuit,
• 43:09just,
• 43:10pursuit tracking,
• 43:12a number of tasks from
• 43:13the Druid, so roadside
• 43:16sobriety tests, walking on a
• 43:18straight line, standing on one
• 43:19leg,
• 43:22Ramaker's
• 43:23critical tracking task,
• 43:25and body sway,
• 43:27from an iPad. We have
• 43:29a very sophisticated body sway
• 43:30measurement.
• 43:32So
• 43:33our combined predictive ability using
• 43:35multiple tests
• 43:37rather
• 43:38than one task,
• 43:39looking at sensitivity and specificity,
• 43:41and I'm
• 43:42can't tell you what the
• 43:43tasks were. Some of the
• 43:45combos gave us really good
• 43:47sensitivity,
• 43:48specificity
• 43:49trade offs in terms of
• 43:51AUCs equivalent to about point
• 43:52nine, which is in the
• 43:54excellent range.
• 43:57And
• 43:58the tasks that did best,
• 44:00some of them did awfully,
• 44:02like no better than chance.
• 44:04The test that did best
• 44:05that we selected
• 44:06correlated
• 44:07highly
• 44:08with impaired driving on the
• 44:10simulator,
• 44:11so they actually have criterion
• 44:12validity,
• 44:13which was crucial for us.
• 44:16So they're sensitive, they're reliable,
• 44:18they have predictive ability,
• 44:19and they have criterion validity.
• 44:23We then picked the three
• 44:25best performing tasks from that
• 44:26study, which I can't tell
• 44:28you what they were,
• 44:29and retested them in thirty
• 44:31one new subjects,
• 44:33in a one day study
• 44:34where we gave everyone placebo
• 44:36first and then the high
• 44:37dose only and only used
• 44:39the lane keeping task for
• 44:41driving.
• 44:42And within that restricted,
• 44:45replication,
• 44:46we got really good replication
• 44:49from that of that first
• 44:50study.
• 44:53So the overall
• 44:55the overall summary of everything
• 44:58is there are quantitative effects
• 45:00of a QTHC on motor
• 45:01vehicle accident risk.
• 45:03They're subtle. They're like a
• 45:05tenth of what you see
• 45:07with alcohol
• 45:08at the point one,
• 45:10BAC range.
• 45:13Field sobriety tests for cannabis
• 45:14impaired driving are a work
• 45:16in process,
• 45:17but you need a battery.
• 45:18No one task is sufficient.
• 45:21Everything should be validated against
• 45:23simulated driving.
• 45:25We found and others have
• 45:27found no clear relationship of
• 45:29impairment
• 45:30to THC,
• 45:32metabolite in blood or saliva
• 45:34levels
• 45:34with big implications for per
• 45:36se laws.
• 45:37And as yet, there's no
• 45:39BAC for THC.
• 45:43We need roadside tests that
• 45:44are cannabis specific
• 45:48that don't depend on knowing
• 45:49the subject's baseline performance. So
• 45:51you need
• 45:52very large normative data across
• 45:55different genders,
• 45:56across different age ranges
• 45:58that, again, relate to marijuana
• 46:01impaired driving.
• 46:03And
• 46:04many of our cannabis using
• 46:05subjects,
• 46:06get buzzed,
• 46:08with alcohol at the same
• 46:09time as they get high.
• 46:11So assessing the combinations are
• 46:13crucial.
• 46:17No one's really tested edibles.
• 46:19That needs to happen.
• 46:22And there's a
• 46:24a problem
• 46:26with testing older subjects.
• 46:28No one has really looked
• 46:29at older subjects
• 46:30to see to what extent
• 46:31they differ from younger subjects.
• 46:33So this is our team.
• 46:36And,
• 46:38so we have a number
• 46:39of nurses and,
• 46:41senior collaborators
• 46:44and,
• 46:45statisticians
• 46:47and
• 46:48imagers
• 46:49and clinical research associates,
• 46:52all of whom collaborate, and
• 46:54we collaborate with doctor D'Souza
• 46:56increasingly on a lot of
• 46:57the work that we plan.
• 47:00And just a brief
• 47:02self serving shout out,
• 47:04to my book. So I
• 47:07think I'll leave things be
• 47:09at that point. We've got
• 47:12about fifteen minutes to answer
• 47:14questions, so I'm happy to
• 47:16take any questions that people
• 47:19have.
• 47:22Thank you so much, Doctor.
• 47:23Carlson,
• 47:24for such a wonderful and
• 47:25thoughtful presentation. We have several
• 47:27questions in the chat.
• 47:29And please feel free if
• 47:30I'm reading your question you'd
• 47:31like to unmute.
• 47:34Bruce Parker has asked about
• 47:35a standard test, which might
• 47:37be useful.
• 47:38Is the reaction time to
• 47:40lift off an accelerator and
• 47:41place the foot on a
• 47:42brake,
• 47:43I think, in the driving
• 47:44simulator?
• 47:46Yes. We we do measure
• 47:48that, and that's very slightly
• 47:50slowed,
• 47:51but it's not one of
• 47:53our better performing tasks.
• 47:56That's
• 47:58we're all in favor of
• 47:59of simple
• 48:02tasks, but that's not one
• 48:03of the more reliable ones.
• 48:06There's very wide baseline variability,
• 48:09and it's not particularly impacted,
• 48:12which makes it
• 48:14a reasonable sounding choice, but
• 48:15not one that's good in
• 48:17practice.
• 48:22And Steven Holt asked if
• 48:23there's any benefit to serial
• 48:24testing of THC.
• 48:26So he says that is,
• 48:27does the rate of change
• 48:28in serum THC level over
• 48:30sixty minutes allow you to
• 48:32extrapolate
• 48:32recency of use?
• 48:35So,
• 48:37if if I'm understanding that
• 48:38right, is should should we
• 48:40be dozing people serially
• 48:42through through the task?
• 48:45Like, giving them repeated doses
• 48:47of THC rather than just
• 48:48one acute dose and then
• 48:49seeing what happens? No. When
• 48:51I had asked the question,
• 48:52I, which was a while
• 48:53ago earlier in your talk,
• 48:54I was asking if if,
• 48:55you know, when when law
• 48:56enforcement takes somebody and and
• 48:58decides that they have a
• 49:00normal blood alcohol level, they're
• 49:01still intoxicated, so they send
• 49:03them to get blood work.
• 49:04If they were to do
• 49:05blood work, you know, three
• 49:07times over sixty minutes, would
• 49:08the rate of change of
• 49:10their THC level,
• 49:12enable them to extrapolate how
• 49:13recently they used cannabis?
• 49:16It would if the person
• 49:18was driving along
• 49:20and taking tokes off a
• 49:22joint while driving,
• 49:24and they were caught in
• 49:25the first five minutes after
• 49:27a dose. Right.
• 49:29But not otherwise.
• 49:31Understood.
• 49:32But, yes, that's a reasonable
• 49:34question. If the person's actually
• 49:36driving and smoking,
• 49:37like the
• 49:38dude, in the picture that
• 49:40I showed you, that that
• 49:41would be valuable, but not
• 49:42otherwise.
• 49:50And Aly Scott asked many
• 49:51of the findings suggest that
• 49:53THC consumption leads to statistically
• 49:56significant worse functioning
• 49:58in many of the tasks
• 49:59that were examined. Can you
• 50:00comment on the size of
• 50:01these effects and whether they're
• 50:02noticeable
• 50:03or practically
• 50:05significant degrade in driving performance.
• 50:08I think our question is
• 50:09about effect size. Yeah. In
• 50:10terms of effect sizes,
• 50:12most of these are moderate,
• 50:15at best.
• 50:17So,
• 50:19I guess the best comparator
• 50:20is where we've given people,
• 50:23alcohol
• 50:24sufficient to get them to
• 50:25a blood alcohol concentration
• 50:27of zero point zero five
• 50:28or zero point one.
• 50:31So they're
• 50:32roughly equivalent in terms of
• 50:33effect sizes of what we'd
• 50:35see
• 50:37with alcohol at about zero
• 50:38point zero five.
• 50:40And the the best
• 50:42the the most impairing ones
• 50:44are about what we'd see
• 50:45at zero point zero eight,
• 50:47but almost none of them
• 50:49make it to that level.
• 50:50So they're they're modest.
• 51:00There's also a practical question
• 51:02about how long the saliva
• 51:04testing is positive for THC
• 51:06versus urine.
• 51:08Oh, versus urine?
• 51:10Yeah. Yeah. Because
• 51:16because THC has to leech
• 51:18out of fat slowly
• 51:21and takes a while to
• 51:22appear in urine,
• 51:24at least in our experience,
• 51:26saliva testing,
• 51:28has a shorter window than
• 51:30urine testing.
• 51:32But
• 51:33to also to be honest,
• 51:35we've not looked at either
• 51:37for, like, several days after
• 51:40testing,
• 51:41which is really what's relevant.
• 51:43So,
• 51:44in
• 51:45I'm not able to answer
• 51:47that from our experience
• 51:49for multiple days after an
• 51:51acute dose.
• 51:57Anyone else?
• 51:58Mohini, can you address that?
• 52:00Hi, Godfrey. Thank you so
• 52:01much. Such a such an
• 52:03excellent and timely talk. I
• 52:04have I have a new
• 52:06driver at home, so he
• 52:07hears about this,
• 52:09all the time, and I
• 52:10should have invited him to
• 52:12this to this talk. I
• 52:13have a question, which as
• 52:14you know is one of
• 52:15my favorite questions. Can you
• 52:16comment a little bit about,
• 52:18sex differences as it relates
• 52:20to either performance
• 52:22or detection,
• 52:23in the context of driving
• 52:25performance?
• 52:29We don't see marked sex
• 52:30differences
• 52:32compared to the person's baseline
• 52:33driving.
• 52:35And
• 52:36we don't see
• 52:37marked sex differences in how
• 52:39high
• 52:40people
• 52:41rate themselves,
• 52:42but we do see differences
• 52:43in blood levels
• 52:45of THC and its metabolites.
• 52:47So it's hard to square
• 52:48those things.
• 52:50But, yeah, what what what
• 52:51do you see? Because you've
• 52:52looked at this more carefully
• 52:54than we have.
• 52:56So in in general, I
• 52:57think what and that I
• 52:58think that might be also
• 53:00what what you're finding is
• 53:01that,
• 53:03women seem to kind of
• 53:05experience
• 53:06the effects of THC even
• 53:07at a lower level. So
• 53:08they don't they don't necessarily
• 53:10need a very high
• 53:11dose of what is being
• 53:12ingested to achieve that goal
• 53:14of,
• 53:15feeling high, and that could
• 53:17be related to differences in
• 53:19metabolism
• 53:19as well.
• 53:21And
• 53:22what we have also seen
• 53:23in chronic cannabis users is,
• 53:27for instance, in one of
• 53:28Cyril, D'Souza's studies,
• 53:31when we had people try
• 53:33to,
• 53:34abstain,
• 53:35it took
• 53:36some women much longer for
• 53:38the urine to test negative
• 53:41compared to,
• 53:43men who were using roughly
• 53:44the same amount. And so
• 53:47I'm just wondering what impact
• 53:49that might have in detection
• 53:51of THC
• 53:52in someone who might be,
• 53:54you know, pulled over for
• 53:56for driving,
• 53:58problems. Yeah. That's really important
• 54:00to know, actually, and that's
• 54:01not taken into account in
• 54:03per se testing at all.
• 54:07Thanks.
• 54:10Thank you.
• 54:14Doctor. Carlson, can you comment
• 54:15on how policymakers,
• 54:17legal professionals, and law enforcement
• 54:19officials could better educate themselves
• 54:21on the current scientific evidence
• 54:23regarding cannabis use driving impairment
• 54:25to inform
• 54:27reasonable policy development, legal standards,
• 54:29and enforcement practices?
• 54:31It's a big question.
• 54:32So,
• 54:34unfortunately,
• 54:35that happens
• 54:36state by state at the
• 54:38moment.
• 54:39But as far as I
• 54:40know, no one's testified before
• 54:42congress
• 54:43on,
• 54:45how laws should be formulated
• 54:48at a national level. And
• 54:50in part, that's due to
• 54:51the confusion
• 54:52over
• 54:55whether cannabis should be legalized
• 54:57or decriminalized
• 54:58or reclassified
• 55:00or unclassified
• 55:01at a national level.
• 55:04So there's
• 55:05confusion at so many levels.
• 55:07It's hard to get
• 55:11a comprehensive national policy
• 55:14formulated.
• 55:15But, locally,
• 55:17legislators
• 55:18seem pretty,
• 55:21willing to reach out to
• 55:22cannabis researchers. That's happened, I
• 55:25I know, in the state
• 55:25of Connecticut.
• 55:28And Connecticut chose not to
• 55:29have per se laws, at
• 55:31least until this time, explicitly.
• 55:35And I've helped.
• 55:37I think can Cyril and
• 55:39others have also helped, like,
• 55:40lobby for
• 55:42some of the legislation.
• 55:46But legislators
• 55:47should feel welcome to reach
• 55:49out to cannabis
• 55:51researchers, get a variety of
• 55:52opinions.
• 56:03And any other questions in
• 56:05the chat?
• 56:07Yeah. We have, one Ellie
• 56:09Scott just asked a follow-up
• 56:10about. Has there been any
• 56:11research looking at the relationship
• 56:12between drivers' self efficacy for
• 56:14driving under the influence
• 56:16and whether confidence mediates performance
• 56:18outcomes?
• 56:21We found
• 56:23almost
• 56:25no relationship between those things,
• 56:28at least on subjective measures,
• 56:31that some people think they're
• 56:33driving perfectly fine when they're
• 56:35not.
• 56:36But what what is truly
• 56:37interesting is
• 56:39compared to alcohol impaired drivers,
• 56:43it's very difficult if you're
• 56:45alcohol intoxicated
• 56:46to make yourself sober up
• 56:48and snap out of it
• 56:50despite trying your best.
• 56:52Whereas some,
• 56:54cannabis impaired drivers
• 56:56seem to be able to
• 56:57do that. We've not tested
• 56:59that explicitly,
• 57:01but that's what our
• 57:04subjects tell us,
• 57:06for what that's worth. Some
• 57:08people say, well, I was
• 57:09really stoned. I was pulled
• 57:10over by a cop,
• 57:12and I managed to just
• 57:13pull myself out of it
• 57:14and pass the sobriety test,
• 57:16given the problems with sobriety
• 57:18tests.
• 57:20So to that extent,
• 57:22some people do seem to
• 57:23be able to
• 57:24get themselves
• 57:27sober by effort of will
• 57:28and with cannabis in a
• 57:30way they're not able to
• 57:31with alcohol.
• 57:32And that's a curious understudied
• 57:34phenomenon that I don't quite
• 57:35understand.
• 57:46I think we might have
• 57:47time for one more question.
• 57:57Alright. It looks like doctor
• 57:58Ranganathan,
• 57:59got that in right under
• 58:00the wire.
• 58:02She's asking if we've heard,
• 58:04unless you'd like to ask,
• 58:06that with dry January there's
• 58:07increased cannabis use. Does this
• 58:09translate in some way to
• 58:10different rates of DUIs?
• 58:13That's a really interesting question.
• 58:15I don't know the answer
• 58:15to that.
• 58:16I I can,
• 58:18I can reach out to
• 58:19some of my epidemiology
• 58:21statistician friends and find that
• 58:24out,
• 58:25but, yes, that's
• 58:26logically the the there may
• 58:28well be, but I've never
• 58:29seen that documented?
• 58:35And,
• 58:37doctor Prosen, if anybody has
• 58:39any follow-up questions,
• 58:43to reach you?
• 58:44Just email me at godfrey
• 58:47dot pearlson at yale dot
• 58:49e d u,
• 58:50and I'm happy to send
• 58:52people
• 58:53either
• 58:54papers,
• 58:55from the
• 58:56literature or answer questions directly.
• 58:58I'm very happy to do
• 59:00that.
• 59:01Thank you. We're adding the
• 59:02email to the chat for
• 59:03anybody who is Oh, okay.
• 59:05Thank you. You back. Appreciate
• 59:06that.
• 59:08Oh, I also have a
• 59:09TED talk on stone driving
• 59:11if anyone wants to listen
• 59:13to that. Oh, sounds great.
• 59:18Thank you so much, doctor
• 59:19Carlson.
• 59:20You're welcome. Thank you. Thank
• 59:21you. Thank you.