Understanding Cohesion and Collapse in Macaques Societies: Social Housing


I am Eric Hutchinson. I’m with the Division of
Veterinary Resources. I’m one of the organizers of this
conference, so I get the lucky job of introducing the speakers this morning,
a lot of the speakers this morning. But one of my favorite things about this
meeting is the diverse experience and expertise that people
bring to this effort. So in order not to fill all of your
time with me babbling about the great qualifications of everybody
that’s here today, we do have in your packets a hand-out that
lists the bios of all of our speakers. So I will try and keep them brief and
with dazzling highlights. First, we have our keynote
speaker Brenda McCown. She has also been, luckily for all of us a
great help in organizing this conference. She has been our go to
person here at UC Davis and our point of contact in organizing this
conference, helped to arrange the venues, helped to arrange that awesome tour that
a lot of you got to go on yesterday. So we’re really thankful to her and
also excited to hear her talk. She has more then 20 years of primate
behavior research experience. She’s head of
the UC Davis Behavior Program. She’s a great researcher
with lots of funding. So if you’re looking for a job,
I’m sure she’s hiring at all times. She’s going to talk to us today about
her research on social networks, and obviously, that’s a pretty relevant
topic for socialization of animals. And without further ado, Brenda.>>Thank you, Eric, and thank you for inviting me here today to give you a talk
on some of the work that we’ve been doing. Looking at social network,
and social stability, in large breeding groups
of Rhesus Macaques. And I wanna go back sort of historically,
very quickly, just to say that when I was asked to join the behavior management
program back in 2004 as a faculty of the School of Veterinary Medicine, I sat
with the associate directors at that time, one of whom is in the audience
here John Capitanio, I said, what do you want me to do. They said well, a number of things, but
one really serious problem that we’ve had over the years,
this pesky little things called cage wars. And these are these cage wars where animals
in this large breeding groups completely collapse socially and
have to be dispanded, basically removing more than
100 animals from the cage. So from that moment on, I began to really think about how
to tackle this kind of problem. And since really,
very seriously, since 2008, with funding from the National Institutes
of Health, we’ve been able to really begin to understand cohesion and
collapse in the cat societies. And we use this for both social management
purposes but also as a translational model for human social systems and
their relationship to health. So I’ll be talking today about a series
of studies that my collaborators and I have been conducting since that time
on network stability of rhesus macaques social systems to improve
animal health and well-being. And as I said, as a translational model,
for humans as well. So why are rhesus macaques
used as a biomedical model? Well it turns our rhesus macaques are the
most common biomedical non human primate model that we have. And it’s partly because They’re
not only genetically and physiologically very similar to us,
but they’re very hearty and adaptable, and
do very well in captive situations. And they’re very much like us
in that they’re a weed species. You look in South and Southeast Asia,
they have a really large distribution and they’re extremely well
adapted to human contexts. In fact, so adaptable that rhesus macaques
are considered to be a nuisance in Asian and Southeast Asia. Great example of a rhesus macaque
stealing an ice cream cone, probably from a poor child. They’re also very good at bartering for
sunglasses as well. So they really are quite
a nuisance in these contexts. And part of the reason why these
guys are a nuisance is probably has to do with their dominance style. Rhesus macaque live at a continuum of
different kinds of dominant style going from despotic to egalitarian,
kinds of communities and rhesus macaque are at
the top of the despotic. So while we keep them in captivity,
they’re very adaptable, but they’re also very despotic. Which means that they have high
degrees of aggression, and lots of severity of aggression, as well. So while keeping them in these breeding
groups is great for a breeding colony doing biomedical research, they’re
very adaptable to these situations and it promotes their species typical behavior,
and it promotes their individual health. We have some serious problems with trauma. Trauma needing veterinary attention, trauma leading to social
relocations of these animals. And as I said earlier, social systems
that fall apart and collapse, what we call social overthrows or commonly
known, by our group, as cage wars. So we began to look at this problem, we
started to look at this notion of what’s going on in these kinds of social groups. And what we found, of course,
is that there’s variation in aggression within social groups,
both in contact aggression and trauma, and this can relate a lot
to the breeding season. A lot more trauma aggression during
the breeding season than outside the breeding season. But what’s more interesting is we
also found variation across groups. So, not all groups show the same amount
of aggression despite the fact they have similar composition and
density in these groups. So, it’s this variation that allowed us to
begin to see how we might be able to pull apart some of the problems that we have
with aggression with these groups and the social overthrows that
occur on a periodic basis. So we needed a conceptual framework from
which we could really begin to understand this processes. And we decided to look at a myriad of
factors that really contribute to social stability and aggression patterns
in rhesus macaques groups. And this includes things like
bio-behavioral organization. For those of you who were
on the tour yesterday, we talked about the BBA program so we can
try to use personality and temperament to understand in the groups how they
contribute to these kinds of problems. We can also look at environmental social
stressors, such as kinship patterns. Density of course would be interesting. Composition of these groups
including things like sex ratio, which we’ll talk a lot about today. And how that interacts with the social
network structure both the individual level, roles of individuals in the network
As well as at the group level, the network structure at the group level. And how that contributes to
different kinds of health outcomes, including wounding and trauma. And I will be focusing today
specifically on wounding and trauma, and not talk about some of the other health
outcomes that we’ve been looking at as part of our translational work. Okay, so
we have a conceptual framework now, right? We wanna look at the myriad of factors
that are contributing to social network structure to individual
roles in these networks, and how that affects wounding and
trauma in these groups. But we can also take two
different approaches. So we can even look at rates of behavior,
rates of aggression. Remember, we saw variation in rates
of aggression across these groups. Or we can take an internal look at
the rates of these rates of aggression and look at the patterns in the relationships. So it’s a little bit different than
just looking at raw rates of aggression. It’s about looking at
patterns of relationships and we need social network
structure in order to do that. So let’s do a little demonstration
of social network structure. I’d like everybody who is coming from
the primate center to raise their hand. Okay, has anybody noticed that kind
of clustered in the middle there?>>[LAUGH]
>>And then I’m gonna pick on
one other group as well. How about IACUC? Can I have the hands up from the IACUC? So what we’re seeing here is that people
are clustering themselves with respect to their relationships, and so it’s these patterns in relationships
that we really want to understand. It gives us more information about
what’s going on in these groups. So the notion of this is
not just clustering, but also this idea that there are direct and
indirect relationships that are important. Does anybody in here know about
the six degrees of Kevin Bacon?>>Would somebody, please,
like to describe what that is? Loudly?>>You have a network context
type people you have contacted. And then, whoever they have contacted,
I always like to reflect, if anybody knows Boris [INAUDIBLE]
who used to be the director of the Russian Primate Center for
60 years. They have indirect contact with Stalin, Khrushchev, Brezhnev and
every Russian leader from there. So that’s the concept of
Kevin Bacon [INAUDIBLE]>>Exactly. So, the idea here is that
you can reach Kevin Bacon through six degrees in
anywhere in the world. And it’s actually true that
there’s been some work done using letters that show that you can reach
anybody in the world in six degrees. It may even be shorter, now,
because we have the Internet. And I think that it’s actually a little
bit shorter then as a result of some of our technological advances.>>[INAUDIBLE] another example. If you see Ed Harris up there. Well, his brother is Bob Harris who’s
lab animal veterinarian who just retired from Rutgers. So, he’s snugged up against
Kevin Bacon up there [INAUDIBLE]>>[LAUGH]. Hey, now, we only have three. One degree, hey. All right. We’re doing good. So, this is this idea,
this is what networks measure, okay? They measure both direct relationships
as well as indirect relationships. And it’s the concept that friends
of friends actually matter, that these indirect paths actually matter. And so, this is just a representation
of that, more formally, showing that we have the direct relationships, but
it’s also these indirect relationships that reveal hidden patterns in these
relationships and how they can influence, in our particular context, aggression,
trauma and social group stability. Okay, so, that’s a nice little network. It’s real easy to see those direct and
indirect relationships. But what about if we pull
up this terrorist network? Suddenly becomes a lot more difficult to
quantify what’s going on with respect to the direct and
indirect relationships, right? So, we need tools. We need a tool set for us to understand
how these networks are structured. And what we’ve developed along with our,
they’re standard network measures, but we’ve also developed some
computational approaches. And I shouldn’t say we, I should say
Fushing Hsieh has developed computational approaches to look at these indirect and
direct relationships in using data-driven, bottom up approaches that have
no underlying assumptions. And I will be using some of these today. I’m not gonna go into detail about them,
but I’ll be using some of these today in order to show you what we have found
with respect to these systems. Okay, so,
we’re talking about rhesus macaques. How many of you were at the tour
yesterday, raise your hands? So, all of you know that we got
to see the large social groups. Lindsey did some positive reinforcement
training and showed us some of the, talked a little bit about social
structure in rhesus macaques. And rhesus macaques society, the origin
is from South and Southeast Asia, their social structure, they live in
large, multi-male, multi-female groups. The females are philopatric, that means
that they stay within the social group and the males leave, they disperse. The females inherit their
rank from their mother, and we see a pretty strong dominance hierarchy
in these animals, although they’re not as linear as we once thought based upon
some of the work that we have done. The CNPRC groups mimic this kind of
structure, except for one caveat to that, and that’s, traditionally, natal males
are not removed from the social groups. And this will become
important in a little while. Okay, so, our study groups are these
one-half acre enclosures that some of you got to see yesterday. They range anywhere from
80 to about 200 animals, although we’re decreasing
the density actively. We’ve studied 16 groups so
far in our population, and some of these groups are stable. And some of these groups
actually had a social collapse either while we observed them or
after we observed them. So, we were able to look at
stability in these social groups. And we did this by looking at behavior
such as conflict, aggression, looking at what we call status signaling,
which I’ll describe more about later, and also affiliation. Okay, so, we’re gonna take this conceptual
framework of looking at the myriad of factors that actually influence social
network structure and individual roles in the network, and look at how it affects
wounding and trauma and social collapse. And, basically, what we found is
three major pathways through which internal external factors
affect network structure and the social outcomes of wounding and
social relocations. The first is matriline fragmentation
which I will not be talking about today. This is simply the idea that
matrilines that have many cousins and not a lot of sisters are not as cohesive. Personality, which I will
only touch on very briefly. And what I’m really gonna focus on for this particular talk today
is the role of males. The really, really important role
of males in these social groups. Because the males are very,
very important in maintaining social stability in these
large social groups of rhesus macaques. And we know this because we can first
take a look at a very general parameter, sex ratio, and we can see that there’s
a negative relationship between the trauma per individual, so it’s this trauma total over the total
number of individuals to standardize it. And we can see that when males are around, we have less trauma in
these groups overall. But not all males are created equal. So, the natal males in
these captive groups can increase in rank if they are coming
from high ranking families. As we can see here, male rank on the y
axis and male’s family rank on the x axis, males that are from high ranking
families are also high ranking. And they gain that rank
through untraditional methods. They’re not doing it because they have,
they have power to become high ranking. They’re doing it through kin alliances
with their sisters and mothers. Their mothers and sisters are supporting
their males in increasing in rank by helping them in their
alliances in interactions. So, high ranking natal males who
get support from their kin actually occupy really central positions, and
that’s what we’re looking at here. You can see, if you, whoops,
here, the animals that are, the natal males that get
frequent kin support and are high-ranking are very
central in their networks. Meaning that they have
a lot of connections and they have a lot of connections to
individuals who have a lot of connections, something we call Bonacich power. Again, friends of friends matter. And, interestingly,
they actually start more fights as they have increases the number of kin
alliances per week, and they also have a higher frequency of intense aggression
if they have more kin alliances per week. So, these natal males, these high
ranking natal males, are problematic. Okay, so, not all males are created equal. So, then why do we have this relationship
between sex ratio and trauma? What’s going on here? Natal males are problematic. It’s not all males that are helping
with this particular type of problem. And to do that, we have to understand
some concepts known as rank and social power, okay? Now, to do this, we have to start
talking about the direct and indirect relationships that
I brought up before, okay? So, you can imagine that we
can have a direct relationship between this individual here and
this individual here. But we also have an indirect relationship
between these two individuals, right? And we can as many, we can have pathways
up to, I think, one of the highest orders for us go up to, or at least what
we can measure, is up to seven orders. But we can have these indirect
relationships where there’s no direct relationship between two individuals. But there are these pathways that link
them, in terms of their dominance or aggression networks. So we’re showing a one-step
pathway here in red, a two-step pathway in blue, and
a three-step pathway in yellow. So you can also imagine, once you start
looking at these indirect relationships, that there could be relationships that,
through different pathways, either show consistency, or
do not show consistency. And this is represented here. So we can see in the graph on the right,
that if you go from F to G, through H, J and I, you have
a consistent relationship from F to G, and you see that if you
go from F to H to G, without going through that other pathway,
it’s showing consistency. So the black are showing
consistent relationships. Whereas, on the left graph, you can have these indirect pathways
also showing inconsistent relationships. Where you have a two-step path between A
and B, in which it goes in one direction, meaning that A is aggressing B, or you can
have B going through a different pathway, showing inconsistency with that path, meaning going through that
indirect path going from B to A. So this is this notion of these
indirect pathways being important, but not just that they’re important, but
the importance of the consistency of those kind of relationships, and
this becomes really important. So, we can talk about dominant certainty,
which is a little bit different from rank. So if we take all those indirect
pathways and direct pathways, weighted by the direct pathways,
of course, but there’s not all individuals have a direct pathway, we can come up with
what we call a dominance probability. So the example I’m showing here is 35502. Ashley what’s the name of that one?>>[INAUDIBLE]
>>Thank you. And 27993,
there’s a really strong relationship that 35502 is dominant to 27993, okay? And on the lower panel here we can
see that in the lower panel here we have dominance probabilities
that are very low, meaning that 38928 is very
subordinate to 2793, okay? So, this is how we read these
dominance probabilities. And we can get rank out of these
probabilities by doing a bunch of shuffling using these really cool,
computational approaches, but we also can get something else and
that’s called, dominant certainty. Dominant certainty is the average
across the all of the individuals you have a dominance certainty with. And it tells you the degree to
which you’re certain about your relationship, okay? In this case, you’re very
certain about your relationship. But here you’re also very certain about
your relationship because you’re really, really subordinate as opposed to really,
really dominant. So if we wanna get a dominant certainty,
what we do is, we minus the 1 from, we minus out from
the lower part of this basic matrix. And we can get a dominant certainty
that ranges between 0.5 and 1. And it tells us how certain those
individuals are about the relationships, not just about their rank. And when we do that, we see a u shaped function of
rank against dominant certainty. So we have rank on the x axis, we have
dominant certainty on the y axis, and you can see that Individuals
that are very high ranking and individuals that are very low ranking
have higher dominance certainty than individuals that are middle ranking. But, we do see individuals that are high
ranking that show low certainty, high ranking that show high certainty,
low ranking that show high certainty, and low ranking that show low certainty. So we’re decoupling rank from certainty. And this becomes important, because if we
look at dominance certainty and trauma, what we find is that the expected number
of traumas per individual decreases with an increase in dominance certainty. So individuals who have high dominance
certainty have lower trauma rates. Not surprising, right? And this is true of not just the rate
of trauma, but also severity of trauma. So individuals with higher
dominance certainty actually show less severe aggression as well. Okay, so now we have an idea of
what dominance certainty is. We have an idea of what rank is. There’s another notion that we need to
understand and that is social power and social power is actually measured
through a very different mechanism. Are called silent bared teeth displays. Now, historically, traditionally,
this behavior, this signaling system has been thought to be a signal
of fear, been called a fear grimace. But what we’ve discovered
over the last several years, is that silent-bared-teeth displays is
actually multiple types of signals, and there’s many different functions for
silent-bared-teeth displays. And one of them, when produced in peaceful
context, where there is no aggression, but literally, animals will walk
up to another animal and give a silent bared teeth display. Frequently, looking at the animal very
closely, as we see in this picture here by Lisa Parr from Emory University,
is what we call peaceful SBT’s and it turns out that it’s a formal
signal of subordination. That Jessica Flack and
Frans de Waal discovered. And, we have and
looked at in pigtail macaques, and we have been looking at in resis macaques. And essentially,
provides a consensus of subordination, a type of voting by group members
that these animals are powerful. So we call this social power. Okay, so that’s also a very
important concept to take with us. Okay, so back to the idea of the roles
of males in these social groups. Well, it turns out that males play
a really, really important role in intervening in the conflict of
others in their social groups. And I mean others that do not provide
any benefit to the male himself. So we see that males provide
more policing behavior. We call intervention or
policing behavior then females do. And we also find that unrelated males are
more successful at these interventions, at policing third party conflicts than
either related or unrelated females. And we think this is part in due that
related males are really not policing their own alpha matraline because they
actually get a lot of support from their alpha matraline. And interestingly,
if we look at social power, from these status signals that I
was just talking about earlier, we see that the predicted intervention
success by males is higher as you have higher social power,
and particularly in males. And everybody’s gonna love this. So the other thing that we find is that if
you take a look at social power against intervention success, and you now plot
some of the bio-behavioral stuff that we do with these animals, in males and not
females, tolerant individuals are better at intervening on third party conflicts
than either bold or unpredictable. And yes, I am going to refrain on
making any comment about our current presidential campaign.>>[LAUGH]
>>I don’t know if I want that recorded though.>>[LAUGH]
>>Okay, so, social power is really important. These SBT signaling systems are really,
really important. It’s so important that we’ve deemed
them what we call at keystone network. It turns out that these SBT networks
are extremely directed and what, transitive, meaning that they
have no cyclic pathways. There is no inconsistency
in these networks. You see a lot, you do see an inconsistency
in the aggression network. But you do not ever seen
inconsistency’s in the SBT networks. It also turns out that
the multi-leveled structure, these indirect pathways that
I’ve been describing to you, is it’s also extremely important
stability in these groups. And it’s associated as we’ve seen
with conflict policing behavior. So here’s some examples of some
really strong, stable SBT networks, keystone networks,
in our stable social groups. You can see that it’s multi-level. There are a number of indirect pathways. There are also direct pathways. A lot of cohesion,
a lot of connections between individuals. And this is what we’re looking for when
we’re looking for stable social groups. In fact, if I briefly show you what
happened in one of our social groups back in 2009, we showed multi-level
structure when it was stable, and then we continued to collect data. And after we finished collecting data,
this group fell apart. We had a social collapse. We see a complete breakdown
of this SBT network prior to social collapse in these groups, okay? Another example I’ll be coming back to
these to discuss a little bit more about how we can look at these kinds of stable,
unstable systems. But again, multi-level structure, lots of connections completely falls
apart, but right before a social collapse. I just wanna point out here because
I have here this is a paper, Fushing et al., 2014. One of our papers in
the International Journal of Forecasting. We actually use this model as a way
to talk about how we might be able to use these kinds of methods to predict
financial collapse in human social systems. So monkeys are good for
a lot of things, okay? So again, back to this idea of direct and
indirect pathways. Very important that we have these indirect
pathways not just the direct pathways. And it turns out that if you have lots of
SBT pathways, these indirect pathways that you also have a lot of certainty and
dominance. So what we’re seeing here in green is
a lot less variation in the dominance certainty for individuals that have
multiple paths in their SBT network. Meaning basically that If
you have more SBT pathways, you tend to have greater certainty, okay? And this is important because the more signaling pathways you have
in these networks, direct and indirect, we find we have a greater conflict
policing in these groups. So individuals that are doing
the conflict policing have a higher number of these power pathways,
okay? So I hope at this point we know how
important SBTs are toward conflict policing, which can lead to stability
in social groups and reduce trauma. So the question we have is do
these high ranking natal males destabilize these power pathways and
the structure of these SBT networks? So recently in the last few years,
we’ve been doing some experimental work, doing what we call natal male knockouts. Kind of like a gene knockout, but instead
you take out a social entity in the group. So what we were doing is we were removing
these natal males, not necessarily specific individuals as opposed to
the role that they were playing. So we removed high-ranking natal males
because of the role we think that they were playing in these groups. And we basically did six weeks of
baseline, six weeks of post intensive, and eight weeks of follow up on
a less intensive schedule. And that’s to test
the idea that whether or not high ranking males actually
destabilize these groups. And the answer to that question is it
depends on the type of natal male, the type of high ranking natal male. In some groups, what we found is that when
we removed the natal male from the group, we did not see an increase
in the complexity of the structure of these SBT networks. And what’s interesting is that in the case
of where the natal male, whoop, sorry. The natal male is beginning
to have his own SBT network. He’s stealing SBT pathways
from the alpha male is where actually see an improvement
in the complexity of the SBT network. Whereas natal males that
are not causing problems, they’re within the SBT
network of the alpha male, does not improve the complexity and
structure of these SBT networks. So actually matters who you are. It’s not just all high ranking males or
nasty, okay? And we could look at this
a little bit differently. In that, we can take a look at
these indirect pathways and the ratio of these indirect pathways with
respect to 2 step paths, 3 step paths, and 4 step paths to see if there’s
improvement in that structure, right? So if you have a greater ratio of
2 step paths to 1 step paths or 3 step paths to 2 step paths,
it means you have more complexity and more indirect structure in that network. And we can do that, as I just said,
by looking at the ratio of those things. And when we do look at the ratio
of these indirect pathways, what we find is that, in the natal
males that are not problematic, we do not see a change,
an increase in the ratio of these paths. But we do in the ones that where
the natal males are problematic. And this one just represents here sort of an ambivalent one where we see
some improvement but not all improvement. And we can then begin to look at
the change in the ratio whether it’s increasing or decreasing. We can begin to look at
natal male attributes and behavior that might contribute to
the trends that we’re seeing here. So we looked at a number of different
kinds of males’ behavior, but I’m gonna focus mostly on group fights,
percent group fights, natal male supporting kin, and the
dominance probability with the alpha male. So remember, the dominance probability is whether
you’re dominant to an individual or not. If it’s higher,
then the alpha male is more dominant to the natal male and
if it’s less, the natal male is more dominant
to the alpha male, sort of, yeah. Okay, so let’s take a look at that. So if we look at the change of
the ratio of 1 to 2 step and look at the percent of group fights
that are started by the natal male, what we find is that when natal males were
not problematic, the complexity did not improve and we see it. So the number of group fights
actually is associated with natal males being problematic and
complexity actually improving in green, complexity not improving in red,
and the equivocal one in yellow. Okay, so
if natal males are starting more fights, then removing them is a good thing. Same with agnostic support
given to kin by natal male. So if the support is given
to kin by the natal male, then we see that it improves if we
remove the animals from the group, okay? The green means that it’s improved. And then finally, a dominance probably
between natal male and alpha male, we find if the dominance probability
is if the natal male is close to the dominance probability of the alpha
male, removing them is a good thing. So it shows the opposite
relationship as the others. If the alpha male is really dominant,
quite dominant, to the natal male, then removing
them doesn’t change anything. Okay, so we’re getting a sense of how
these individual males, natal males and alpha males,
are interacting with each other. So I now wanna turn to looking at
social power and group stability. So we’re looking sort of at
the individual level what’s going on. So let’s turn to looking at group stability
and social power, and here, we’re going back to this notion of SBT networks,
and stable and unstable structure. Just to give you an example of sort
of how we look at these things, we had several groups that,
actually, we watched for a year, and then they would socially
collapse at the end of that time. And there were other groups in which
we would be watching the animals, and they socially collapsed while
we were actually watching them. So we really have an interesting
set of groups that we can look at with respect to this. So and what we tried to do in terms of
looking at SBT networks and structures, we also wanted to look at how two different networks would
actually correspond to each other. So not just that we have an SBT
network that we’re worried about, but what happens to the SBT network in
relationship to the aggression network? And this is a tool that we use
called Joint Network Modeling, which basically allows us to look at
the number of dyads that actually show these different kinds of patterns. So here I’m showing bidirectional
aggression with bidirectional status, doesn’t happen all that often. [LAUGH] And these, we count basically the number of
dyads that fit this fit this pattern. Another example of this would be to
remove the bi-directional status and just look at bi-directional aggression. And here’s an example of that, where we
have two bi-directional aggression but no status signaling, okay? So we can count the number of dyads for
each of these different things. And we can get a graph,
or basically a table that tells us the total number observed
dyads under these different contexts, noting that most of them had zeros,
no interaction at all, okay? And we can then do an expected dyad,
the expected number of dyads. We expect to show this just by chance,
okay? And this is this notion of independence. And we can actually add these up using
a chi-squared technique that allows us then to look at what happens when we
start making other kinds of assumptions. So we start with a Null model,
which is the chi-square value that you get from just looking at
the observed minus the expected, or basically the absolute value of that. And then we can start to add what
we call constraint functions, where we can actually say, well you know,
two-way aggression is really rare and we can see how it changes
the chi-square value. If it’s important,
it will decrease the chi-square value. We can then ask,
what [INAUDIBLE] second function? Two way status signalling is rare. If that’s important,
it will also decrease this chi-square. So both two-way aggression, and two-way status, they do decrease
the chi-square function a bit. But,the third function, aggression and
status, are in opposite directions. That means that I give a status signal
to a animal that’s very dominant to me. Shows a huge affect on
the chi-high square value. So it’s extremely important
to group stability, okay? That you see aggression and
status in the opposite direction. So what happens if we take a look
at stable and unstable groups? Let’s first start just with stable group. Don’t worry so much about
the absolute value of these things, it’s the pattern that’s occurring
that we’re interested in. In the stable groups we see
the same pattern over and over and over again, over many different periods. These are all just different periods of
data collection that you’re seeing, and it shows the same pattern. There’s two-way aggression is
contributing a little bit, two-way status signaling a tiny bit, but
the really big thing, the really big result is that there is opposite direction
of aggression and status signaling. When we look at the unstable groups, what happens is that begins
to completely fall apart. We see that actually two way
aggression actually increases here, it becomes more important, and we start
to see sort of the disintegration of the importance of the opposite
aggression and status signaling. And it’s much more,
actually it’s more illustrative here, where you can see that
starting to fall apart. It no longer has this huge drop. In some cases,
you see it happen immediately. And then in this third case we’re
looking at here, we just saw gradual disintegration of this joint modeling
of aggression and status signaling. So in summary, Peaceful SBTs represent
a keystone network that leads to dominance certainty and
provides opportunity for conflict intervention that leads
to social group stability. These Peaceful SBTs are mostly
given to high ranking males. Although I’ll note that we do have
another kind of SBT that is given during the breeding season, and
that’s by males to females. So it switches around a little bit when
there’s stakes for mating opportunities, but we think its a completely
different kind of signal. In fact, females, males almost never give
females Peaceful pSBT, it’s almost always females giving high ranking males or other
males giving high ranking males SBT’s. High ranking males, we’ve shown can
disrupt these Peaceful SBT networks. They steal SBTs from higher-ranking
unrelated males, and these disrupted networks can lead to the disintegration of
Peaceful SBT networks, and decoupling of aggression and status networks that
is predictive of social collapse. So we can use this to
predict social collapse and prevent trauma associated
with these events. So I hope I’ve shown how important
the role of males are in these groups and particularly the high
ranking unrelated males. In fact, we think they’re so important
that we’ll be starting a new R24 come this summer, where we will be trying
to introduce males into social groups. Yes I know Ashley, I know. Yeah. [LAUGH] I know you’re worried. But it’s really quite important, so we
wanna learn how we can get good, tolerant, high ranking,
unrelated males into these groups. And as I’ve said, we’ve done some other
network stuff that I’m not presenting today, showing other kinds of pathways
that contribute to this kind of stability. I also hope I’ve shown the importance
of looking at the myriad of factors that interact with these
social networks and individual roles in these networks to
produce these kinds of health outcomes, trauma, wounding, social relocations,
as well as actually a number of other kinds of health outcomes
that we are currently looking at. And I’d finally like to say that
the system science network approach, instead of just basic rates of behaviors, really has given us greater insight
though the processes leading to these higher rates of trauma and
social group instability, and because it’s the beginning of
the social housing conference. Social housing, of course, is critical for
better welfare and research outcomes. But as I’m showing this morning, I think
we’ll see throughout the venue here, it can be quite challenging and complex. And with that, I would just like to
acknowledge all my collaborators. The observation team that does
amazing work out there, and, of course, the Primate Center staff and
funding sources. Thank you.>>[APPLAUSE].>>Thank you, Brenda. We have a few minutes for questions, if anybody wants to ask
questions before the next talk. Sorry I sprung this on you. [LAUGH]
>>[LAUGH]>>Hey you’re taking questions. [LAUGH]
>>[LAUGH] I can take questions->>I just was wondering if you might have an idea about How much data, how many hours of data
collection do you actually need? Cuz I know your team is out there, but for just regular management,
do you have an idea about, yet. [LAUGH] About how much data somebody
like the behavior team could actually collect, or obviously focusing
on SBTs would be the thing. But do you have an idea
about how many hours?>>Yeah,
I have a couple of ideas about that. Let me first say that we, a master’s student of mine actually did
a simulation looking at the amount of data that’s needed to produce these
kinds of social network measures. And he found that it’s about 50% is enough
of what we collect in order to do that. Now, we collect a lot of hours. So it may be that it would
require about half of that. So I’m not remembering exactly how many
that is, but I can certainly look that up. Well, they’re out there from basically
eight to four, right, four days a week. So basically half of that. But I think there are ways
to use this in another way. I mean you don’t necessarily have
to collect SBTs on everybody. Why not collect SBTs on natal males and
alpha males? That might be enough to give
us a sense of what’s going on, because it’s what we call ego networks,
right. You can get a big network out
of what we call an ego network, just focusing on one individual. But you’re still getting all
the connections to others as well. So we’re doing this with
the infant study right now. So it might be that you could focus that
on something that is actually really doable from a management perspective. Or you really focus on SBTs and
not on everything else. We’re now looking at grooming networks and
things like that, so we’ll see how important those are as well. But we’ve really been focused
on this particular question. And then maybe you won’t have
to collect as much data. We collect data on everything,
conflict, affiliation. So focusing on the SBTs
might help as well. You might not need to have as many hours.>>How do you think this applies to same
sex groups as opposed to a breeding group?>>That’s a good question. We haven’t actually looked at that. That might be a future
study that we could do. I don’t know how much SBTs go on
in female groups to each other. They tend to really focus their
SBT behavior on males, but it would be interesting to see. Because you can imagine that that network
has to also function, in its own way. And it may be in females grooming networks
are more important when you have same sex. And I don’t know about in all males. But, or even like in a harem kinda group. I mean, it would probably be
in somewhat different dynamic. But I think that there might be
some application to these kinds of techniques there, yeah.>>Well I certainly have a new
appreciation for the **** eating grin.>>[LAUGH] But that being said,
two questions. The natal male, the one that’s
the troublemaker, when you remove him, can you put him into another situation,
and does that behavior change, is he always gonna be sort
of the trouble maker? And, the social structure as far as what you’re seeing just exists
in a captive environment. Do you see it in a natural environment,
whether or not in a captive environment?>>Great question. So the first question,
what I would say is, I mean, it’s always been sort of my dream to,
in a captive situation, is to remove natal males, at the time
when they would normally disperse, put them into bachelor groups In
corn cribs or something like that. And then once we know how to
introduce males into groups, we introduce those males into groups
where they don’t have related kin. They don’t have kin. And then they’re gonna have
to fight their way up. Then they’re gonna have
to fight their way up. And if they don’t make it,
then they won’t be high-ranking. I suspect that in a novel situation,
they’re probably not gonna be troublemakers, because the kin alliances
alone tell you what’s going on there. Second question, a really good question, because we’re just about to
embark on doing field work. I just send out two post
docs to Malaysia and India to start a study on
human Macaque conflict. And we are gonna find out how
applicable this is to that situation. Rhesus macaques, nobody’s really
studied wild rhesus macaques. They live with people. So I actually suspect that the natal males
stick around in the commensal groups. Too many good resources. But we’re gonna find out. So we’re really excited about that. I’m actually going out there in less
than a month to help them set up.>>All right, last one.>>Hey, Brenda, thanks for all that
information, it’s really very interesting. Do we have a feel for,
from the time that the silent bear teeth networks start breaking down,
to the time that we get collapse? Are we talking a week, a month,
days, what are we talking about?>>It varies, it does vary. In the cases where it collapsed
afterwards, it could be anywhere form a couple of months, but
we also see it within a few weeks. So you want to monitor it enough
to know that that’s happening, so that you can go ahead and
try to change the social group out, so that it doesn’t have that problem,
remove who you need to remove. But we don’t have a really
good idea on that yet, just simply because of the sample size. There is some variation.>>Thanks again to Brenda. [APPLAUSE]

Leave a Reply

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