Documents
Resources
Learning Center
Upload
Plans & pricing Sign in
Sign Out

Lecture 4 - iSites

VIEWS: 2 PAGES: 133

									Studies on Insect
    Cognition


          Lecture 4
          Psych 1090
  Now, we’ve spent a few
lectures discussing animal
        cognition,


 but concentrating on
 creatures mostly like
  apes, monkeys, and
       parrots…
  Creatures with either large
   brains or brains that are
 organized somewhat like ours

but what about the insect world?


  Until fairly recently, “insect
  cognition” was considered an
            oxymoron
For example, ant brains are among
  the largest per body weight in
             insects…

 An ant brain has about 250,000
           brain cells

mushroom shaped brain appendages
 have function similar to the gray-
     matter of human brains.
  But a human brain has 10,000
        million brain cells

  It has been estimated that an
  ant's brain may have the same
    processing power as an old
      Macintosh II computer

   Something that carries out a
program, but can’t think on its own
   Behavior that seemed
intelligent in creatures like
   bees was dismissed as
something rotely specified

 immutable and inflexible

impervious to environmental
         influence
     And, to some extent, a
  significant amount of insect
  behavior is indeed like that…

If you put a ‘foreigner’ scent on a
   worker ant in some colonies,

  the other workers attack, and
 nothing can be done to stop the
             attack
     But, as we will see, other
   aspects of insect behavior do
   seem to reflect some levels of
           intelligence….

  learning, flexibility, adaptation
     to various circumstances

and, at least sometimes, at primate
             -like levels
We’ll concentrate today on bees
          and spiders…


Not because ants—or a number
    of other insects—are
        uninteresting

  But just as representative
           critters
Jumping spiders are particularly
           intriguing


 because they stalk, chase, and
essentially hunt other critters…

a far cry from the view of a lump
  of protoplasm sitting on a web
   waiting for a meal to appear
   In general, the spider we’ll
      discuss today, Portia,
  Uses active mimicry to catch
other spiders in their own webs…

As we’ll see, Portia lands on webs
 and mimics the mating or prey
  behavior of the prey spider
munching the occupant that comes
          to investigate
  But Portia is also able to adapt
 to other types of prey and their
   situation in a given habitat…

   In at least in one area where
another type of prey spider exists,

 Portia has evolved a strategy to
          hunt it as well
 The species, Euryattus, are
 also rather unusual, in that

 only the juveniles spin webs


Instead they inhabit curled up
  leaves that they suspend
     Euryattus females also build
          suspension nests
  that the males visit and on which
   they perform special vibratory
               dances
  The female is lured out and either
   mates or drives the male away

And Portia capitalizes on this behavior
    So to test exactly what Portia
   could do, the researchers gave it

Ø a Euryattus female in her nest

Ø a vacant Euryattus nest as one control

Ø a choice between a juvenile Euryattus
and a juvenile of another related species


Ø a Euryattus juvenile in its web
   The experimenters also used
Portia of different ages, raised in
 the lab w/o experience in nature

  in order to see what could be
  changing with maturation w/o
           experience

 that way they could separate out
learned behavior from what would
       just appear with age
 They also tested male and
       female Portia



     Conceivably, some
difference could exist in the
behavior based on nutritional
 needs of egg-laying females
 They also tested Portia from
 areas in which the Euryattus
     spiders did not live

  that comparison would allow
   them to determine if the
behavior had evolved only in the
  specific area in which it was
              used
When given just the suspended
 Euryattus female in her nest

Only adult Portia females of the
 type that were sympatric with
    the Euryattus in the wild

  succeeded in capturing and
     killing the Euryattus
   And only that same subset
      of Portia in the lab

    engaged in the specific
“shuddering” behavior exhibited
     by the male Euryatta

   despite never having seen
     these males in the lab
  If you looked only at final
          outcomes,

you’d figure that this behavior
 was just something that had
  evolved for this particular
           situation

and that developed without any
      learned component
 But you also have to look at the
   specific processes in which
        Portia engaged…..


Ø Portia ‘waiting at the door’ after
an initial failed attempt

Ø Portia tracking whether it had
been observed by Euryattus and
freezing
Was Portia learning something
 from these interactions?

How much of the behavior was
  fixed and how much was
         flexible?

  Also, look at the defense
  strategies of Euryattus…

  some of which succeeded
Note that Portia males and
juvenile females attempted
  the successful behavior
          patterns


 But did not succeed and did
not seem to ‘learn’ from their
           failures
Tentatively, such data suggest
some kind of interplay between

     some innate, fixed
        program and


some maturation of the ability
  to both learn, adapt, and
         persevere
   But let’s keep going….

What if you give Portia some
 empty Euryattus nests?

Is Portia fooled in some way?

Will it perform the behavior,
  hoping that its prey will
      eventually appear?
Well, yes, female Portias do go
onto the leaves of the empty
  nests and stick around…

 So that tells us that Portia’s
 behavior is not triggered by,
say, the scent of a prey animal

But rather the observation of
     an appropriate nest
      What about males and
    juveniles of the sympatric
              Portia?


and what about allopatric Portia
 (those w/o Euryattus in their
         natal area)?
 Sympatric Portia adult females
went and stayed more frequently
   than other Portia species

  Sympatric adult males acted
 similarly, tho’ females ‘waited’
               more

So males either didn’t learn when
driven off or didn’t care to wait
 Note that earlier studies had
   already shown that males
 were less efficient predators

 But if they don’t need extra
      energy to lay eggs,


that might be evolutionarily fine
Allosympatric Portia pretty
   much did not respond


Suggesting that some genetic
component had evolved only in
        those Portia


 that had this type of prey
         available
 Of interest, however, was that
  the juvenile sympatric Portia


   didn’t engage much in this
            behavior


   was it simply because the
behavior pattern had to mature?
 Or was it some “knowledge”
that they were just too small
   to take on the Euryatta

Euryatta females are about
      the same size,


 and actually are predators in
       their own right….
 Of course, one can imagine an
  evoltionary “just-so-story”


In which those Portia for which
the behavior matured too early


 were themselves eaten and
 taken out of the gene pool…
So we really can’t argue for a
level of conscious decision on
  the part of the juveniles


And we can’t yet do any tests to
 separate out decision versus
        genetic wiring
 Note that if Portia faced off
  Euryattus that were not in
        nests or webs

  Portia succeeded when the
   Euryattus were small or
            medium

But didn’t bother with a large one
Whether that was because
the larger Euryattus could
   more easily get away


and other prey was available


         is unclear
   Interestingly, too, is that
    Portia didn’t seem very
     interested in juvenile
    Euryattus in their webs


  Remember that Portia’s usual
predatory technique is to mimic a
         mate on a web
    Somehow Portia recognizes
     the species on the web as
               either

 (a) too small to be worth pursuing

                or

  (b) juvenile and not interested in
pursuing a potential mate on the web
  Note, too, that the Euryattus
 somehow did recognize Portia as
     something predatory….

and tried to fight it off right away

   Portia is the only spider that
   will engage in these particular
         behavior patterns…
     other than conspecifics….
  Most other prey of Portia do not
           recognize it


Or at least not in time to escape from
   Portia once it is on their webs…


   Has Euryattus evolved some kind
       of “you-me” distinction?
The authors suggest some kind of
    evolutionary ‘arms race”

 which means that each time one
species evolves some technique to
          its advantage,

 the other species evolves some
 technique that overcomes this
     advantage, ad infinitum
    What is clear, however, is
   that even in what appear to
    be set behavior patterns,


     elements of decision and
          choice appear

arguing for at least some behavioral
      flexibility and cognition
   In fact, Portia is quite the
    expert when it comes to
  flexibility and at least some
       forms of learning…

And even if we don’t want to call
the behavior advanced cognition,


     it’s quite impressive…
Again, the critical issue is that
      Portia has evolved

    not to sit and wait for
   something to hit its web,


   but to actively hunt other
        jumping spiders
And, what is more important,

     is that Portia is a
         generalist…


     a critter that is not
  specialized for just a few
        types of prey
Portia engages in what appears
 to be classic trial-and-error
           learning…

   try a bunch of different
      behavior patterns


see which one seems to work the
    best in a given situation
  and then concentrate on the
       winning strategy!

 But what makes Portia unusual


is that it forgets or erases what
  worked the most recent time

 so as to be able to start from
  scratch in a novel situation
     So, one might argue that
       failure to recognize
      something familiar is a
             drawback

 and evidence of stupidity rather
         than intelligence

but in the life of Portia, such is not
               the case
    Because even if Portia finds
     the web of a similar prey
              item,

 The specific conditions are likely
            to differ….

and if Portia misreads the situation,

   it might itself end up as dinner
   In demonstrating this
behavior, Jackson and Wilcox
 again used Portia that were
       raised in the lab

  Thus the individuals had no
 previous experience with the
specific prey items used in the
          experiments
The experimenter first put Portia
  on a three webs of the same
     species of prey spider


And showed that Portia adjusted
   its behavior to each of the
different situations for the same
               prey
 Thus showing that it wasn’t
  something specific about
   the web that triggered
      Portia’s behavior

But rather the actions of the
    individual prey items

that led to the different types
       of Portia’s actions
  Thus Portia was sending
   out a set of signals…

 determining what the prey
was doing in response (which
   differed in each case)


 and adjusting behavior in
         response
  And, yes, one might argue that
 there are just so many different
combinations and that these could
         be programmed…

 But programming (hard-wiring)
   takes a lot of neural space

 And the spiders don’t have that
 much; learning is more efficient
 Next, the experimenters tagged
  prey spiders with magnets so
 that the prey could be made to
         move on demand

  They reinforced some random
  movement of Portia’s with the
         prey movement

  And saw that Portia repeated
these randomly reinforced motions
Portia also could be ‘duped’ by
fake movement if it could see
         a prey spider


And if a signal that was initially
   reinforced was no longer
  reinforced, Portia stopped
          repeating it
 And Portia could be duped into
 repeating a signal if it saw, but
   not felt, a response by the
           prey spider

Such data suggested that Portia
 used visual and vibratory cues

and could quickly alter use of cues
  And, too, the experimenters
 choose only about five signals
out of the 100 or so that Portia
          could make


   but repetitions were still
  limited to those that were
          reinforced
   Other data (Tarsitano and
Jackson, 1997, Animal Behaviour
53, 257-266) suggest that Portia
      has a “cognitive map”

That is, if given an overview and
then two indirect routes to prey

 Portia more often chooses the
          correct one
   And we’ll talk a lot more about
   cognitive maps and what they
       mean in bees in a bit

 But remember that such ability will
help Portia take a circuitous route to
               its prey

which could be very important in not
         getting detected
 Interestingly, when Portia of
 purportedly the same species
   but of different habitat
       were compared…

One with high prey diversity and
  one with low prey diversity

  and neither with any hunting
           experience
  The Portia that came from the
   high prey diversity area used
  more trial-and-error behavior

  suggesting that what was maybe
genetically fixed is the USE of such
               behavior

and not a particular set of responses
  Another study (Jackson, Carter,
  Tarsitano, 2001, Behaviour 138,
            1215-1234 )

    showed that Portia will use
    trial-and-error learning to
        escape confinement

suggesting that such learning can be
      adapted to other tasks
 But Portia’s “cleverness” doesn’t
      stop at trial-and-error
             learning….

Portia also uses various background
    noises to mask its stalking
             movements


   Thus Portia has to be keenly
    aware of its environment
In the very basic experiments,
Jackson and Wilcox put Portia
   onto the web of its prey


  then disturbed the web with
  either wind or a magnet that
 mimicked the prey of the prey
species (e.g., an insect caught in
            the web)
Portia consistently moved closer to
      its prey during times of
            disturbance

  and there were no sex or age
      differences involved


 Now, it could be that Portia just
   reacted to web disturbance
   So the researchers tested
  whether Portia was attending
 to what its own prey was doing
     during the disturbance


Remember, Portia sometimes was
on the web of a spider that could
    attack and eat it as well….
   First, they made sure that
   Portia would not respond to
   disturbance if the web was
      empty of its own prey


   or if it was on the web with
something that didn’t need stalking
           such as a moth
         Interestingly, if
    disturbance was constant,

    Portia was generally more
successful, but sometimes gave up


  Possibly because the prey will
    sometimes leave the web
   Basically, the wind interfered
     with the preys’ ability to
            detect Portia

 but not Portia’s ability to detect
  the motion of the prey, likely

because Portia uses visual as well as
          vibratory cues
 Further work by Jackson and his
  colleagues (Ethology 106, 2000,
 595-615) showed that Portia will
also create its OWN smokescreen,


Setting up pulses of brief, strong
 rocking motions to confuse the
  prey spider and hide its own
            movement
Sci AM
   So, now that we know at
    least some spiders are
   smart, what about bees?

 The article we read was a bit
 old, but provided a really nice
     review of bee behavior


I’ll add some new stuff as well…
 First, let’s state that a honeybee
   brain has only about 960000
   neurons (I’m assuming this is
  what was meant in the ant wrt
              brain cells)



So it’s got a lot more than the ant,
but it’s also a lot bigger than an ant
We’ll see that bees go beyond simple
  stimulus-response associations

    And that they seem to draw
inferences, at least with respect to
 what is ecologically valid to them…

  Note that, like ants, they have
mushroom bodies, which are sort of
  like human grey matter—areas
      related to ‘intelligence’
  So, these seem to be
ideal critters in which to
examine insect cognition
 Nevertheless, Gould doesn’t look
 at cognition quite the same way
        that I define it…

I want a subject not only to figure
 out the correct answer to a given
               task

but to be able to do so for a wide
         variety of tasks
  But bees seem to do fine on a
   number of different tasks
    related to their survival


 And, as Gould states, there is
definitely the need to access the
   capacities in terms of the
         ecological niche
  One of the big issues in bee
behavior was how the bees stored
     their representations…


And not only of a particular flower


   but of the area in which they
              foraged
According to some data, animals
  recognize mirror images as
   identical to the original…

Note that mirror images are NOT
 simple rotations of the items…


 Any arboreal animal had better
     understand rotation…
 And, of course, the issue of
what exactly one is asking the
  animal is truly important



  Remember the study that
   couldn’t understand why
pigeons were sorting blue and
   green things together?
 So asking whether the bee
       can distinguish



            from



is not the same as asking if it
 thinks they are quite similar
      Is the animal ‘stupid’
    because is ‘suffers’ from
    mirror-image ‘confusion’?

    or is it ‘smart’ because it
 ‘understands’ the ‘relationship’
between mirror image ‘reversal’?


Depends on how the experimenter
     is defining the issue!
  When researchers found that
  bees couldn’t understand 90
           rotations

    the data didn’t make any
            sense…

Until they realized that they were
using a vertical format specific to
         the laboratory….
And that bees didn’t view flowers
vertically, but did so horizontally

 And, as Gould mentions, flowers
have a specific type of symmetry,
 as least as far as the bee sees


And, luckily, bees could easily be
       trained to win-stay
 Obviously, researchers have to
    be very careful, the more
  different an animal is from a
             human,


 To design experiments that will
reproduce the world of the animal


 rather than that of the human
   Now, we obviously didn’t read
  about the symbolic bee dances

 And that is because I think that
  the topic is well-covered in a
number of other courses.. and we’ll
        see some video…

And, if needed, a nice review is in
    Griffin’s “Animal Minds”
 And cognitive maps are just
      as controversial

Various papers argue that bees
 aren’t really using such maps


 But rather some form of path
  integration, concerning the
    distances and directions
     But let’s see the issues
    starting w/ the material
      reviewed by Gould….


 Let’s start with the idea of a
         cognitive map…

A mental representation that has
       various landmarks
 So that you can decide the best
  way to get from one point to
             another


Even if the route involves a detour
 or if you have never taken that
           path before
      So, the idea, is that if I
     know something about the
          Harvard campus,

   And the usual route between
   Wm James and the Oxford St
    garage is closed because of
           construction

I can still figure out an efficient way
     to get to my car after class
The issue
     is
 whether
  the bee
  can use
landmarks
   Does the bee notice various
 aspects of its environment as it
       searches for food…

   And does it remember these
    various aspects over time?

And can it integrate these aspects
  over a fairly large distance?
  Although the idea makes a lot
      of sense for a human,

the question is how much sense it
       makes for a bee….

And Gould provides strong support
for his thesis; I won’t go over the
           details here
The real issue is how a brain
that is so small, even relative
     to the body weight,

stores this kind of information


 and that is something that we
have not yet completely worked
               out
     In fact, another set of
 researchers, Kirchner and Braun
   (Animal Behaviour, 1994, 48,
  1437ff) argue that bees don’t
         have this map…


They interrupted the bees’ flight,
put them in a wind tunnel turned at
        a different angle
   And when the bees were
  released, they didn’t act as
      tho’ they had a map,


but rather flew and danced as tho’
    the extra distance and the
direction from the tunnel was real

How do we deal with these data?
Well, one issue is that the wind
tunnel was a real experience for
           the bees….

  It was open on top, but the
      sides were striped

The arrangement was like this:
 Open to
  the sky
gave them
    the
direction,
which was
 90 off
 the path
 Conceivably, the experience was
  somehow part of the cognitive
     map that they stored….

And, as Gould states—and we also
 know from homing pigeons—the
bees will use whatever is available

odor, landmarks, color, shapes, etc.
 And they use them in whatever
  order is most likely to help

And, conceivably, they update this
     information as needed…


Support for this hierarchy comes
from other work by Gould on maps

    Ones that ‘fake’ the bees
      Bees
      were
X
    ferried
    out to a
    station
     in the
      lake
 When these bees returned and
 danced to tell the others of the
           food source


 They were ignored because the
other bees “knew” that no flowers
  grew in the middle of the lake
     When
       the
      scout
      bees
    were put
X   on a boat
    near the
      shore
    and then came back and
  danced about that resource



they were followed because that
    more or less made sense
  according to what the bees
         already knew
   More recently, Sherman and
   Visscher (Nature, 2002, 419,
  920-922) have shown that the
 bees really attend to the dances
       when food is scarce…



  food-location information in the
dance is presumably important when
      Food sources are

 Ø hard to find
 Ø variable in richness
 Ø ephemeral


Such that any extra clues would
     be extremely useful
    Remember, most of these
           studies,


 even if carried out in open land


still use hives that are opened and
       shut and maintained by
             researchers
  So, we next look at a paper that
     puts the bees into a really
      controlled experimental
             situation…

To see if it can understand ‘same’ and
               ‘different’

      according to the rules we
        discussed last lecture
Well, even in the abstract, the
        authors admit

  that what they will show is
   match-to-sample versus
     nonmatch-to-sample

 which we know isn’t really a
 concept of same/different
 But how complex a concept can
       these bees learn?

  Here the bees were, as were
   pigeons, trained to see A

And then asked to choose A or B


Then given C, with choice of C or D
 The interesting thing was
that bees got the match-to-
sample idea in only about 60
          trials….

 Which was an order of
magnitude faster than the
        pigeon…

   Now, think of why…..
 If you are a bee, colors and
    shapes are CRITICAL
     indicators of food….

 Such is not really true for a
            pigeon

 One problem with this paper is
 that we do not know how many
trials were needed for transfer
  Even
tho’ the
data are
 good,
we need
  first
 trials
 The bees were not, however,
rewarded on the transfer trials

   Which suggests that they
   probably were not learning
    anything about the novel
             stimuli

 Although the were ‘retrained’
after each set of transfer trials
 And what seems weird to
 me was that the bees had
  a 90 degree rotation on
    the vertical lines….


Which other researchers had
 shown was really difficult
     So, although the bees
   probably haven’t got a true
    sense of same-different

 They do, with their very tiny
           brains,

manage an ecologically relevant
matching/nonmatching paradigm

    and faster than pigeons
Recently, Dyer and his colleagues
  (J. Expt’l Biology, 2005, 208,
 4709ff) argued that bees can
    recognize human faces…


Something for which they would
 have absolutely no ecological
       predisposition…
He and his colleagues fastened a
  portrait above each of four
            feeders


  They used portraits of men's
faces from a standard test used
to diagnose people with cognitive
            deficits
   One picture was above two
   feeders with sugar water

Others, a stylized cartoon and one
real one, were above feeders with
     quinine, which bees hate


Bees learned to go to the feeder
       with the one photo
 The researchers then switched
  the photos around, giving the
    bees the other pictures

Bees consistently chose that one
        picture EXCEPT


when the photos were upside down
 Data which are consistent with
   the original work on bees
  having trouble with vertical
           rotations…

Now, did the bees really learn to
       recognize a face?


   or just a specific pattern?
     Probably just the latter


 Given that the expression never
             changed


Nor the ‘normal’ angle from which
          it was viewed
  Nevertheless, it was a
  fairly complex form of
      discrimination…


And that was really all the
 researchers wanted to
      demonstrate
    Now, remember that most of
     this material has been on
           honeybees….
 And it isn’t clear that bumblebees
   or stingless bees will respond
        quite the same way…

In fact, work on bumblebees suggest
slight brain differences compared to
              honeybees
 although the researchers aren’t
 entirely clear about what these
 differences might mean (Brain,
Behavior, and Evolution, 2005, 66,
              50-61)


 They do remind us that honeybee
brains increase in size with age and
       foraging experience…
  which suggests some role for
           learning ….

But they also suggest that larger
brains are needed for creatures
     that fill multiple roles…


 which honeybees do sequentially
      In honey bees, workers
     perform different tasks
      depending on their age:

young workers care for the brood,

older ones guard the nest entrance

  and the oldest workers forage
 Thus the oldest ones are the ones
   that experience the greatest
     variation in environment,

  And probably need the greatest
flexibility in their behavior patterns

    And only foraging honeybees
        have been tested….
So, with all that
said…let’s look at
some bee video…

								
To top