Yale Psychiatry Grand Rounds: November 6, 2020

November 06, 2020

Information

Dean Mobbs, PhD, Professor of Cognitive Neuroscience, California Institute of Technology

"Space, Time, and Fear: Survival Decisions Along Defensive Circuits"

ID5851

To CiteDCA Citation Guide

00:00It's a real pleasure to welcome my friend
00:05and colleague Dean Bobs from Caltech.
00:08Dean is a professor in the chance enter.
00:13In humanities and social Sciences
00:15in computational neural systems
00:17program at Caltech, he did his PhD
00:20with Chris Frith and Ray Dolan.
00:22So we share a certain degree
00:24of academic heritage.
00:25He and I and we also overlapped in Cambridge,
00:29where we became friends.
00:30One of the things I really admire
00:33about Dean is that he was a
00:35really early adopter of bringing
00:37principles of behavioral ecology
00:39to human cognitive neuroscience,
00:41and I think we're going to hear
00:43a lot about that today.
00:45Some other notable things.
00:48Dean began his career before he
00:51became a committed neuro scientist
00:54as a painter and decorator,
00:57which is rather an unusual career path.
01:01And another notable thing is that
01:03Dean's grand rounds is actually
01:04a rescheduled grand rounds.
01:06He was meant to be with us in New
01:09Haven on the day that everything
01:11shut down in March.
01:13This exchange between Dean
01:14and I began on Twitter,
01:16where I invited him to come speak
01:19to us because he saw a picture of
01:21some Donuts that I bought from Donut
01:24crazy and the hope was that he get
01:27to share those when he came here.
01:29We haven't quite managed that,
01:31we rescheduled.
01:31Today in the hope that we
01:33still get to do that,
01:35and maybe we will still in the future,
01:37but I'm sorry Dean.
01:38It's so early for you and
01:39no Donuts this time around,
01:41but we're delighted to have you.
01:43We can't wait to hear your talk,
01:44so thank you.
01:45Please take it away.
01:48Great, thanks for.
01:50Thanks for that wonderful introduction.
01:53So what I'm going to talk about today is
01:57really the some of the older than some of
02:00the recent research that we've put forward.
02:04Looking at how we can separate fear and
02:07anxiety circuits in the brain using.
02:09Tasks the altar in the spatial
02:12temporal frequency of the threat
02:15of the proximal or distal distance
02:17to a threat and As for dimension.
02:21This is really been influenced by a lot of
02:24work from the field of behavioral ecology,
02:28so I'm going to talk a bit about today's some
02:32theoretical background about really well.
02:36Why I got to this point and give
02:38you some background in some of the
02:41more contemporary theories of how we
02:44really think about fear and anxiety.
02:46So I want to start by thanking all of
02:50these individuals in these funding
02:53bodies for their collaborations
02:56and friendships over the years,
02:59and particularly with colleague demo
03:02service that Google Deep Mind we
03:06still collaborate and where we began
03:09to work together on some of these
03:13early studies of fear and anxiety.
03:16My lab.
03:17Which was originally at Columbia University
03:20law move to Celtic four years ago.
03:23Song she only took mainly about his work.
03:27Along with these collaborations or
03:29collaboration won't fund my original
03:32PhD supervisors Chris Frith and.
03:34Write down in my again another
03:36long term collaborator.
03:37I talk a lot about some of our theories,
03:40which is Peter, Diane and then some.
03:42My colleagues from the field would be able
03:45to colegi and colleagues from Princeton,
03:47particularly Daniel Door
03:48Bensimon typically shut.
03:49That came to University,
03:50ought to acknowledge people before gets tax.
03:53I think they're playing important role in
03:56the research that I'm going to talk about.
03:59And of course my lap.
04:01Now that I'll talk about some of the
04:04more recent work that we've done.
04:07OK,
04:07So what I want to do is sort of
04:10start by going over the really the
04:14the keyholder core sort of theories
04:17of how we define an approach fear.
04:20In in the modern era,
04:22and there's lots of theories out there.
04:25We can go back to many old theories
04:28of fear and anxiety,
04:30but there's been a real burst of theoretical.
04:33Approaches to failing sides.
04:35This is really captured in this article
04:38that I was the moderator on an edge.
04:40New songs that came out last year.
04:43So although recommend if you guys are
04:45interested in trying to delve into these,
04:48there is a little bit more than I
04:50do recommend reading this paper.
04:52You can get a link to the PDF on my website.
04:56Which will advertise at the end as well.
05:00So really, there's been.
05:03So that four key contemporary theories.
05:08We can really go from left to right
05:11there where we have two extremes,
05:13many of yak panksepp and what planks
05:15that proposed was that there's really a
05:18dedicated set of evolved circuits that
05:20are associated with different emotions
05:22which you refer to is primal emotions.
05:24What we're interested in here,
05:26and really 2 which is fear and panic.
05:29But he also said that there's
05:31a place circuit and out wide
05:33rage circuits you can circuit.
05:35The circuit can care circuit and this is
05:38really kind of the extreme in terms of.
05:41All this overnight circuits
05:42that we believe exist,
05:44or at least some of them exist in the
05:47animal and maybe in the human brain as well.
05:51If we shift all the way
05:53over to the other side,
05:54we can see at least a phone by its
05:57conceptual act theory where she says
05:59there's actually no dedicated circuits.
06:01These circuits that we owe these brain
06:05regions that determine different
06:06emotions are not universal and this
06:09is conglomeration of brain regions
06:12that combine to create emotion and
06:14and what Lisa is saying is that
06:17emotions are created in the cortex.
06:20The output is through motor systems and.
06:23I related systems and what we have is
06:26is called across circuit that makes up
06:28the emotion in the moment and this is
06:31an interesting analogy she uses for this is.
06:34To say that if you take eggs,
06:38flour, sugar, salt,
06:39I'm not the cook as you can tell,
06:42but you mix them in different proportions.
06:45You can create different types of foods.
06:48You can have pancakes, bread and so on,
06:52so this actually argues that
06:54our emotions are created,
06:56particularly in the perceptual realm,
06:58and particularly in humans.
07:00So what you have here then,
07:02is these two distinct sort of theories.
07:05One where there is a dedicated
07:07hardwired circuit, one where there's
07:10no dedicated wide second OK.
07:12Now if you go over to fans Lowe's
07:15theory which you can see in the green.
07:19Fans, though,
07:20argues is a set of this.
07:22Defensive behaviors are organized around
07:25distinct set of modes of danger or what
07:29he calls levels of threat imminence.
07:32I'll come back to that in a minute
07:34because this is really the model that's
07:36been the most influential to me because
07:38it has one foot in the College in
07:40one foot in behavioral neuroscience.
07:42And then another recent model
07:44is by Journey do and AMP line,
07:46which they call the two systems theory.
07:49Well,
07:49what you have is a defensive set
07:52of defensive survival circus.
07:53When talking about fear now
07:56specifically focused on fear.
07:57You have a set of circus associated
08:00with survival behaviors and then on
08:03top of those circuits you have these I
08:05order mental or meta representations of fear.
08:08The conscious representations
08:09of fear and what joag is here,
08:12which has been a bit controversial is
08:15that fear is the conscious feeling
08:17that we have the behavioral output.
08:20The fight,
08:20flight freeze in behavior that comes
08:22out of the defenses survival circuits.
08:25Is more of a a pre programmed
08:30emotive response OK.
08:32So.
08:32The controversial thing for most
08:35people who work in the field
08:37of animal research is that.
08:40You cannot study fear in animals,
08:42because if fear isn't I order
08:45conscious representation of, say,
08:46your bodily state, for example,
08:48or the threats in the environment.
08:51You can't ask an animal if it feels fair.
08:54You can only measure these
08:56sorts of fearful self report,
08:58and it's therefore we can
09:00only measure them in humans.
09:02Or we can measure in animals
09:04is defensive Savarese Circus.
09:05The output of these circuits,
09:07in the form of fight flying,
09:09freezing behavior?
09:10I'm going to argue against that thing.
09:12I think it's.
09:13I agree with this separation between
09:15these circuits of the conscious state
09:17and that the behavior survival circuit
09:19output and fight flight freeze and behavior.
09:22So mentioned.
09:23But the definition the way defines fear I
09:26think doesn't really give us any further.
09:28I think that we need to have a more
09:31concrete representation of what fear is,
09:34and I think this fits.
09:35This is where fans those throughout the
09:38imminence continuum model comes in.
09:41In this paper,
09:42I wasn't really allowed to
09:44talk about my theories,
09:45'cause I was the moderator,
09:47but well,
09:48I'm going to talk about really
09:51over the course of this talk.
09:53Today is really the direction
09:55in which I've gone again,
09:57really influenced by Franz Liszt,
09:59but also your journey Deuce work and.
10:02You can read about that in a recent
10:04paper that came out earlier this year,
10:05and again, I'm going to talk
10:07about it anyway in this talk,
10:08but if you're interested in that paper,
10:09there it is.
10:11OK, so how do we define fear or we
10:14take more of a dimensional approach?
10:17If you go back to Darwin,
10:20Darwin proposed that fear takes
10:22the graduation from mere attention
10:24to extreme terror and horror.
10:26But what we know is that more
10:29contemporary theorists proposed to
10:31fear is an emotion that results in the
10:33presence of an oily, imminent threat.
10:35OK, if the threat is here in this room.
10:38Now,
10:39if it's a Tiger and it's sitting
10:41next to you and it's growling,
10:43that's going to evoke a fear response.
10:46Anxieties about when the stimulus is
10:48abstract or remote in time and space,
10:51it's about something that's
10:52going to happen in the future.
10:54OK, that's how we separate from anxiety.
10:57There's something terribly
10:58this happening now.
10:59Anxieties,
10:59something tell what's going
11:00to happen in the future?
11:02OK,
11:02So what you can see it all of a
11:04sudden is with these definitions is
11:06that we're looking at time and space.
11:09In some respects.
11:10We're looking at fear as a
11:11more proxamol threat.
11:12Anxiety is a more distal threat,
11:14so we've used those sorts of definitions
11:16to be able to separate fear and
11:19anxiety using the top stats that we use,
11:21and I'll be talking about in a minute,
11:23but they separate these emotions
11:25by creating stimuli or threats
11:27that will distort.
11:27Approximately.
11:30So let's go back to the drawing board.
11:34Let's try and understand the why.
11:36Fear and anxiety too important
11:39to survival in nature.
11:41And then our prize winner Nico
11:43Tinbergen propose it becomes difficult
11:45and even in some cases impossible
11:47to say where ethnology stops.
11:49A new Physiology begins.
11:50And in a recent paper that we published,
11:53review replies that we need to understand
11:56our fear as well as our decision processes.
11:59You must consider the evolutionary and
12:01ecological conditions that give rise to them,
12:03because if we don't do that,
12:06this will lead the field of
12:08affective neuroscience study affair
12:09ungrounded and they had to interpret.
12:11And the natural and one crate paradigms
12:14we need to create paradigms that relate
12:17to the similar types of decisions that
12:21would be executed in the real world.
12:23Just go to some of the theoretical background
12:27beyond some of the paradigms that we're
12:29going to show in this this talk today.
12:32So in an issue that mean Joe
12:35Ledoux edited couple of years ago,
12:37now concurrent opinion be able to
12:39Sciences I wanted paper looking at
12:41how we could separate fears and those
12:43fears and applaud sense based upon
12:45the conditions of the environment.
12:47So what I proposed in that paper was
12:50the first of all we need to understand
12:53the natural conditions of the Volk.
12:55Those survival behaviors, OK,
12:57and there under those natural conditions
12:59we can look at things like the traits
13:01of the threat, what type of threat.
13:04Is it on the table?
13:05Temporal spatial properties are threat
13:07we can think about other things as well.
13:10About is an environment where is
13:12open field or is it a forest?
13:14For example,
13:15we need to think about what
13:16are those natural conditions.
13:18And once we understand those
13:20natural conditions,
13:20we can then begin to understand what are
13:23the optimal strategies of that Organism.
13:25Two of aid that predator OK.
13:28So we can try and workout what
13:30those strategies are.
13:31The actions and reactions are to those
13:33natural threats in the environment,
13:35and once we understand the
13:37environment and the strategies,
13:38we can begin to understand the computations.
13:41And all the animals optimizes it behavior,
13:43but also we can think about our
13:45other Internet connected survivor
13:47circuits may modulate that.
13:48So for example,
13:49a strategy to a Veda predator
13:51will be different if you have to
13:53protect your offspring as well, OK?
13:56So we can think about our other survival
13:59circuits may impact these or survival
14:01guys may impact these strategies.
14:03And finally,
14:04once we understand those strategies,
14:06we can understand what parts of
14:08the brain are involved in those
14:11strategies and then apply computation.
14:13Must understand that more closely.
14:16So really,
14:16this is the general principle
14:18of the approach I've taken in my
14:20lab to try and think about when
14:22we develop our paradigms.
14:23How does that relate to the natural
14:25world and how can we create paradigms
14:27particularly overly that allow us
14:29to change your allow the individual
14:31in those environments to change
14:33their strategies this summer?
14:35We're currently working
14:36on in more depth as well.
14:38OK, so I've mentioned fans Lowe's that
14:42Eminence Continuum originally called
14:44a predatory imminence continuum,
14:46and what this proposes,
14:48is that there's really 3 core
14:51levels of threat imminence.
14:54We start off with this first one here,
14:56which we called the preferred activity.
14:57This is where the animal spends
14:59most of its time.
15:00This is where it fills the safest. OK.
15:02So if we're looking a little Birdy,
15:04are a little bird is in the nest OK?
15:06And it's the safest place at the bird
15:08is feels it could be I should say.
15:10If we move down here then we can
15:13see that the bird is fallen or
15:16gone to the bed of the forest.
15:19It OK.
15:19Now this is where there's a
15:21potential to actually encounter
15:23a threat in the natural world.
15:26They call this the pre encounter context.
15:29Just imagine now that little bird spots and
15:32natural predator like a cat for example,
15:34which we can see just there.
15:35Now we switch over into what's
15:37known as the Post Encounter.
15:39This is where there's a predator
15:41in the environment.
15:42There's no direct interaction
15:43between a predator and prey.
15:45Then finally, a predator wakes up,
15:47spots Little Bird, and begins to attack.
15:50We refer to this as the
15:52circus strike context.
15:53OK, now what's really interesting
15:55about these contexts?
15:56OK,
15:56they allow us to think about fear
15:58and anxiety slightly differently,
16:00particularly in the context
16:02of natural behavior.
16:03But also what we know is that
16:05the behaviors of the strategies
16:07alter across these different
16:09levels of threat imminence.
16:10What you typically see
16:12in pre encounter OK is.
16:14Cautious behaviors and otherwise
16:16you see such things as well as
16:19known as intermittent locomotion's
16:20the other will move along.
16:22Look around OK and keep doing
16:25these voluntary sort of pauses OK,
16:27you also see thicker mataxis Indiana
16:30artificially invites with the
16:32animal will go around the borders
16:34of the environment, but also.
16:36You see such things as increased vigilance.
16:39I doubting, for example,
16:40they're trying to increase their
16:41vigilance to detect approach to
16:43before the predator detects them.
16:44OK, so you see these cautious behaviors.
16:48If we switch over into the posting
16:50counts at the moment that are
16:52priced spots in natural product.
16:54The classic response that you
16:56will see is freezing behavior.
16:58And then once the process begins to attack,
17:00the animal will continue to
17:02freeze for awhile,
17:03but as this threat comes closer,
17:05you'll see switch into flight behavior
17:07and therefore there's even the last
17:09thing they can do if it strapped,
17:11for example, is it?
17:12We're going to fight behavior.
17:14So what you can see is the spatial
17:17temporal distance, the context.
17:18Will revoke different defensive behaviors.
17:20OK,
17:21so our goal then is based off
17:23of this model is can we create?
17:27OK can we create?
17:28Experiments that model these different
17:31environments and different behaviors.
17:33It's tough because if you're using
17:35classic Pavlovian conditioning or
17:36fear conditioning paradigm,
17:38you cannot do these types of experiments.
17:40At least you cannot do them in the
17:43way that we can get behavior OK.
17:48So why does this matter in some respects
17:51to definitions of fear and anxiety?
17:52Well, if we look at the top here in eight,
17:55again, this is taken from the paper
17:57I just mentioned in ticks that we
17:59probably should earlier this year.
18:01If we look at the top here,
18:03we can see fans Lowe's.
18:05So remanence continue myrene and
18:07we go from the the green here,
18:09which is a safe all the way
18:11through to circus strike.
18:12Over here we also talk here about
18:14these more ambiguous sort of zones
18:16where you get in this sort of switch
18:18order event defensive transition zone.
18:20The reason why we mentioned that
18:22in this in this model is because we
18:24think this is a good point at which
18:27you should be studying in clinical
18:29disorders such as general anxiety which
18:31we're working on at the moment an.
18:34What we've kind of show me Trey
18:36anxiety at least,
18:37and some of our preliminary day with.
18:40General anxiety is that this which seems
18:42to be earlier in those individuals or oily,
18:45anxious, OK,
18:46so this zone gets a little bit larger.
18:49OK, in those individuals.
18:50If we go down to beer though,
18:53the important part really of this
18:55slide is that what we can begin to
18:58do is using this time and space to
19:01threat is to think about theoretically.
19:03Think about how we can separate fear
19:05anxieties into different subcomponents,
19:07and how we shift from more sort
19:09of motor programs from fight,
19:11flight freezing behavior all the
19:13way through to a cognitive process
19:15is so if we start at the end here,
19:18yeah,
19:18this started circa strike here
19:20in the dark pink OK.
19:22When a sweat is very close OK,
19:24we begin to use lesser cortex and
19:26this may be the same in animals
19:28and we begin to just rely more
19:30on our sort of innate reflexive
19:32reactive responses and this is
19:34observed clear clearly in disorders,
19:36such as panic when a threat is
19:37very close to a pray will go into
19:40some form of uncoordinated flight.
19:42They'll just bump all over the
19:44walls and so on,
19:45just they just want to get
19:47away from this thread.
19:48Let's win the threat is very, very close.
19:51But if it has a little bit more time.
19:54It can direct its fear in a way
19:57that allows it to optimize it.
20:00Reactive response,
20:01whether it wants to go into fight
20:04response or freezing response OK.
20:06If you give the sweat and little
20:08bit more distance from the from
20:10the the pray you begin to see
20:12what we call cognitive fear.
20:14This is where now the animal's a
20:16little bit more time to make decisions
20:18about where it's going to escape.
20:20OK, it can appraise its behaviors,
20:22it can strategize its escape and so on.
20:26Now,
20:26what's important here,
20:27if you focus here on the pink,
20:30they match up these colors the
20:32way that we define fear.
20:33OK is when you're under
20:35potential or you are under
20:37attack from a threat.
20:38So we define fear by the context
20:41in which the Organism is in OK.
20:44We don't define it by the conscious
20:46state as the dude as we find it.
20:49We define it by the conditions in which
20:52the Organism is in the same of anxiety.
20:54So if we go over then to do
20:56post encounter throughout,
20:57there's a threat present,
20:59but not attacking.
21:00We call this encounter anxiety OK,
21:02and again we're seeing now,
21:03but more of the cortex and cognitive
21:05conscious processes coming online
21:07because there's more time to think, OK?
21:09The animal can begin to think about
21:11what's the best action and the
21:13outcomes of the action in the future.
21:15Again, it can strategize and
21:17it can go into these.
21:19Free app to determine its freezing
21:21state or when it's going to flea,
21:23and again with the cognitive
21:24fear and encounter.
21:25I should mention,
21:26this is where we begin to get
21:28more of the feeling state OK,
21:30you have this perception
21:32that we owned by the state.
21:34When were in pre encounter we
21:36have these anticipatory anxiety.
21:37OK this is where we get the
21:39cautious behaviors OK prospection
21:40again strategizing worry something
21:42may come in the future so we
21:44can begin to worry about it.
21:45And then if we're looking at days,
21:48weeks or even months away from a threat,
21:50we're going to intermittent locomotion.
21:51This is where we're probably
21:53thinking about it.
21:54Talk a big talk that we have to
21:56give in a months time and every
21:58now and again we think about it.
22:00We get a bit of intermittent anxiety.
22:03We think about it, but.
22:04Quickly we can suppress those feelings
22:07because we know it's in the future.
22:10OK,
22:11now I'm not saying these are all independent.
22:15Aspects of these emotions to
22:16me there more of a continuum,
22:19and they're defined by the amount
22:21of time in which the Organism
22:23has multiple organ.
22:24We talk about humans in general,
22:26but potentially other animals.
22:28How much time that we have
22:30to anticipate that threat?
22:31OK, so as you can see,
22:34is that what we suggest is that
22:36if you can change the time
22:38and space to the threat,
22:40you can begin to separate these
22:43different components of fear and anxiety.
22:45I'm going to focus a little
22:46bit already today is is escaped
22:48decisions in our flight initiation,
22:50distance task and really show how
22:52time can have an effect upon these
22:54differences between cognitive,
22:55reactive fear.
22:58OK, So what are these strategies?
23:00If we go back to our model and we talked
23:03a little bit about these strategy,
23:06we talked about the contexts
23:07already and fans. I saw him in it.
23:10Spell immigrants continuum.
23:11So it's very early for me.
23:13I'm still not drinking my coffee.
23:15And what we've proposed at the cross these?
23:19Levels of threaten image.
23:21There's five key survival strategies that
23:24humans and potentially other animals perform.
23:27We call them prediction strategies,
23:29protection strategies or prevention
23:31strategies, threat oriented strategies,
23:33threat assessment strategies,
23:34and finally reactive defensive strategies.
23:36OK, so we take all of these strategies
23:40and we put them into a model.
23:43It looks something like this into
23:46a classic kind of box model so.
23:49If we look here on the left,
23:52we can see fans Lowe's Eminence
23:55context switch,
23:56go from safety Pre Encounter Post
23:58encountered a circus joy and what we've
24:01proposes that joint safety and pre encounter.
24:04The animal is trying to predict
24:06what's going to happen in the future
24:09through imagination and simulation.
24:11The threat to appear in the vault,
24:13where does it expect the threat
24:15to appear in the bottom OK?
24:17If it feels that. The sweat is big enough.
24:21What it would do next is going
24:24to prevention strategies.
24:25It can do this in two ways.
24:27It can either change the environment
24:29to protect itself from use
24:31construction or niche constructions.
24:32I say in America and what that means
24:35is that you can build a big wall,
24:38for example,
24:38living the nest to protect yourself from
24:41those threats so they can't get to you.
24:43OK,
24:44the classic example low energy
24:45is herding group living OK.
24:47Living in groups is a way
24:49to protect yourself against.
24:50Threats in the environment.
24:52OK,
24:52so you go through these prediction
24:54systems if you think there's a
24:56potential threat to encounter,
24:58you're going to these prevention systems now.
25:00If we switch over now into post
25:03Encounter stimulus appears in
25:04the environment that point.
25:05We're not sure if it's a threat or not.
25:09We go into these threat oriented strategies.
25:11This is where our attention can
25:13be driven towards specific parts
25:15of the environment.
25:16If we don't have a prediction about
25:18something appearing in that environment,
25:20for example.
25:21With foraging and something
25:22appears in our peripheral vision,
25:24it will capture our attention for you.
25:27Bottom Up attentional systems and
25:28this could either be a threat
25:30or it could be nice threaten.
25:32That would be a prediction error.
25:35Or if we're pretty good at
25:37our prediction strategies,
25:38and we expect that in a certain area of
25:40the environment a threat might appear.
25:43Maybe there's a Bush over there.
25:45We may pay more attention for Tip,
25:47top down attentional systems
25:49towards those past event where
25:50they may encounter the threat.
25:54What will happen next?
25:56It will even ignore that thing in
25:58the environment. It's not a threat.
26:01Maybe it's a conspecific a plastic
26:03bag floating in the night for example,
26:06or they're going to deeper processing
26:08through threat assessment strategies.
26:09This is where the.
26:11Can humans as well we use more of
26:14assessment of the value of that threat.
26:17They were tracked the movement predictions
26:19of that movements of that threat.
26:22They were searched for safety. OK.
26:24Where can I escape to what's the best place
26:28to escape 2 and execute that action OK?
26:31Finally, the sweat begins to attack them.
26:34A danger threshold is breached and
26:35they go into these more circus drive,
26:38defensive or reactive strategies.
26:39We refer to them. Now.
26:41This is where it begins to see
26:43more innate reactions just fight
26:45and flight responses as well.
26:47OK, we also begin to see those
26:48ramping up of analgesic responses
26:50in the midbrain regions,
26:52such as the packets are growing,
26:54which I'll come to in a minute, OK?
26:57So we have these five different
26:59strategies that can be used across these
27:02different levels of threat imminence.
27:05These continuously being updated
27:06by a set of overlapping potentially
27:09independent learning systems we
27:11can think about encounter learning.
27:13When I encounter the threat has
27:16been a more Pavlovian,
27:17maybe instrumental types of learning.
27:20But also I can learn vicariously
27:22for living a group or still my
27:24friend being attacked by a threat
27:26or somebody tells me the story
27:28about somebody being attacked in a
27:30certain part of the environment,
27:32can learn vicariously about those threats,
27:33will come back to that in a minute.
27:37We can also bridge information
27:38that we've never experienced.
27:40We can think about.
27:41Well,
27:42well, lot.
27:43She's example of David Umi says
27:45that through our imagination we
27:47can think about a gold mountain.
27:49We've never seen a gold mountain.
27:51But what we can do is we can imagine gold.
27:55We can imagine the mountain.
27:57We can combine those two parts
27:59of our imagination to create the.
28:02In our minds eye, a goldmountain.
28:06So we can do this is probably
28:08not likely done,
28:09but what we can do a lot of the times
28:12that we can make inferences about
28:15what a potential threat might do,
28:18although we've never experienced that threat,
28:20doing it OK.
28:21So that's where imagination stimulation
28:23systems are very important.
28:25And finally what we have is a
28:27monetary system where we can have this
28:30cognitive appraisal reappraisal or
28:32cognitive control of these systems OK.
28:35Where we can down regulate it to these
28:37challenges with various levels of success,
28:39the monetary system will and
28:41reappraising the process system would
28:43probably play more role in these
28:45pre encountering posting counter
28:47responses as you go down to do more
28:49reactive types of strategies then
28:51you may see that the consciousness
28:53have less of control over them
28:56although they may have some.
28:57What's interesting about this?
28:59This model, though,
29:00is what's different in humans.
29:02I think that there's really
29:04a number of things,
29:05but I think there's two key things
29:07that are very different about humans,
29:10which almost make us the optimal speech
29:12of aiding and avoiding predators.
29:14And that is through this wonderful
29:17imagination system prediction strategies and
29:18this wonderful system of vicarious learning.
29:20We can learn from others if
29:23you can learn from others.
29:25Or you can imagine encountering
29:26threaten in the future and you avoid it.
29:29That's the most optimal defensive
29:31strategy that any Organism can have.
29:33The every Organism tries not to actually
29:36encounter the predators themselves.
29:37We have these wonderful systems
29:39to protect us against that.
29:41But as we know, this can get us in trouble.
29:44OK, we can simulate threats that are
29:47known that actually don't exist.
29:49OK, we can learn about threats
29:51from watching the news.
29:52For example, we can watch.
29:54A shooting in a different part of the
29:57world and feel that we're in threat
29:59in California or on the East Coast,
30:01for example.
30:02Although that place is thousands
30:03of miles away.
30:04So it seems in today's culture
30:06where we probably have more time
30:08to think that we used to.
30:10Well, actually,
30:10before we have less time to
30:12think that we used to,
30:14but we have more time to think about
30:16bad things because of all the bad
30:18things that we read about in the media
30:21and all of the information online and
30:23can really impact these vicarious
30:25learning systems and give us a skewed.
30:27View of the world and we know this
30:29occurs in police officers and so
30:30on are very skewed view of the
30:32world when they go from Chrome
30:34to Chrome to crime for example.
30:35Just simple example there.
30:37So this is really the sort of model
30:39that we've been trying to use.
30:40The great thing about creating
30:42these types of.
30:43Models is that they create a framework
30:45for you to be able to test them empirically.
30:48And I should say 1 final thing is,
30:50is there.
30:51That is one thing I think is we're
30:53missing again,
30:54as I mentioned a little bit earlier,
30:56but one thing it's really been
30:58missing is the way that we approach
31:00the study of fear and anxiety,
31:02and particularly humans is that we
31:03study it as a way of saying look
31:06is this context of.
31:07For example in Pavlovian conditioning.
31:08But we rarely ever do is look at how
31:10we switch between these different
31:12defensive states.
31:13OK,
31:13and we can use time and distance
31:16as a way to be able to do that.
31:18OK.
31:22Oh, I so let's think Libor,
31:23then about threat,
31:24imminence and human defensive circus.
31:25And so I'm taking me a little bit
31:27of time to get to the actual date.
31:29I know you probably been.
31:30We're going to see some of the data,
31:32but I think it's important that I kind
31:34of justify why I do these experiments.
31:37Just a quick note here that if you
31:39go into newest since and you type in
31:42fear and anxiety what you'll find is
31:44almost identical brain regions become
31:46inactive for those two terms and.
31:49The argument is that they are different
31:51systems of the brain and this is this
31:54is some respects that confounded the
31:56way that we approach fear and anxiety,
31:59and it's it's a conference because people
32:02tend to not that after but tend to use
32:05interchangeably use those two terms,
32:07and when they're talking about feelings,
32:09ieti the tops of paradigms,
32:11they tend to use such things
32:13as fearful facial expression,
32:14that the perception of fearful folks
32:16pushing which we know really does
32:18evoke activity in the amygdala.
32:20But what we're shows a lot of
32:22research that we do is we don't
32:24necessarily always see the amygdala
32:25in our response as we look at the
32:28transition between fierce days.
32:31And again in this recent paper,
32:33we propose that there's a
32:35defensive circuit extends from the
32:37prefrontal cortex and hippocampus.
32:39We think of those two regions of
32:41the core of what we would call an
32:44anxiety circuitry along with the bass,
32:47electronica, singular, and.
32:48And this extends into the
32:50midbrain and hypothalamus,
32:52particularly the midbrain of
32:54the regions of the Parker Dr.
32:57Gray, which were associated
32:59with passive and active coping,
33:01freezing and fly behaviors.
33:04Nothing too much details,
33:06but we become more influenced by
33:08some of the work will be conducted in
33:10these population codes that rather
33:12than thinking about one region is
33:14involved in one thing and the other
33:16regions completely switched off,
33:18it's really sort of mixture of
33:20all of these different population
33:21codes across these circuits that
33:23will be associated with different
33:25defensive responses where we have
33:27our toes dipping a little bit in.
33:29Lisa Feldman Barrett theory,
33:30but we believe that their
33:32structure to this OK?
33:36So how does thread transfer
33:37along these defensive circuits?
33:38I want to give you a simple example.
33:41Now of the posting counter
33:42circa strike through US minute,
33:44you may have seen this video.
33:47And I must warn you in this video,
33:50nobody was was killed.
33:55OK, so this is some individuals who.
34:01The filming on top of another
34:03phone which you can see here,
34:04is the tour guide and what they
34:06know is that in that grass.
34:08OK, there is a Tiger,
34:10so we know this is posted accounts
34:12you can't see it, but we know that
34:14there's a Tiger in that grass, OK?
34:18So all of a sudden now we
34:21switched over to Circus Strike.
34:25OK, the sweat is attacking OK.
34:30Alright, there you go.
34:31You can go to attackfootage.com if
34:34you wanna find out more from it.
34:36So what we can see here then is
34:39the join the posting counter.
34:41There's a threat there,
34:42but you know you can't see it.
34:45We don't know at that point
34:47if it's attacking, OK?
34:48We switch over now to circus.
34:51Joy OK,
34:51and the intensity of that threat going
34:54from the distal to proxamol is what
34:56we're interested in in their first study.
34:59What's changing in the brain when
35:01the threat is here versus here, OK?
35:05And a vicarious learning mechanism.
35:06There is a stick is not a
35:09great weapon against a Tiger.
35:11OK, so I'll skip over this one pretty quick.
35:15It was somewhat primitive study
35:17back in the glory days of fMRI.
35:20What we created was it was a task
35:22which is a bit like PAC man.
35:25OK, subjects in the MRI scan.
35:27It was scanning their brain and what
35:29they can do using the keypad is to move left,
35:33right,
35:33up and down in a 2 dimensional maze.
35:36To avoid this red dot OK.
35:39Now what we're interested in is
35:41parametrically what happens in the
35:43brain when the threat is further away
35:45versus when the threat is closer to them.
35:47OK, what we found is that when the red dot,
35:50which I should say if it can't,
35:53if the red dot captured them,
35:55they will receive an electric shock.
35:57OK, we removed the conditions and create
35:59a time to time delay between shots,
36:01and we're not looking at stock effects here,
36:04OK?
36:04What we?
36:05Discoveries when the threat was further away,
36:10this activated regions of the
36:12venture medial prefrontal cortex,
36:13particularly extending to the subgenual
36:15regions of that of the prefrontal cortex.
36:18We also found,
36:20because we did,
36:21our resolution image in here with
36:24the 1.5 millimeter slices we.
36:27Found that natural pastor Mixer also active,
36:29although we bet will look to
36:31talk about that in the paper.
36:34Mainly because people at that
36:36time didn't believe that you
36:37could separate using everybody
36:39why you couldn't separate the.
36:41Regions of the amygdala.
36:42So after threats,
36:43which is to start coming closer
36:45to the subject,
36:46we see switching out to the midbrain.
36:49Parker Dr Gray.
36:50OK, those you don't know much about the park.
36:53Dr Gray is the region that's associated
36:55with fight flight and freeze in behaviors,
36:57along with hypothalamus,
36:58which is believed to also be involved
37:01in instigating those responses
37:02along with the amygdala output.
37:04And we know that if you go
37:06into rodents and
37:07you stimulate specific columns of the
37:09power grid Tegra, you will evoke.
37:12Active coping such things as fight
37:14and flight, and if you go to the
37:16eventual parts of the back doctor
37:19value will see freezing behavior
37:21of those regions are stimulated.
37:23OK, so it was noise to show that
37:25there were these switches between
37:28cortical and mid brain regions
37:30associated with more strategizing of
37:32the prefrontal cortex is I believe,
37:35and more reactive responses in
37:37the Midlands threat came closer.
37:39Now we followed this up with
37:41the paper we published a couple
37:43years later and general cognitive
37:45neuroscience and what we found was
37:47we festival replicated this finding.
37:49But what we also found is that
37:51when subjects made more refer
37:53to Hispanic related motel is the
37:55more miss presses they play.
37:56When they were being attacked
37:58by this virtual predator,
37:59we found that that they rated
38:01themselves is really more panicky
38:03when the threat was closer,
38:05but also that correlated with
38:07increased activity in the midbrain.
38:08Popular Tegra, along with the door.
38:10So raffle nuclear,
38:12which is really been implicated
38:14in panic disorder as well.
38:16OK, so.
38:17That was quite primitive task,
38:21very simple task.
38:21We wanted to see again is that what
38:24happens if we create a task that seems
38:26more realistic to the individuals
38:28where we place a transfer closer or
38:30further away from the subjects foot.
38:32So what we did is we put people
38:34in the MRI scan it and we convince
38:36them that we move in this transfer
38:38the closer or further away from
38:40their foot in this open top box
38:42here OK is a picture of the actual
38:45tarantula is a pink salmon bird
38:47eating tarantula from Brazil.
38:49And the subjects in the MRI scan and
38:51they believe in Avaya live video feed.
38:53They can see me moving this transfer
38:55closer or further away from their foot.
38:58Now the good thing about doing
39:00these types of problems that we
39:01could decorrelate space and time
39:03so I would do St Box 5.
39:05Now go to Box 1, box 3,
39:07box 4, box 2 and so on.
39:09So I could decorrelate space and time,
39:11he couldn't really do that when
39:13you're being pursued by a threat
39:15that's actively attacking you.
39:16It's difficult to separate those
39:17two components in the task.
39:19But we could do that day and
39:21that's what we did.
39:22So the subjects looking down and they
39:24can see this translate in either
39:26closer or further away from their foot.
39:29In these open top aspects.
39:30Apart of these boxes,
39:31we put,
39:32the curtain is so they can't look down.
39:35And a lot of people say to me,
39:38did you get a lot of movement
39:40during this task?
39:41You know people will probably scared when
39:43the transfer was closed there footin.
39:45I say no because we told them if they
39:47move their foot they might overtake
39:49the transfer and it might get upset
39:51and run up the scanner and point them.
39:53So we told him to keep still so we didn't
39:55get too much movement in this
39:57task is a bit more difficult.
39:59Control movement in these other
40:00virtual tasks. Again we call that
40:02stuff out and request that stuff out.
40:04It's natural picture of a subject
40:06in the MRI scan of the box.
40:08OK, what do we find?
40:09We reorganized the conditions and
40:11we want to look here purely and
40:13parametrically at what happens in the
40:14brains that transfers place closer
40:16and closer further to their foot, OK?
40:18So what do we find?
40:20We find that the midbrain regions and
40:22the password or so singular regions
40:24come online as the threat is placed
40:26closer and closer the subjects for we
40:28didn't do I resolution image in there,
40:30so we couldn't specifically say if
40:32it was the power conductor grade.
40:34But when we looked at the peak
40:36of those voxels,
40:37it was right there in the power to great.
40:40But you know,
40:41we like to side with caution
40:42when we make these decisions.
40:44So we generally refer to as the midbrain.
40:46But again, we're talking about this.
40:48Switch to the midbrain regions.
40:51What happens when the when the tranches
40:54move further away from the foot we
40:57found a different set of regions,
40:59particularly in the automated
41:01prefrontal region and we.
41:03Proposed at the time that this may be
41:05related to some form of safety signal
41:08that when the threat is being placed
41:10further away from there for their feelings,
41:13feeling safer and safer and we follow
41:16this up now with three experiments
41:18and theoretical paper that we're
41:20just about to submit to ticks.
41:22And a meta analysis in there that shows
41:25that when we look at this region across,
41:29I think now about 15 studies are purely
41:32safety signals or safety type paradigms,
41:34that this region seems to be the core
41:37region in the perception of safety.
41:40OK, what we also find this very interesting
41:43different from our previous study,
41:45is that.
41:46Or similar to our previous study,
41:48the Packmaster.
41:48She says that what we find is
41:50that when it's a danger signal,
41:52when the sweats distant,
41:53but it's it's more of a danger signal,
41:56we see increased activity in the
41:57posterior pass ultimate pre funded
41:59courses and I'm going to show you
42:01our last experiment that shows this
42:03distinction on the task we refer
42:04to as a margin of safety task.
42:09We want to also look at what happens
42:11in the brain when the individuals and
42:13monitoring over longer periods of time,
42:15the movements of the threat.
42:16OK, so I'm gonna give you a simple
42:19example of what we did is we
42:21look for example just a box 3.
42:23That is more complex than this,
42:25but we looked at Box 3.
42:27What we was interested in is,
42:29is the threat moving from
42:31a previous position,
42:32say of Box 5 year or box 4 OK?
42:35Or is it moving from box 1
42:37or box 2 to 2 box 3?
42:39So if it's moving from keeping the
42:42spatial position identical OK,
42:43we just look into the history of
42:45the movements of the spine and
42:47how they impact their decision.
42:49OK, or their perception of the thresher.
42:51There's no decisions in this.
42:54And what we proposed is that
42:56this long term monitoring of
42:58the threat as it's ramping up,
42:59it's it's movements to become
43:01close to the subject.
43:02This should activate more
43:04of these anxiety circuits,
43:05particularly regions of the bed.
43:07Next right term analysis.
43:08Exactly what we found.
43:09We found that for this comparison here,
43:12we just subtracted as it's moving
43:14closer versus moving away again,
43:16keeping the spatial position the same.
43:18We found that the bed extra
43:20term analysis was increasingly
43:21active for this type of response.
43:23Therefore we proposed.
43:24Associated with increased,
43:25sustained and increased vigilance
43:27of that threat.
43:28OK over longer periods of time.
43:31So what we are shown in this again,
43:35this is building our model.
43:37Dimension of space and Time Model
43:39affair is that space and time to throw
43:42out determine the defensive responses.
43:44OK on the brain regions associated with that.
43:47History of the threats,
43:48movements,
43:49and proximity will activate regions
43:50such as the bad News Journal Terminal
43:53and talk about his predictions
43:54that we call expectancy errors.
43:56OK,
43:57we found that when individuals
43:58rated the spiders being more scarier
44:00than what they originally thought,
44:02that activated the amygdala.
44:04OK,
44:04supporting this idea that the the
44:07amygdala may be associated with the
44:09detection of threats in the environment.
44:12So and also things like surprise as well.
44:15And these are all going on at
44:17a different temple levels,
44:19but they were going on at the same time.
44:23But you could argue here that that the.
44:28The.
44:29These studies don't really
44:32explicitly measure decisions OK.
44:35So we went back to the ecology
44:37literature were very much influenced
44:39by this quite famous theoretical paper
44:41here called the economics of fleeing
44:44from Predators by Edenburg and deal.
44:46And this was a paper.
44:48It wasn't the 1st paper,
44:50it was a paper that really made the
44:52concept of flight initiation distance
44:54famous in the field of behavior ecology.
44:57What is flow initiation distance?
44:58Is the distance at which prey will
45:01flee from an approaching threat.
45:03OK,
45:03so is your decision variable
45:05the economic component to it?
45:07Is related to Audi flee from a
45:09threat when you're performing
45:10other survival behaviors and
45:12this very much captured here.
45:14So as an example of fluctuation,
45:16distance to zebra is keeping its distance
45:18from the predator and the distance to safety.
45:21OK, now if the predator begins
45:23to move closer to our prayer,
45:25there will be a certain point where the
45:28danger threshold is breached and the prey
45:31will flee towards its safety refuge.
45:33OK, there's many different models of this.
45:35I'm giving you the simple one here.
45:38When it makes these decisions, there's.
45:40A cost of not fleeing if it doesn't
45:43flee is eaten by the predator,
45:46but there's also this is the economic
45:48component to the task or cost of fleeing.
45:51OK, so what's as every doing here?
45:53Well, maybe it's doing some
45:54other server will be a dream.
45:56Maybe it's trying to mate.
45:58Maybe it's feeding OK, so.
46:01It's not going to run every
46:03time it sees a lion.
46:05OK, in the environment it just to get
46:07to certain point where it feels that OK.
46:10I need to give up this once if I
46:13will be able to be eating to protect
46:15myself against the predators.
46:17And depending on the internal
46:19states of the zebra, is it hungry?
46:22Thirsty for example,
46:23how far away is it from the refuge?
46:25How fast is the predator?
46:27There will be an optimal point which we
46:30called East area that will determine.
46:32The the time at which the prey
46:34will flee from the predator.
46:37So the backdrop to this is what we
46:39propose is that this will allow
46:41us to still each two stops of fear
46:43that I mentioned earlier,
46:45the first of what we refer to as reactive.
46:48Here we refer to this or define.
46:50This is a quick phrase.
46:51It coordinated reaction in response
46:53to an imminent threat that is or
46:55proceed to be directed towards
46:57Organism and where there is little
46:59time to cognitively comprehend
47:00the danger of the situation,
47:02this reactive fear is about trying
47:03to optimize your defensive response.
47:05Do I freeze or do I flee?
47:07For example, OK,
47:08you've got that time to do that.
47:10This is different from panic and panic
47:12is the next level up in some respects,
47:15but you don't have time to
47:16make those decisions.
47:17Now those decisions are not going
47:19to be more conscious decisions.
47:20They're going to be more over
47:23reflexive Nate type of decision.
47:25Cognitive fell on the other hand is
47:27where we begin to particular focused on,
47:29you must say is there's a conscious
47:31feeling of terror which results from
47:33the presence of threat that is or
47:35perceived to be directed towards organisms,
47:38and where there is not any time to
47:40strategy or just keep it also comprehend
47:42forbidding nature of the situation.
47:44OK, now now the threat is attacking,
47:47but it's distant so you can begin
47:49to say to yourself, OK?
47:51Feeling very good,
47:51this makes me feel terrible or
47:53getting butterflies in my stomach,
47:55but also account that's the better
47:57direction to flee than that direction.
47:59OK, you can strategize.
48:03So we create a very simple task
48:05where and this
48:06is the song to my former grad
48:08student is now at NIH as a postdoc,
48:11and we create is very simple
48:13sort of platform here.
48:14OK, what we have is two types of predators
48:17and early attack and later attack,
48:19which allows us to look at fast and
48:21slow escaped decisions so the subject
48:23controls this triangle just there OK.
48:25And the longer this subject allows
48:27that triangle to be in a position,
48:30the more money they run.
48:31This is the economic component.
48:33Of the escapes,
48:34the goal is subject is to flee from these
48:36virtual predators without being caught,
48:38but also trying to maximize the amount
48:40of money that they earned more there
48:42and what they can do is just press a
48:45burn to escape to this exit at anytime,
48:47and it always flees at the same speed.
48:50We keep this first version very simple.
48:53There's two different types of predators.
48:55OK, as I mentioned that one
48:56that will attack early is here,
48:58and the one that were attacked
48:59late will be here,
49:01the red one and then always start here.
49:03They will oscillate towards the subject.
49:04They'll go back and forth,
49:06will always do this sort of oscillation,
49:07you know, I used to be a boxer years ago and.
49:11The thing you do in boxing
49:13is you faint a lot,
49:14so you pretend you're going to punch.
49:16So we thought we just keep the subject on
49:19edge by trying to do a lot of this fainting.
49:22OK,
49:22so the subject so that is moving
49:24towards a service they always start
49:25here and they move towards the subject.
49:28Now this blue one will attack at some point
49:30within this Gaussian and the red ones.
49:32Some point in this Gaussian,
49:34OK.
49:34We didn't want it to always be the same
49:36position because the subject will quickly
49:38always going to attack and they'll
49:41be optimum very optimal in their escape.
49:43You wanted some element of uncertainty,
49:45but we kept them as a normal
49:47distribution and or Gaussian so
49:48they could get better at this task,
49:50OK?
49:51We will see what we do is reverse.
49:53Thereafter trials they would be.
49:54The exit would be this side and they
49:56would start out after Charles it would
49:58switch and that made it a bit more
50:00difficult for them to to learn it.
50:03If they are caught by the virtual
50:05predator air, they will receive the
50:07lectric shock and they will lose any
50:09money they are in that round, OK?
50:13So we had all the various control
50:15conditions to control for any time
50:17and confounds in the experiment,
50:18which are the first 3 bars
50:20that you can see here.
50:22OK across all of the different
50:24conditions and what this shows here.
50:25These blobs in the brain Air Show is.
50:29When they made their decision,
50:30the two seconds before they
50:33made their decision.
50:35To escape from the approaching threat,
50:37and this is when it was
50:39the fast attacking threat.
50:41OK, and what you're seeing is that when
50:43they're making it fast decision to
50:46the quick or early attacking threat,
50:48you see this increased activity in the
50:51midbrain encompassed in Parker, Dr.
50:53Gray and the mid singular cortex
50:55which we know already connected to
50:57each of these two regions as well.
51:00Again,
51:00so mentioned pages associated with
51:02these flight escape responses may
51:04be more complex motor processes
51:05going on there in the singular,
51:07and we do find that when we apply a
51:10Bayesian decision model here to optimal,
51:12they're joined their task.
51:14Also was probably the right word to use,
51:18but what we find is that there's.
51:21Correlates with activity in the in the
51:22mid singular but not the paradoxical Gray,
51:24so this may be important in driving
51:26their their decision making processes,
51:28and we can see across all the
51:30different conditions will start
51:31in mid attacking threat which I'm
51:32not going to talk about this one.
51:34Keep the story bit simpler,
51:36but what we can see is the blue
51:38increased activity in these two regions,
51:40but in the red one it's a slow
51:42attacking threat.
51:42We don't see these regions come online.
51:44I think about this.
51:45The context of previous study that
51:47I've shown you where we just look at
51:49special temporal distance, it seems to.
51:51To to mirror that.
51:52But yeah,
51:53of course we're looking more at sort
51:55of temporal distance and spatial.
51:57When we're looking at the slow
52:01attacking threat,
52:02we find a different profile.
52:05You find that the.
52:06Parts of the prefrontal cortex and
52:09epic campus come online and positive
52:11leasing that when they're making
52:13a slower escaped decision again,
52:15we're controlling here against
52:17all of the control conditions.
52:19For any timing confounds OK and we can
52:21see here that particular campus and I guess,
52:25positive singular,
52:25not so much the media prefix.
52:27So all these are significant
52:29against each other against their
52:31control conditions and against the
52:33relative fast attacking conditions.
52:34We find that these regions come along.
52:37An argument here was that this
52:39may be associated with more.
52:41Of strategic escape from that
52:43threat or using their memory system.
52:45So imagine where that threat may
52:48have attacked him before more
52:50information processing is going on.
52:53So our prediction then was that.
52:56To become anxious if anxiety is a
52:58future stay, it's a slow state.
53:00We made this prediction that when
53:02subject, so escaping from
53:03a fast attacking threat,
53:04anxiety or individual difference
53:06in translite should have no effect
53:08upon their behavior or no circuits.
53:09We should see the most prominent effect
53:12when they've got time to think and their
53:14damn pioneers talked a lot about this.
53:16Where to become anxious?
53:17You need to have time to think.
53:20You don't need to have time to
53:22think that you're anxious.
53:23OK, so our prediction is that anxiety
53:25would only have an effect upon
53:27the slower processor system which.
53:29Targets these or is a vote by
53:31these hippocampal and prefrontal
53:33regions of the brain which we know.
53:35Josh Gordon.
53:36Others show another work on that.
53:39And that's what we found for
53:42fast attacking conditions.
53:43Again,
53:43we control for variance and everything.
53:45Here we add in their mid
53:47conditions and so on.
53:49We still find that for fast attacking
53:52threats trainings on T as no effect.
53:54OK, however we see that I try and
53:56just individuals when they're
53:58encountering a slow attacking threat.
54:00OK, it predicts or not predict.
54:02So she's word predicts,
54:04but correlates with then fleeing
54:05earlier from the threat.
54:07OK, I am just individuals flee earlier.
54:09The slow attacking threat OK?
54:13What do we see in the brain?
54:15Well,
54:15if we look at correlations
54:17with track anxiety with the
54:19fast attacking threat before,
54:20it doesn't correlate with anything.
54:22If we look at slow attacking threat,
54:24we find significant effect of increased
54:26activity in the upper campus eventually.
54:28Proven coaches and the
54:30install so limited singular.
54:31That those those regions correlate with
54:33increasing car like with trade anxiety.
54:36If we run a PPI to look at
54:39functional connectivity or coupling
54:40between those brain regions OK,
54:43then we find the strength of the
54:45connectivity between hippocampus
54:46and vengeance medial prefrontal
54:48quarters increased as as they
54:50scored higher on trait anxiety.
54:52OK, and again,
54:53this fit really beautifully with the work
54:56of Josh Gordon and others in this group,
54:59showing that this circuit
55:01may be associated with try.
55:03Anxiety.
55:05OK, so I'll go through this one pretty quick.
55:09We have what 5 minutes feel is that correct?
55:12Yeah alright yeah thank you.
55:14Yeah OK,
55:15so here we're looking at what's
55:17called spatial marginal safety.
55:19Another paradigm or theory,
55:20that sort of we stole from the
55:23field of behavioral psychology
55:25and what we want to look at is how
55:28people make decisions about how
55:30close to be to safety in the face
55:33of uncertain attack distances.
55:34And imagine the safeties just simply
55:37defined here as the distance which
55:39prey will move from safety refuge, OK?
55:42And we can see that if the prey is in
55:45a volatile or identity predatory environment,
55:48it would always for which
55:50closer to its safety refuge.
55:52OK,
55:52if it's in a situation where it's
55:54not seen approach to for a long time,
55:57it knows the environment well,
55:59it can predict the environment very well.
56:01It will move further away from its safety.
56:04Refuge, OK?
56:05So we create a a fairly simple task
56:09here again, where subjects are given
56:12contingency, Avaya libeled oil shock,
56:15they encounter three different
56:17virtual predators that asked are
56:19confident they are escaped from the
56:22predator and then what happens is they
56:25after make a decision here about.
56:27How close they want to place their
56:31triangle to the safety refuge, OK?
56:34What happens next is they?
56:38Make that decision and the execute that
56:40exists decision so they made the decision.
56:42They can't move.
56:43The triangle should say in that condition
56:45and then what will happen is they can
56:47move the triangle to move it closer.
56:50Now the reason is this is because we
56:52use motivate panels to say we want to
56:55control for any motor confounds OK.
56:57So we have a pure decision where
56:59they look at the screen and Kate
57:01what decision you're going to make
57:02and then they execute that decision.
57:04That's what we're interested in.
57:06Is that when they're going to
57:07make that decision,
57:08they then see the outcome and
57:10they see if they escaped or not,
57:12but they see the position at which the
57:14predator would have attacked them OK.
57:16So this is actually a speed up version.
57:19Some respects of the flight
57:21initiation distance task.
57:21Now we didn't want to have the same flight
57:24initiation distance task here because
57:26we would have rather enough conditions.
57:28It would have extended the
57:29length of the of the experiment.
57:31So this task was already 2
57:33hours long per subject.
57:34So we do once before hours long post subject.
57:37So what we did is we just
57:39showed in the outcome.
57:40But what we want them to do is
57:42begin to learn overtime where they
57:44think this red one for example.
57:47Will attack them begin to build
57:49a model of wherever the tax.
57:51So therefore,
57:51when they make their next imagine
57:54safety decision,
57:54they'll take in consideration
57:56where attack them before,
57:57but we make this more difficult.
57:59We have two Gaussians dear OK,
58:01and we have one which is one of
58:03interest which is electric kurtick
58:05positive ptosis distribution and
58:07electric kurtick distribution is 1
58:10where you have an increase in outlaws
58:11it's more difficult to predict.
58:13OK let's occur tick lepto means skinny
58:16account so you can see it skinnier.
58:18But what you see here?
58:20Is that there's more outliers?
58:22OK,
58:22so that makes it more difficult to predict.
58:25Um?
58:25What we are here, then,
58:28is a matched variance.
58:29OK,
58:29but normal distribution and then just a
58:31normal distribution with our ferrets.
58:33So this was the easiest to.
58:35Blue is easiest to predict.
58:36The green second easiest and the
58:38red the most difficult to predict.
58:41So Long story short here,
58:43what we see is that subjects place
58:45themselves closer to safety when
58:47they encounter the more uncertain
58:50leptokurtic virtual predator.
58:51So it worked.
58:52We also found we only have 20
58:54subjects in this explain because
58:56it was we won't focus on individual
58:59differences for two hours per subject.
59:02So we have a nice dense data
59:04certain as powerful
59:05data set, but we don't really have a good
59:08large datasets look individual differences,
59:10but we still looked at it.
59:12We didn't put this in the paper,
59:14but what we found is that trait
59:16anxiety predicted out close the
59:18subjects replacing close to safety,
59:20particularly for the leptokurtic thread.
59:21Now we didn't find a
59:23significant correlation here.
59:24For the other types of throughout
59:26the other two Gaussians.
59:27But as you can see, there's a trend,
59:29and I think if we have run 200 people
59:31on this report would have found all
59:33of them with the all of them would
59:35have correlated with try anxiety.
59:37But again, when you've gone end of 20,
59:39we only see the effect here.
59:41So we might extend on that as another paper.
59:45OK, So what do we find in the brain?
59:47I'll go over this cooks.
59:48I know we don't have much time so.
59:50We again as we do,
59:52we use machine learning approaches
59:54from multivariate panelists and what
59:56we're interested in is just the choice.
59:58We're not interested where they make
01:00:00a safe choice or a dangerous choice
01:00:03in this first pass of the data,
01:00:05we just interested what's going on their
01:00:08brain when they make these decisions.
01:00:10OK, and what we find is that some more
01:00:13cognitive fear anxiety circuit coming online,
01:00:15hippocampus and venture medial
01:00:17prefrontal cortex regions.
01:00:18Now we were very interested in this,
01:00:20defined in here.
01:00:22OK, because.
01:00:24What we had predicted that when they're
01:00:26making safety decisions is going to be
01:00:29more this anterior region when they're
01:00:31making more dangerous conditions.
01:00:33More posterior mentioned in the
01:00:35original studies and meta analysis
01:00:37we found we found this region.
01:00:39There seems to be associated with
01:00:41safety signals and potential
01:00:43safety decisions as well.
01:00:44OK or more predictable environments.
01:00:46So what we did is we looked at these
01:00:49two regions separately with hippocampus,
01:00:52posterior immediate venture,
01:00:53me from cortex and anterior.
01:00:55Make venture mini proofing cortex.
01:00:58I will be following up a campus
01:01:01seems to be active for all of them.
01:01:03There seems to be a trend towards the.
01:01:07Epic Campus being more active
01:01:10for the Leptokurtic uncertain,
01:01:12but still there was not significant.
01:01:16Are all significantly above our threshold OK?
01:01:22We found that when it was the
01:01:24more uncertain threat,
01:01:25the only significant above
01:01:26threshold was the posterior,
01:01:27and when it was the easiest
01:01:29one to predict the greenest.
01:01:31So this should be the other way around.
01:01:33This cream is the easiest ones,
01:01:35that one 'cause it's my fault.
01:01:37We found that the anterior is more active,
01:01:40so we want to prove that even further.
01:01:43So what we did then is we run it again.
01:01:47But in universe analysis and we
01:01:49wanted to look at the use in those
01:01:52two regions of the venture media,
01:01:54prefrontal cortex of seeds.
01:01:56Where is their connectivity to and
01:01:58what we found that the posterior
01:02:00passed eventually.
01:02:01Prefrontal cortex seem to be oh
01:02:03central increased coupling with
01:02:05the amygdala and their campus,
01:02:07but For the more anterior parts
01:02:09of the venture River
01:02:11cortex, it seems to be the chordate
01:02:13seems to be more active if we run
01:02:16a parametric modulator on when
01:02:17they made a dangerous decision,
01:02:20the more dangerous the further
01:02:21away they went from safety.
01:02:23Posterior parts of the
01:02:24venture media from cortex.
01:02:26If they made more of a safety decision
01:02:28towards safety venture anterior parts
01:02:30of the venturing readable from the
01:02:32cortex that preliminary model where
01:02:34was was that the more predictable
01:02:36threats will activate more of these.
01:02:39Safety signals that we see
01:02:41in the anterior regions,
01:02:42and this seems to have contact too.
01:02:45Coupling with the stratum and more
01:02:47posterior parts seem to be associated
01:02:49with more unpredictable threats that
01:02:52increased connectivity hippocampus,
01:02:53and the amygdala.
01:02:54And we ran a simple model on this,
01:02:57showing that these two regions also seem to
01:03:01be such J with prediction errors as well.
01:03:05So summary is that fair alongside
01:03:07her dynamic process involving the
01:03:09complexity of defensive circuits,
01:03:11this still or slow gradual attacking
01:03:13threats with activate more.
01:03:14This cognitive fear circuitry,
01:03:16which involves adventure mode equivalent,
01:03:17cortex,
01:03:18hippocampus,
01:03:18posterior singular and parts of the middle,
01:03:21again recalling fear because we define it
01:03:23by the context that you're under attack.
01:03:26Proxamol or fast attacking threats
01:03:28seem to activate more of this
01:03:30reactive fair circuit current,
01:03:32including the midbrain pack,
01:03:34doctor Gray and the motor circuits
01:03:36and more avoidance decisions in these
01:03:38new circuits seem to be searching
01:03:40more perspection and anxiety.
01:03:42When argument is that we found again is
01:03:45that maybe these I order representations
01:03:47of fear maybe in front of block.
01:03:50That's what Joe reduced stats
01:03:52for us to test in the future.
01:03:55Again, we thank all these wonderful
01:03:58people and my lab at Caltech.
01:04:00Thank you.
01:04:04Thanks so much Dean. That was fabulous.
01:04:08Does anyone have any questions?
01:04:15I'll start if there aren't any.
01:04:18So first of all,
01:04:20I love the laptop kurtick finding it
01:04:24makes me think that 2020 might be
01:04:28a leptokurtic easier for everybody.
01:04:30But you know the really nice
01:04:33model that you set up.
01:04:35I wonder if there are other ways
01:04:37to sort of leverage the differences
01:04:38between the different modules and
01:04:40one of the things I was thinking
01:04:42of that sort of came to mind as
01:04:44you were laying out the model.
01:04:46Are the differences between sort of
01:04:49experienced and instructed extinction
01:04:51that we see in the cognitive world,
01:04:53so the idea would be that you've learned,
01:04:57say, that in this leptokurtic
01:04:59environment that you ought to go quick.
01:05:02But but you,
01:05:03having learned those those
01:05:05statistical distributions,
01:05:05you tell them that now you're in
01:05:08a safe environment,
01:05:09but it's not the same context anymore,
01:05:12and the extent to which people believe you,
01:05:14they can actually very rapidly switch their
01:05:17assumptions about the underlying statistics.
01:05:19That works really nicely with JSR responses,
01:05:21for example,
01:05:22and I wondered how you might
01:05:24think about that in the context
01:05:26of your model.
01:05:27That I think that's beautiful.
01:05:29I think you know what we clearly
01:05:32see these differences overtime.
01:05:33In terms of the changing,
01:05:35once they learn that we speak
01:05:38into see down regulation,
01:05:39I think showed the work on the
01:05:42tarantula that overtime, the pH,
01:05:44PG and middle of the old circuit
01:05:46just begins to dampen down overtime,
01:05:49you know, and of course they're
01:05:51learning that is a dynamic system.
01:05:54And I think that doing that
01:05:56type of paradigm would be,
01:05:57you know, really fast,
01:05:59not thought about it,
01:06:00in that in that in that perspective you
01:06:02know one of the questions are typically
01:06:05related to that is basically there.
01:06:07What if you've got this small verse of Nate
01:06:10response that from fleeing from a thread,
01:06:12but then you switch people around
01:06:14to actually doing this paradigm and
01:06:16slightly different way you switch
01:06:18them around to say well look,
01:06:20it's normal for you to feel that
01:06:22when you see through it to avoid it.
01:06:25But actually to avoid the threat.
01:06:27You have to approach it around,
01:06:29you know.
01:06:30I think it's so similar type of.
01:06:33While I'm thinking about your question
01:06:35is that you know our plastic is this
01:06:37system and is that what the hippocampus
01:06:39and medial prefrontal cortex does?
01:06:41It allows you to be plastic.
01:06:42Say, OK, I can approach a threat if
01:06:45it's the most optimal strategy to.
01:06:47Of course, then you have these
01:06:48urges and drives that end,
01:06:50you know, know, know.
01:06:51But we all feel it.
01:06:52You know if you have not
01:06:54done the power she drunk,
01:06:56but I'm sure if I was to my brains
01:06:58telling me don't do it my midbrain,
01:07:00but my prefrontal cortex is saying do it.
01:07:03'cause you save and.
01:07:04You know it's not something that we
01:07:06would normally experience for evolution.
01:07:08I don't think that those says,
01:07:10but we have a way of being able to
01:07:12overcome our threats through changing
01:07:14the environment or approaches to it.
01:07:16Yeah, absolutely, that's a great idea.
01:07:18Maybe should talk more about that.
01:07:20Yeah, I'd love to see you in
01:07:23your tarantula experiment.
01:07:24It was a real tarantula, yeah, So
01:07:26what we did is we showed
01:07:28them trying to in the box,
01:07:30but then we actually switched out to videos.
01:07:32We lost a few people about five
01:07:35people out of 25. In that study.
01:07:37Didn't believe it but we got it
01:07:39and that was really the beginning
01:07:42experience event with you.
01:07:43Kind of got down the protocol but
01:07:45we got it down to convince them
01:07:47and we got everything so that we
01:07:49got 20 good people who believe
01:07:51that I was moving the tarantula.
01:07:52Now the reason why we didn't
01:07:54put the transfer in there was
01:07:56number of different reasons but.
01:07:57That were simple such things as you know,
01:08:00it was difficult to get Abby to
01:08:03actually put ranch in there with them.
01:08:05And also we didn't quite know how he's
01:08:08going to react in the magnetic field.
01:08:11Control the movements and
01:08:13direction every across subjects.
01:08:14So if you want to look at any variation
01:08:17across subjects you know it could just
01:08:19be in that it was the spider was moving
01:08:22more in one day and less on another day.
01:08:25So we want to have some control over it,
01:08:28and that's the way that we did it.
01:08:31We have actually had my two transfers
01:08:33here there just just there actually
01:08:35can see what where we wanted.
01:08:37Some more studies on transfers,
01:08:39but the problem is is just
01:08:41trying to get control over.
01:08:43In attic.
01:08:44And we're not quite sure would
01:08:47function in a magnetic field.
01:08:49It might just freaked him out.
01:08:51So we we try to keep it as real and realistic
01:08:55and the Irbe originally rejected it in.
01:08:59Cambridge,
01:09:00but then they think they once
01:09:02they the wire be wasn't.
01:09:03They didn't reject it
01:09:05because of this transfer.
01:09:06The last thing big thing was
01:09:08that I was lying to the subjects.
01:09:11I I was not allowed to say you're
01:09:13going to see a real translor.
01:09:14I just said on the screen you're going
01:09:16to see a tarantula moving closer to
01:09:18your foot that's not lying to them.
01:09:20It is lying, but it's not really.
01:09:22Because they were seen that ranch,
01:09:23I just didn't tell him if it was real or not,
01:09:25but this is what I mean in the beginning.
01:09:27It was sort of like this.
01:09:29You have to get the pro club down and.
01:09:31And I think we had one person
01:09:33in psychology student,
01:09:34and you never want you never
01:09:35want to scan psychologist years
01:09:37because they question everything.
01:09:38Yeah, we're terrible subjects.
01:09:39Yeah, subjects.
01:09:40I don't know what the hypothesis is that
01:09:42this study you know, like, no don't.
01:09:44We don't want you to think.
01:09:46Just do it.
01:09:48Does
01:09:48anyone else have any anymore questions I
01:09:50could I could keep talking to you all day?
01:09:53Does anyone else have any anything
01:09:55that they'd like to ask him?
01:09:57Maybe you could raise your hands or
01:10:01just unmute yourselves and speak up.
01:10:03OK.
01:10:09I have a quick question so
01:10:12I was very interesting.
01:10:15You're talking was fantastic
01:10:17and so you talked about the the
01:10:20connectivity between the Pfc stratum.
01:10:23And like active escape conditions. So
01:10:26I was curious whether some
01:10:28of the same circuitry that's
01:10:30involved in working memory is also
01:10:33involved in some of this fear.
01:10:35Learning behavior, particularly
01:10:37something like extinction.
01:10:39Yeah, so for the so two questions
01:10:41because the first part is yes.
01:10:43You see the join,
01:10:45particularly on avoidance,
01:10:46which is what we're kind of looking at.
01:10:49A bit of a Joule. Imagine a safety.
01:10:52Question experiment is that we
01:10:54kind of look into the neuber
01:10:57between escape and avoidance.
01:10:59You're making an avoidance response.
01:11:03Which is really related to
01:11:04an escape response later,
01:11:06so it's annoyed between those two
01:11:08and what you see is the animal
01:11:10that shows that the stratum is
01:11:12involved in avoidance behavior.
01:11:14We bout to the second power working memory.
01:11:17Yes, I think it is just this is Jose
01:11:19and he thinks that the higher level
01:11:21cognitive process is really important.
01:11:23Part is the working memory system.
01:11:26So absolutely I think it is important
01:11:29because it's there where we're,
01:11:31you know,
01:11:32processing information online and I
01:11:35would have to do the experiments to
01:11:38be able to look at these effects.
01:11:40We talked a little bit about him,
01:11:43similar experiments to look at the
01:11:46relationship between working memory
01:11:48and some of the decision somewhere
01:11:51completely different paradigm
01:11:53with Thomas Love is in my lap.
01:11:55We were doing a lot these
01:11:57experiments online now,
01:11:58so we're just going to behavior,
01:11:59but it would be sort of interesting
01:12:01to see if we can predict the
01:12:03working memory ability and capacity
01:12:05influences these prices.
01:12:06And we did do one study with.
01:12:09See,
01:12:09that's what 7 Tim Dalgleish where we
01:12:11looked at working memory training in
01:12:13the ability to regulate in appraise
01:12:15your threats and the environment.
01:12:17And we found that that that not
01:12:20quite sure they all the brain
01:12:22training stuff is held up,
01:12:24but we did show in effect there of that.
01:12:27So yeah I think what can we just play
01:12:29an important role but again we've
01:12:32not tested that directly ourselves.
01:12:39Anymore questions.
01:12:44Well, thank you again Dean.
01:12:45Thanks for getting up early in
01:12:47joining us and thanks for
01:12:49sharing your beautiful
01:12:50work. Yeah thanks Phil.
01:12:51And yeah I'm going to do some more
01:12:53coffee and going to keep you want
01:12:55me ioffer the politics and given
01:12:57the talk, I wasn't sure
01:12:59whether there was an approach.