LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN Box to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:lCIRC/DateDuo.pBS-p.15

 

l

 

SIMULTANEOUS DISCRIMINATION LEARNING AND ITS NEURAL
CORRELATES IN THE CUTI'LEFISH SEPIA OFFICINALIS
(CEPHALOPODAzMOLLUSCA)

By

Miranda Alderson Karson

A DISSERTATION
Submitted to
Michigan State University
In partial fulfillment of the requirements
For the degree of

DOCTOR OF PHILOSOPHY
Department of Zoology

2003

SIM

ngh
western big
learning in 1

discriminan

 

to exit a tw ‘
was then cc
Spatial Cues
both the WS
cWill sub}
ham t0 mal
visual. and :
performanu
Where bOih

Exit. After n

Immunohistc

ABSTRACT
SIMULTANEOUS DISCRIMINATION LEARNING AND ITS NEURAL
CORRELATES IN THE CUTTLEFISH SEPIA OFFICINALIS
(CEPHALOPODAzMOLLUSCA)
By

Miranda Alderson Karson

Highly controlled behavioral experiments, immunohistochemistry, and
western blot analysis were used to examine visual and spatial discrimination
Ieaming in the common cuttlefish (Sepia oficinalis). A simultaneous
discrimination maze was designed and cuttlefish were trained against preference
to exit a two-choice maze using either visual or spatial cues. Maze perforrnanoe
was then compared to that of a motor control group, which received visual and
spatial cues, but were not reinforced for choosing between the cues. Cuttlefish in
both the visual and spatial groups yielded improved exit performance relative to
control subjects after repeated training in the maze, indicating that cuttlefish can
learn to make simultaneous discriminations based on either type of cue. Both
visual- and spatial-cue trained cuttlefish also yielded significantly improved
performance compared to cuttlefish previously trained in a similar experiment,
where both visual and spatial cues were relevant to selecting the appropriate
exit. After maze training, cuttlefish were sacrificed and brains prepared for either
immunohistochemistry or western blotting to examine calexcitin expression.
Calexc'rtin is a protein that yields increased expression during associative
Ieaming in mollusw as well as some vertebrates. Cuttlefish central nervous

systems immunolabelled for calexcitin yielded calexcitin distribution within

 

indicaIEG g

cuelrainec
Cuttlefish yl

than in the

 

illl‘lherquar
all CUTilefist
thelmmunc
theomic lol
Whitefish |
”ghtand Ie
the initial p,
lobes. This
OVErall, n a
Earning in
spafial

and

relate direc

Ieaming related-brain areas including the optic lobes, vertical lobe, subvertical
lobe, inferior frontal lobe, and superior frontal lobe as well as motor regions
including the median and dorsal basal lobes. Cell counts, comparing intensely
calexcitin stained cells in cuttlefish from the visual, spatial and control groups
indicated greater levels of intracellular calexcitin within the optic lobes of visual-
cue-trained cuttlefish than did spatial-cue trained or control cuttlefish. Control
cuttlefish yielded greater levels of intracellular calexcitin in the subvertical lobe
than in the visual and spatial cue trained cuttlefish. Western blotting was used to
further quantify calexcitin expression in the supraoesophageal and optic lobes of
all cuttlefish using homogenate, cytosolic and microsomal fractions. Similar to
the immunohistochemical resutts, greater quantities of calexcitin were found in
the optic lobes of visual trained cuttlefish than in spatial cue trained or control
cuttlefish. Experiments also indicated differences in expression between the
right and left optic lobes. However, these differences did not appear to relate to
the initial preference of the cuttlefish nor to differences in size between the optic
lobes. This may suggest lateralization of function between the optic lobes.
Overall, it appears that calexcitin expression is involved in visual discrimination
Ieaming in cuttlefish. The experiments here indicate that cuttlefish show both
spatial and visual discrimination Ieaming ability, and calexcitin expression may

relate directly to associations of visual stimuli.

 

THANKS MOM!

 

This
number ot I

rest of my 5

 

and Robert
thanks to P
Laboratory
dunng EXp
Mollie Tub
tank mainr
Nelson at
N‘aiyland
laboraton
much Of I
With he.

“elated it
to thank

years F

To all 0‘

ACKNOWLEDGEMENTS

This project would not have been possible without the help of a large
number of people. Many thanks to my advisor, James Atkinson as well as the
rest of my graduate committee, Heather Eisthen, Thomas Getty, Roger Hanlon,
and Robert Root-Bemstein for their helpful comments and support. Additional
thanks to Roger Hanlon and the Marine Resources Center (Marine Biological
Laboratory, Woods Hole, MA) for use of laboratory space and technical support
during experimentation. Particularly, I am thankful to Janice Hanley, Bill Mebane,
Mollie Tubbs, and Nicole Gilles and for their daily assistance in animal care and
tank maintenance. I am very appreciative of Dan Alkon, Deanna Buck and Tom
Nelson at the Blanchette Rockefeller Neurosciences Institute, Gaithersberg
Maryland, for their incredible generosity and assistance in training me in various
laboratory techniques. I am grateful to Jean Boal for her continued guidance,
much of my knowledge and interest in cephalopods is a direct result of working
with her. I am also appreciative of recent correspondence with John Messenger
related to interpretation of the results of these experiments. Finally, I would like
to thank my family and friends for their support and encouragement over the
years. Financial support for this project was provided by Sigma Xi, the Marine

Biological Laboratory and the Blanchette Rockefeller Neurosciences Institute.

To all of the above, my heartfelt appreciation.

TABLE OF CONTENTS

List of Tables

List of Figures

Chapter 1: Introduction to the Cephalopods
1.1 Classification
1.2 Evolution of Coleoid Cephalopods
1.3 Natural History of Sepia ofl‘icinalis

Chapter 2: Cephalopod Learning
2.1 Sensitization/Habituation
2.2 Associative Learning
2.3 Spatial Learning

Chapter 3: Cephalopod Brains
3.1 General Description
3.2 Description of Supraoesophageal Lobes
3.3 Role of Supraoesophageal Lobes in Learning
3.4 Development of Behavior in Sepia
3.5 Proposed Functions of Vertical Lobe
3.6 Comparisons with Vertebrate Hippocampus

Chapter 4: Calexcitin
4.1 Associative Learning in Hermissenda
4.2 Protein Kinase C and Associative Learning
4.3 Role of Calexcitin
4.4 Comparisons with Other Models
4.5 Calexcitin in Cephalopods

Chapter 5: Hypotheses

Chapter 6 Behavioral Experimentation
6.1 Summary
6.2 Materials and Methods
6.3 Results
6.4 Discussion

Chapter 7: lmmunohistochemistry
7.1 Summary
7.2 Materials and Methods
Results
7.3 Localization of Calexcitin
7.4 Comparison of Calexcitin Expression
between experimental groups
Discussion

vi

(DAN—‘-

11
14

19
20
21
24
29
31
32

33

36
38
40

41

45
46
46
49
53

57
58
58
60

63

 

 

Chapter 8,
81 i
8.2 l

 

83¢
8.4[

Chapter 9:
9.1 '
9.2l
9.3l
9 4
9.5

7.5 Patterns of Expression
7.6 Group Comparison

Chapter 8: Western Blots

8.1 Summary

8.2 Materials and Methods
Fraction Preparation
Protein Content Determination
Electrophoresis

8.3 Results

8.4 Discussion

Chapter 9: General Discussion
9.1 The Function of Learning in Cephalopods
9.2 Relevant and Irrelevant Cues
9.3 Individual Variation in Behavior
9.4 Calexcitin and Visual Associations
9.5 Time Course of Learning and Calexcitin
Expression
9.6 lnterocular Transfer
9.7 Future Directions

Appendix A: Tables
Appendix B: Figures
Appendix C: Reprint
Experimental Evidence for Spatial Learning
in Cuttlefish (Sepia oflicinalis)

Literature Cited

vii

66
7O

73
74
74
74
76
77
79
80

85
89
90
91

95
97
98

114

137

164

 

 

 

Table 1
Brain lobe

Table 2
Initial desc

l
TNE3

Reoealeo' ‘
maze peric
exDer-Intera

Thb4

ngg
andlastb
Cuttlefish

Table 5

Mean 1 5.
middle ar
Cuttlefish

Table 5
Mean nu
brain re
Table 7

LIST OF TABLES

Table 1
Brain lobes associated with the supraoesophageal mass

Table 2
Initial descriptions of the cuttlefish

Table 3

Repeated Measures Analysis of Variance results of cuttlefish
maze performance across the first, middle and last block of
experimental trials

Table 4

Mean 2 standard error of maze escape times for the first, middle
and last block of five trials during training for male and female
cuttlefish in the control, visual, spatial and combined cue groups

Table 5

Mean 1 standard error of overall percentage of error for the first,
middle and last block of five trials during training for male and female
cuttlefish in the control, visual, spatial and combined cue groups

Table 6

Mean number of intensely stained calexcitin cells in various cuttlefish
brain regions post-training

Table 7

Kruskall-Wallis One Way Analysis of Variance results for differences
in the mean number of strongly calexcitin stained calls in various
brain regions

Table 8

VVIIcoxon signed ranks test results, comparing the mean number of
strongly calexcitin stained cells in the left and right optic lobe for each
experimental group

Table 9

Data on the initial tissue weight (grams) of the pieces of tissue used
for western blot analysis.

viii

99

100

101

102

103

104

105

106

107

Table ID I
Bradford m
for each hall
of the left c
of each cu.

 

Table II

Mann Whit
on western
microsoma
SUDraoeso

Table 12

Wilcoxon SI
WeSiem blo
experiment

Table 10

Bradford method results used to determine the protein concentration

for each fraction (homogenate, cytosolic and microsomal fractions)

of the left optic lobe, right optic lobe, and supraoesophageal lobes

of each cuttlefish. 108

Table 11

Mann Whitney U test results comparing mean calexcitin expression

on western blot membranes among the homogenate, cytosolic and
microsomal fractions of the left optic lobe, right optic lobe and
supraoesophageal lobes between each experimental group. 112

Table 12

lMlcoxon signed ranks test comparing mean calexcitin expression on

western blot membranes in the left and right optic lobe for each

experimental group. 1 10

ix

 

Figure I
Adiagrarn

Figure 2
Saggital St

Figure 3
Transverse

Figure 4 _
Diagram Ill
associative

Figure 5
Schematic
Viewed hor

 

Figure 8
Mean esca
block of flu.

Figure 7
Mean Over.
biOCk of fiV'

F'lgure 8

COmDBTlSo
between Ct

CU9~trai ”QC

Figure 9
orimiltal
' OffiCIna/I‘

FiQUre 10

FigUre 11

Drizorltals

HOUZOnta, S

LIST OF FIGURES
Figure 1
A diagram of the brain of an octopus.

Figure 2
Saggital Section of the supraoesophageal mass of Sepia ofi‘icinalis

Figure 3
Transverse section of the deep retinal region of Sepia officinalis.

Figure 4
Diagram illustrating the major cellular components underlying
associative memory in Hermissenda.

Figure 5
Schematic drawing of the Simultaneous discrimination maze as
viewed from above.

Figure 6
Mean escape time 1 standard error over the first, middle and last
block of five trials for cuttlefish in each experimental group

Figure 7
Mean overall error = standard error over the first, middle and last
block of five trials for cuttlefish in each experimental group.

Figure 8

Comparison of mean number of trials to criterion level performance
between control, visual cue-trained, Spatial cue-trained and combined
cue-trained cuttlefish.

Figure 9
Horizontal sections of the plexiforrn layer of the optic lobe of
S. oflicinalis.

Figure 10
Horizontal sections of the medulla of the optic lobe of S. officinalis.

Figure 11
Horizontal sections of the superior frontal lobe of S. officinalis.

Figure 12

Horizontal sections of the posterior inferior frontal lobe of S. officinalis.

115

116

117

118

119

120

121

122

123

124

125

126

Figure 131
Horizonta s

 

Figure 14
Horzontar :

Figure 15
Horizontal s

Figure 16
Horizontal s

Figure 17
Horizontal s

Figure 18
Horizontal s
8- Ofiicinali:

Fhue 19

PiClein C00
Concentratr
PIOtein C00

Figure 20
NmOCEHUIO
lOr CaieiCIt;

“913621
Nieceiruro

F"lime 22
NEUOCEHUIO
i”mimosa

 

 

Figure 13
Horizontal sections of the vertical lobe of S. officinalis.

Figure 14
Horizontal sections of vertical/subvertical lobes of S. officinalis.

Figure 15
Horizontal sections of the subvertical lobe of S. oflicinalis.

Figure 16
Horizontal sections of the anterior basal lobe of S. ofi‘icinalis.

Figure 17
Horizontal sections of the median basal lobe of S. officinalis.

Figure 18
Horizontal sections of the posterior wall of the dorsal basal lobe of
S. officinalis.

Figure 19

Protein concentration curve of Bovine Serum Albumin at different
concentrations read at 595 nanometers used to determine the
protein concentration of samples for western blotting.

Figure 20
Nitrocellulose membranes loaded with homogenate and immunostained
for calexcitin.

Figure 21
Nitrocellulose membranes loaded with microsomal fractions and
immunostained for calexcitin.

Figure 22

Nitrocellulose membranes loaded with cytosolic fractions and
immunostained for calexcitin.

xi

127

128

129

130

131

132

133

134

135

136

Chapter One

Introduction to the Cephalopods

 

1.1 Classi
Cep
charactenz
diversity of
groups rese
SCOTOgy ant
are consro'e
The
before the f
cePhaIOpod
COleoidea.
Wilde thre
and the sep
Oephalopw
Spaces uset
iItclude the ‘
the CO'SOIGS

ChemiCal “it
1972)_

 

 

NeUt.
CO
”SeqUEnC

OCOMQUOn El

 

1.1 Classification

Cephalopods are members of the phylum Mollusca, an ancient group
characterized by an evolutionarily plastic body plan that allows for a tremendous
diversity of form. The cephalopods are interesting molluscs in that several
groups resemble modern teleost fish in terms of their morphology, physiology,
ecology and behaviors (Hanlon and Messenger, 1996). Thus, the cephalopods
are considered "advanced" invertebrates.

The cephalopods appeared in the late Cambrian, several million years
before the first primitive fish (Teichert, 1988). There are two extant divisions of
cephalopods, the ancient Nautiloidea (the nautiluses) and the more modern
Coleoidea. The coleoids diverged from the nautoloids in the Ordovician and
include three major extant orders, the octopods (octopuses), teuthoids (squid),
and the sepioids (cuttlefishes) (T eichert, 1988). The evolution of the
cephalopods has been largely linked to adaptations allowing for low-pressure gas
spaces used to maintain neutral buoyancy (Packard, 1972). Such adaptations
include the formation of chambered shells (Nautilodea) and those observed in
the coleoids including drastic reductions in the external shell, dynamic lift,
chemical lift, swim bladders, and various behavioral modifications (Packard,
1972).

Neutral buoyancy has yielded a number of interesting behavioral
consequences. First, reductions in the shell have allowed for more rapid

locomotion as well changes in defensive strategies (Packard, 1972). Coleoid

cepha ope
preda' ors

pull th air e
relate 1 to -.'
ceoha opo
nautili 585
Thus. hera
two or Iers

StnICtI res

 

large r Iifac

  
  

adapt: tion
00leoi ls h

mUSCL lam

cephalopods rely heavily on camouflage, speed, or shelter use to evade
predators, compared to the nautiluses, which like many other shelled molluscs ‘
pull their entire body into a shell when threatened (Wells et al., 1992). Directly
related to reductions in shell size are changes in predatory behaviors. Coleoid
cephalopods are fast-swimming, active predators, whereas slow-moving
nautiluses are primarily scavengers (reviewed in Hanlon and Messenger, 1996).
Thus, there are a number of obvious differences in foraging habits between these
two orders and there are a number of related differences in the brain and sensory
structures that are evident (Budelmann, 1994). Specifically, nautiloids have very
large olfactory lobes, and lens-less "pinhole camera" eyes interpreted as
adaptations associated with scavenging (Young, 1965 and 1988). In contrast,
coleoids have large optic lobes, complex camera eyes, increased eye
musculature, and greatly reduced olfactory lobes (Budelmann, 1994). In this
way, coleoid cephalopods are well suited for a highly visual, predatory lifestyle.
However, many octopuses have well-developed chemoreceptors on their suckers
and complex chemotactile memory systems (for review see Young, 1988;
Budelmann, 1994).

The complexity of the coleoid sensory and processing systems relative to
that observed in the nautiluses correlates with the coleoids' increased reliance on
complex external stimuli (visual, chemosensory, tactile, etc.) used to locate food
(Sanders, 1975; Young, 1988). For instance, the coleoid eye shows a
remarkable convergence with the vertebrate eye and is able to discern much

more detail (of objects at a variety of distances) than the eyes of other

invertebraf:
that is sens
the coleow:
lateral line
deep water
(Budelmar:
cephaloooi

Details of ti

 

 

 

are descnb
12 Evoluti
Man
Shell-less. ;
PaiEOZoic‘
Nautilus) w
1988)- The
eIristed bet
SuggeSted
during the I
(Packard, 1
Significant '
returned to
the TriaSsic

hYDotheSQE

invertebrates (Young, 1963). Additionally, a mechano-sensory lateral line system
that is sensitive to local water movement in particular directions has developed in
the coleoids, but is not present in the nautiluses (Budelmann et al., 1991). The
lateral line system allows for prey detection (or predator avoidance) at night, in
deep water, or in turbid water when normal visual processing is difficult
(Budelmann et al, 1991; Budelmann, 1994). Thus, it is clear that the coleoid
cephalopods receive a much broader range of sensory input than nautiluses.
Details of the cuttlefish brain and mechanisms for processing of sensory input
are described in Chapter 3.
1.2 Evolution of the Coleoid Cephalopods

Many hypotheses address the potential influences driving the advent of a
shell-less, active, carnivorous life-style in the coleoid cephalopods. In the early
Paleozoic, the buoyant, chambered-shell forms (somewhat similar to modern
Nautilus) were likely the most mobile and successful cephalopods (Teichert,
1988). There has been a great deal of speculation that intense competition
existed between cephalopods and vertebrates. Specifically, it has been
suggested that ancient fish exerted strong predatory pressure on cephalopods
during the Devonian period, forcing them to occupy niches in deeper water
(Packard, 1972). This may have favored morphological changes allowing
significant reductions in shell size. Shell-less cephalopods may have later
returned to shallower water and competed directly with the teleost fishes during
the Triassic (Packard, 1972). Support for this comes from Aronson (1991) who

hypothesized that predatory teleost fishes directly affected the distribution of

octopuses

 

further fav.
cephalooo.
Tek
because tr

(Packard.
direct com
Webber, 1
almost be
Camoufla

Vertebrar

have {9

plan, w

This Q1
sQth
SlZe E
1988

patcl

Dh)’:

octopuses. In addition, it is likely that predation pressure from the teleost fish
further favored adaptations promoting greater speed and maneuverability in the
cephalopods (Packard, 1972).

Teleost fish and cephalopods probably shared a number of marine niches
because they share similar size ranges, predatory styles and other behaviors
(Packard, 1972). However, the extent to which fish and cephalopods were in
direct competition with one another has been argued extensively (O'Dor and
Webber, 1986). Regardless, there are certain cephalopod adaptations that are
almost certainly directly related to the presence of fish. For instance, skin
camouflage on color-blind cephalopod bodies is probably ”aimed at fooling
vertebrate eyes" (Packard, 1988).

As mentioned before, the ability to avoid teleost fish (or to eat them) may
have required increased speed and maneuverability in the cephalopod body
plan, which in turn required a decrease in the shell and increased coordination
between various sensory and motor structures and muscles (Budelmann, 1994).
This coordination was achieved in large part by the development of a
sophisticated centralized processing unit (a complex brain). As cephalopod brain
size and complexity increased, behavioral repertoires also increased (Young,
1988; Budelmann, 1994). The ability to locate and relocate shelters or foraging
patches, acquire new hunting skills, and generally gain knowledge about the
physical world are all significantly aided by diverse forms of complex Ieaming

(e.g. associative Ieaming, spatial Ieaming, discrimination Ieaming, etc.).

 

TheI

have eight

There are l
between ti
structures

and decapl

 

1.3 Natura

The
thiOughQUj
Atlantic 0:
polnt of At.
living In 0
(no dBEpI

many DO;

meter de
19831. Bc
dunngt,

A
“Page
Substtat

Onthe\

The cuttlefish (order Sepiodea) and squids (Order Teuhthoidea) each
have eight arms and two tentacles are often referred to as the "decapods".
There are several other notable morphological and anatomical differences
between the octopuses and decapods including body shape, fins, and sensory
structures. Thus, anatomical and behavioral comparisons between octopuses
and decapods may shed further light on the adaptive radiation of the coleoid line.
1.3 Natural History of Sepia ofiicinalis

The common cuttlefish, Sepia officinalis Linnaeus, 1758, is distributed
throughout the North Sea, English channel, in the Mediterranean Sea and the
Atlantic Ocean along the west coast of Africa extending to the southem-most
point of Africa (Boletsky, 1983). Adult Sepia are primarily benthic organisms,
living in coastal waters and on the continental shelf in relatively shallow waters
(no deeper than 150 meters). During the summer, temperature increases induce
many populations of S. oflicinalis to migrate to shallower waters of about 30-40
meter depth where they aggregate with conspecifics for spawning (Boletsky,
1983; Boucaud-Camou and Boismery, 1991). The orientation mechanism used
during these migrations has not yet been described.

After a female cuttlefish has mated, She lays her eggs, putting each egg
into a separate case. Egg cases are then attached in clusters to objects near the
substrate (Boletsky, 1983). Eggs take about two months to develop depending
on the water temperature: development is more rapid in warmer water. Female
cuttlefish can lay eggs several times toward the end of their life, but most begin

to senesce shortly after spawning. Male and female cuttlefish have a short

lifespan c‘
approxima
the range ;
adults and
many of t:
exhibit ad 1
between cl
to develo;
there are 5
difficulties
Messengel

Pos

t‘3ml3‘3lalur

 

 

 

a diSilFlCt if
i

lifespan of about one to two years. Hatchlings have a mantle length (ML) of
approximately 50 mm and can grow to a ML of 45 cm but most adults remain in
the range of 20 to 30 cm ML (Boletsky, 1983). Hatchlings closely resemble
adults and have only a very brief planktonic phase. Hatchlings rapidly display
many of the same body patterns shown in adults (Hanlon and Messenger, 1988),
exhibit adutt—like predatory behavior (Wells, 1958) and readily distinguish
between other hatchlings and prey items (Boletsky, 1983). Many details related
to development and Ieaming of adult behaviors in cuttlefish remain vague, as
there are significant difficulties studying adult behavior at sea as well as
difficulties rearing (and thus observing) hatchlings in the laboratory (Hanlon and
Messenger, 1996).

Post-hatching growth in Sepia is largely related not only to water
temperature, but also to feeding conditions (Boletsky, 1977 and 1979). There is
a distinct increase in growth rate at approximately 100 days (juvenile stage) in
25°C water. Male cuttlefish can reach sexual maturity when they are as small as
6-8 cm mantle length (ML). Females may mature at anywhere from 11-25cm
ML. Thus, males have earlier maturity and a longer duration of reproductive
activity than females (Boletsky, 1983). During maturation many aspects of
cuttlefish behavior, such as hunting and aggression and changes in hunting
behavior appear to be a direct result of Ieaming (i.e. Boletsky, 1977; Agin et al.,
1998; Dickel et al., 2000). As more details of the natural history of S. oflicinalis
are clarified, we will undoubtedly understand more about the function of many of

their behaviors (including learning).

  
  
  

extreme c
behaviora
must lea -.
organism .
Messenge
aspects 0
bathing,

discusse

Wells (1962) suggested that Ieaming may allow cephalopods to cope with
extreme changes in body size during their rapid development thus promoting
behavioral plasticity. For instance, young octopuses that outgrow their dens
must learn the location of new dens, or young cuttlefish must learn that an
organism that was once a predator is now a valuable food resource (Hanlon and
Messenger, 1996). Learning is undoubtedly an important factor in several
aspects of the natural history of coleoid cephalopods. Examples of cephalopod
learning, possible functions of Ieaming, and the underlying neurobiology will be

discussed in greater detail in the next two chapters.

Chapter 2

Cephalopod Learning

  

Ber:
of lower vel
more corn
species to

(Packard.

 

learning h
neurobiolc
molluscar
CEDhaIOpI
allowing f

wi“
Earning '
OmODUsi

habituati

Because cephalopods' behavioral abilities have been compared to those
of lower vertebrates and because they are evolutionarily distant from species
more commonly used in Ieaming experiments, cephalopods are worthwhile test
species for functional explanations of the evolution of complex nervous systems
(Packard, 1972; Budelmann et al., 1997). Thus far, most research on complex
Ieaming has focused on vertebrate species or eusocial insects. However,
neurobiological models of associative Ieaming are often based on simple
molluscan systems (eg. Aplysia and Hermissenda). Investigations of Ieaming in
cephalopods could provide important insight into the evolution of cognition, by
allowing for comparisons of simple and complex behaviors

within a single taxon (Phylum Mollusca). Previous studies of complex
learning in cephalopods have focused primarily on Ieaming in octopuses.
Octopuses have been shown to exhibit many forms (or levels) of learning:
habituation, sensitization, operant conditioning, and associative Ieaming have
been demonstrated (Thorpe, 1963). Several extensive reviews of cephalopod
Ieaming literature are available (i.e. Sanders, 1975; Mather, 1995; Hanlon and
Messenger, 1996).

2.1 Sensitization and Habituation

Simple forms of Ieaming such as habituation (increased threshold to
response) and sensitization (decreased threshold to response) have been
demonstrated on several occasions in the coleoid cephalopods. For example,

when encountering a novel object, octopuses generally pass it under their web

10

 

meb,
where the
normal a."
pass the 0

studies ha

 

(69.6qu
NEVIS} Sh:
reRated -

dishablt'u;

Punishm.
SUbsequ
Wells 1g
Mesa-an.

2.2 “133

(formed by the layer of skin that joins the eight arms) and then to their mouths
where they will examine the object. If the same object is presented repeatedly to
normal and blind Octopus vulgaris, neither group will continue to examine or to
pass the object to the mouth after a few trials (Wells and Wells, 1956). Other
studies have demonstrated habituation to various additional stimuli in Octopus
(e.g..,Goldsmith, 1917; Boycott, 1954). Similarly, bay squid (Loligonuncula
brevis) show a marked reduction in escape jets or defensive body patterns with
repeated presentation of a model of a fish (Long et al., 1989). L. brevis will
dishabituate to this same model after a threat stimulus or a one-hour break (Long
et al., 1989). The other form of non-associative Ieaming, sensitization, has also
been demonstrated in cephalopods. During sensitization, rewards or
punishments that are offered before presentation of a test stimulus will affect
subsequent attack behavior for visual, tactile and olfactory stimuli (Wells and
Wells 1958, Chase and Wells, 1986, reviewed in Sanders 1975 and Hanlon and
Messenger, 1996).
2.2 Associative Learning

Associative Ieaming, involving associations between particular events
resulting in long-tenn changes in behavior, has also been demonstrated in
cephalopods (Sanders 1975; Mather, 1995; Hanlon and Messenger, 1996).
Associative Ieaming involves more complex processing of stimuli than
habituation or sensitization (Thorpe, 1963). For example, octopuses will Ieam to
attack one of a pair of stimuli when the attack is reinforced with either a positive

food reward or a negative electric shock (Boycott and Young, 1955). In these

11

cases. lea
erratic tha
1955). Vt
presented

with octop

 

better part
This traini
abilities. ‘

decapods

cases, Ieaming appears quite rapid, though performance is a great deal more
erratic than in vertebrates trained under Similar conditions (Boycott and Young,
1955). Visual discrimination training, in which separate visual stimuli are
presented simultaneously or successively, has also been successfully employed
with octopuses, though simultaneous presentation of stimuli yields Significantly
better performance than successive presentation (Sutherland and Muntz, 1959).
This training protocol has been used extensively to examine visual perceptual
abilities. \fIsual discrimination Ieaming has also been demonstrated among
decapods as well (i.e. Messenger, 1977; Allen et al., 1985).

Tactile discrimination training has also been successfully demonstrated in
octopuses. O. vulgaris will readily accept and pass an object to its mouth
associated with a positive stimulus while rejecting objects that it learns to
associate with negative reinforcement (Wells, 1978; Robertson et al., 1994).
Similarly, avoidance learning has been demonstrated in the octopus Eledone
moschata, which will learn to avoid hermit crabs that would normally be prey
when these hermit crabs are carrying particular species of painful sea anemones
(von Uexki‘rll, 1905; Boycott, 1954).

Associative Ieaming has also been demonstrated in cuttlefish (Messenger,
1968 and 1977). Through associative conditioning, Sepia offlcinalis will Ieam to
not strike at prey items presented in a glass tube (Sanders and Young, 1940;
Messenger, 1968). This inhibition of predatory behavior is categorized as
associative Ieaming as striking tentacles against glass is negative reinforcement

(painful) for the cuttlefish (Messenger, 1977; Agin et al., 1998). S. officinalis with

12

tentacles tr

 

response 1*
Res
learning, h
andYoung
Karson etr
simultane:
Deflorrnar
rewarded
SUCCessn
earner re
9990an
“Wolvlng
OCTOpus
reducec
imDTOVE
trained
l.ndi‘ildi
task. r

s‘°Cial

tentacles that have been surgically removed show a much slower decrease in
response with repeated exposure than do normal cuttlefish (Messenger, 1973).

Reversal Ieaming, which is considered an advanced form of associative
Ieaming, has been demonstrated in both octopuses and cuttlefish (e.g.. Boycott
and Young, 1958; Wells and Wells, 1957; Mackintosh and Mackintosh, 1964;
Karson et al., 2003). Mackintosh and Mackintosh (1964) trained octopuses in a
simultaneous visual discrimination task. Once the octopuses attained criterion
performance, the task was reversed and the unrewarded object became
rewarded. Octopuses more quickly attained criterion level performance across
successive reversals and also attained earlier success on later reversals than on
earlier reversals. Cuttlefish have also been demonstrated to show improved
performance across serial reversals of a simultaneous discrimination problems
involving visual and spatial stimuli (Karson et al., 2003). Similar to training
octopuses, cuttlefish more quickly attained criterion level performance and
reduced numbers of errors across successive reversals. However, significant
improvement was only seen across those reversals in which the cuttlefish were
trained in the direction contrary to their individual preference.

Another form of complex Ieaming is observational Ieaming, where an
individual leams the solution to a problem by watching a conspecific Ieam the
task. Fiorito and Scotto (1992) first suggested that octopuses are capable of
social or observational learning. In Fiorito and Scotto's study groups of
"demonstrator" octopuses were trained to discriminate between visual stimuli.

Once the demonstrator octopuses yielded errorless performance on the task,

13

they were
the field of
shown to a1
visual stim
behavior 0
al., 2000).
octopuses
Qeneratior
result may
se. lnstee
Observer
than the l
2.3 Spat]
Ev
studies L
hOme GE
1933 an
OCIOpUS
monthS
M0005
“9'“ th.

Hank)”

i r

they were then tested without a reward while naive "observer" octopuses were in
the field of view. Observers were then tested without reward, and they were
shown to attack the correct visual stimulus significantly more than the incorrect
visual stimulus. However, in a similar study of cuttlefish, watching predatory
behavior of conspecifics did not yield any observational learning effect (Boal et
al., 2000). These findings have not been successfully replicated. Moreover,
octopuses (unlike squid) are primarily solitary and have no overlap of
generations. These two pieces of evidence suggest that Fiorito and Scotto's
result may not be a case of imitation brought about by observational Ieaming per
se. Instead, perhaps, there was some type of "local enhancement" in which the
observer octopuses attention was directed toward the correct stimulus more so
than the incorrect stimulus (Thorpe, 1963; Hanlon and Messenger, 1996).
2.3 Spatial Learning

Evidence for spatial Ieaming has also been provided by several field
studies using octopuses as subjects. Many octopus species forage away from a
home den to which they return repeatedly for shelter (e.g. Ambrose, 1982; Boyle,
1983 and 1988; Hartwick et al., 1984; Mather 1991; Forsythe and Hanlon, 1997).
Octopuses may occupy a single den continuously anywhere from one day to five
months before moving to a new den (Hartwick et al., 1984). Furthermore, these
octopuses are able to relocate their dens from large distances (40 m, out of sight
from the den) and after extensive time lapses (Mather 1991; Forsythe and
Hanlon, 1997). Octopuses will then change home dens and again use spatial

information relocate these sites after foraging (Mather, 1991 ). Detailed tracking

l4

data frorr
1991; F0."l
COTTCETTTII‘.

Stt

octopuses

 

1991: MOI
maze writ
task as v
Boat et 5
learn to
DTOblen
Sole [Qt
using 2
the 8U
reduC‘
TOT 5:

QCTO;

'98 rr
Tea:
3 v
83{

me

data from octopuses in open water is available from only two studies (Mather,
1991; Forsythe and Hanlon, 1997), thus, one cannot draw conclusions
concerning spatial learning ability from the field data alone.

Studies of detours and maze Ieaming have supported the hypothesis that
octopuses rely on spatial Ieaming (Wells, 1964; Walker et al., 1970; Mather,
1991; Moriyama and Gunji 1997; Boal et al., 2000). Training of O. maya in a T—
maze with escape as motivation successfully demonstrated rapid Ieaming of the
task as well as reversal Ieaming when exits were switched (Walker et al., 1970).
Boal et al. (2000) demonstrated that octopuses explore their surrounds and also
learn to solve a maze within one day and retain the solution to the spatial
problem for at least one week. In this experiment maze escape was used as the
sole reward. Moriyama and Gungi (1997) examined spatial Ieaming in octopuses
using a wall/obstacle maze but provided food as a reward. In this experiment,
the authors cite the fact that octopuses changed actions and significantly
reduced their detouring time around the obstacles after training as good evidence
for spatial Ieaming. Thus, data from spatial Ieaming experiments indicate that
octopuses Ieam relatively quickly and retain the learned information.

In laboratory experiments, cuttlefish may have also demonstrated spatial
Ieaming (Karson et al., 2003; see Appendix C). The cuttlefish in this experiment
readily learned to escape from a straight alley maze, which could be considered
a very simple form of spatial Ieaming. In more complex mazes it was difficult to
ascertain whether cuttlefish learned to follow spatial or visual cues to find the

maze exit, as the cues were linked (Karson et al., 2003). However, these

15

experimer
maze base
Wh
muem
understoor
unhefishr
do not ap

swab

experiments did not rule out the possibility that cuttlefish can Ieam to escape a
maze based on spatial cues.

While learning to avoid painful stimuli has clear utility for cuttlefish in
nature, the utility of other forms of Ieaming (i.e. spatial Ieaming) are not as well
understood. For example, unlike octopuses that use home dens for defense,
cuttlefish rely primarily on camouflage or burying themselves in the substrate and
do not appear to use home dens at all. It is clear, thus, that cuttlefish do not use
spatial Ieaming in the same manner as octopuses.

A number of hypotheses could address the utility of spatial Ieaming in
cuttlefish. Cuttlefish live in complex, three-dimensional environments, and many
species demonstrate particular patterns of spatial use within these habitats
(O'Dor et al, 2001). Cuttlefish may use spatial Ieaming to repeatedly negotiate
vertical barriers, to relocate particular foraging patches or to relocate mating
aggregation areas. For example, S. apama often settle on or near particular
rocks or ledges, forage away from these rocks and then return to the same
location (O'Dor pers. comm., 2002). Spatial Ieaming may be used to relocate
particular features of a habitat (Sanders and Young, 1940). Additional
hypotheses are drawn from field studies of S. Iatimanus, which gather in
breeding aggregations. During these aggregations males position themselves on
specific limestone mounds and wait for females (Corner and Moore, 1980).
When females arrive, males leave their mound to approach the female. After
pursuing the female, a male will return to the mound from which it started the

pursuit of the female. Spatial learning may play a role in helping the male

16

relocate tl

these mc.

  
 
 

Se'
officinalis
involved II
are short-
unlikely it
remains
can solv

fUTICtlon

T
neurob
CaPabl
brain t
FUfihe
3830C
99nd

TOta [ i

relocate these mounds. However, the males may use some other cue to locate
these mounds as well.

Seasonal onshore/offshore migrations have been documented in S.
officinalis (Boletsky, 1983) and it has been suggested that spatial Ieaming is
involved in locating sites for mating aggregations (Norman, 2000). Because they
are short-lived (one-two years) and because Sepia is generally semelparous, it is
unlikely that Ieaming is involved in these migrations (Karson et al., 2003). It
remains clear, however, that by exploring the types of problems that cuttlefish
can solve using Ieaming and the cues they use to solve these problems, further
functions of Ieaming in the natural history of cuttlefish may become apparent.

Though complex learning studies have focused on octopuses,
neurobiological evidence also suggests that cuttlefish, like octopuses, are
capable of relatively complex forms Ieaming. Cuttlefish have a larger overall
brain to body size ratio than do octopuses (Maddock and Young, 1987).
Furthermore, in S. officinalis the vertical lobe, a brain region involved in
associative Ieaming and memory of tactile and visual stimulli (see Chapter 3) is
significant larger. The vertical lobe of S. ofiicinalis occupies roughly 24% of the
total brain volume, whereas the vertical lobe of Octopus vulgaris occupies only
about 13% of the total brain volume (WIrz, 1959; Maddock and Young, 1987). In
S. officinalis Specific improvements in Ieaming ability have been directly
correlated with increases in the volume of the vertical lobe during post-
embryological development (Messenger, 1973; Dickel et al.1998). For example,

healthy adult cuttlefish Ieam to avoid shrimp presented in a glass tube, whereas

17

  
   

juveniles a
at al., 200‘
associate:
learning a.

that obser

juveniles and senescent adults do not (i.e. Chichery and Chichery, 1992; Dickel
et al., 2001). Such findings suggest that vertical lobe size in particular could be
associated with Ieaming (refer to Chapter 3). Thus, the possibility remains that

Ieaming and memory play a significant role in cuttlefish natural history equal to

that observed in octopuses.

18

CHAPTER 3

Cephalopod Brains

19

 

nervous sy

ganglia the
central l'lei
laterally tr

lterature

3.1 General Description

Cephalopod brains are the largest and show the most complexity among
all invertebrates. They are characterized by large numbers of elaborate sense
organs, motor neurons and effectors (Young, 1988). The cephalopod central
nervous system is significantly larger and shows significantly more fusion of
ganglia than is observed in other molluscs (Young, 1988). In all coleoids, the
central nervous system is situated around the esophagus, and optic lobes extend
laterally from either side (Figure 1). There are several extensive reviews of
literature related to cephalopod nervous systems (e.g., Bullock, 1965; Boyle,
1986)

The central nervous system of coleoid cephalopods (including Sepia) is
highly centralized between its two eyes and is enclosed in a cartilaginous
cranium. Cephalopod brains consist of many fused ganglia, and are organized
into numerous lobes connected by complex sets of tracts and commissures
(Young, 1971). The esophagus runs through the brain dividing it into a
supraoesophageal mass, a suboesophageal mass and the perioesophageal
magnocellular lobes. The buccal lobes are anterior to the rest of the brain and
are connected to the inferior frontal lobe by large buccal ganglia. Short optic
tracts found on both sides of the supraoesophageal mass connect the optic lobes
to the central nervous system. Figure 1 shows the anatomical relationships
among the major components of the coleoid brain. Each lobe of coleoid brains

and their peripheral ganglia consist of a central mass of neuropil containing many

20

   
  

axons an.
and glial :
Th
motor oerj
Boycott.
motor CE
input he
and me
lnsteac
mebn
BOYCc

3.2 D.

Supr:
AT Th
l‘ronl

SUDI

TUnc
def:
labe
rem

[Obs

axons and collaterals, but few cell bodies. Perikaryal layers with both neuronal
and glial cell bodies surround the neuropil of the lobes (Boycott 1961).

The suboesophageal mass contains a number of intermediate and lower
motor centers, and is thus plays a role similar to the vertebrate spinal cord
(Boycott, 1961 ). In contrast, the supraoesophageal mass contains the higher
motor centers, sensory centers and Ieaming centers and is largely controlled by
input from the optic lobes (Boycott, 1961). Because of its involvement in Ieaming
and memory, the description here will focus on the supraoesophageal mass
instead of the suboesophageal mass. Anatomical details related specifically to
the brain of Sepia oflicinalis have been described in great detail (Cajal, 1917;
Boycott, 1961), but will be briefly outlined in the next section.

3.2 Description of the Supraoesophageal Mass

Figure 2 shows a photograph of a mid-saggittal section of the
supraoesophageal lobes of S. oflicinalis with the major recognized lobes labeled.
At the top of the mass, there are three distinct lobes, the vertical lobe, inferior
frontal lobe and superior frontal lobe (Fig. 2). These higher lobes of the
supraoesophageal mass have been associated with motor control as well as
Ieaming and memory of visual, chemical and tactile stimuli (Young, 1971). The
function of this set of lobes in Ieaming and memory will be described in greater
detail in section 3.3. Directly underneath the vertical lobe lies the subvertical
lobe and under the subvertical lobe lies the precommisural lobe (Fig 2). The
remaining portion of the supraoesophageal mass consists of the anterior basal

lobe (divided into anterior and posterior portions), the dorsal and medial basal

21

 

lobes. whi:

 

lateral bas.

basal lobe

mflllected
i“button
brachio.c
All of the
The ante
Suggest
1979).
the Ver.

(YOUTK

9309i
Qangg
is Tug
snag,
105%

f
'SS‘C

lobes, which form the posterior wall of the supraoesophageal mass, and the
lateral basal lobes, which flank either side of the medial basal lobe (Fig. 2). The
basal lobe system appears to control motor function including the
chromatophores, swimming, fin movements, etc (Boycott, 1961).

This highest portion of the supraoesophageal mass (often referred to as
the superior frontal-vertical lobe system) is particularly well developed in Sepia
(Boycott, 1961; Maddock and Young, 1987). The lobes of this system are
connected to one another by many afferents and efferents. Afferent sources of
input to the superior frontal lobe include the optic lobes, inferior optic lobe,
brachio—cerebral tract, vertical lobe and subvertical lobe (Young, 1979 and 1988).
All of these afferents enter through the lateral portion of the superior frontal lobe.
The anterior and posterior portions of the superior frontal lobe are distinct
suggesting that processing within the superior frontal lobe is divided (Young,
1979). Efferents of the superior frontal lobe from the posterior portion connect to
the vertical lobe, while the anterior efferents connect to the subvertical lobes
(Young, 1971 and 1979; Agin et al., 2001).

Similarly, the inferior frontal lobe receives afferents from the buccal
ganglia, mantle and arms and sends efferents to the superior frontal lobe, buccal
ganglia and brachial lobes (Young, 1979). In oct0puses, the inferior frontal lobe
is highly specialized and plays an important role in chemical/tactile Ieaming and
memory (Young, 1971). In cuttlefish, the complexity and function of the inferior
lobe is significantly reduced, though it continues to play an important role in the

feeding system (T able 1; see also Young, 1979).

22

The

 

and Youn-
mrxture c‘
tracts (Y o
neuropil ,-
the neuro I

lobe itself

 

input from
(Young, 1
neurOpil ;
frOlltal lo
(1988). r
leceiving
re9Ulatjr

T
basal) a
basal lc
senSOry

eyes ar

The vertical lobe is well developed in the decapods (Wu, 1954; Maddock
and Young, 1987). There are three distinct classes of cells in the vertical lobe. A
mixture of small amacrine cells and "large cells" in the vertical to subvertical lobe
tracts (Young, 1979). Medium and large neurons are scattered throughout the
neuropil (Young, 1979). Additionally, there are smaller cells in the periphery of
the neuropil, most of which are amacrine cells, projecting only within the vertical
lobe itself (Young, 1979). The vertical lobe receives a majority of its afferent
input from the superior frontal lobe, the subvertical lobe and the optic tract
(Young, 1979). A small number of the larger cells sit in the central portion of the
neuropil and carry information (efferents) back to the subvertical lobe, superior
frontal lobe and the inferior frontal lobe (Young, 1979). According to Young
(1988), the frontal/vertical lobe Circuit is largely involved in prey selection,
receiving input from the tentacles and the optic and buccal lobes, and then
regulating attack behavior.

There are several basal lobes (dorsal basal, median basal and anterior
basal) and these are located toward the posterior of the brain (Figure 2). The
basal lobe system is thought analogous to the vertebrate cerebellum, receiving
sensory input from the eyes and statocysts and regulating motor behavior of the
eyes and the whole animal (Boycott, 1961; Young, 1977). These lobes have no
direct connection with the vertical superior-frontal lobe system (Young, 1977).
The basal lobes are the primary processing point for afferent input from the arms
and from the buccal mass, but also receive input from the optic tract (Young,

1977). The posterior portions of the dorsal and medial basal lobes connect to the

23

 

 

consisting

a medulla
1994). Tl
radial fibe
(Boycott.
centers E

lobes T81

optic lobes and control eye movements specifically (Boycott and Young, 1956).
However, the other lobes in the basal lobe system primarily send their efferent
output to motor centers in the suboesophageal mass, such as the chromatophore
centers and pedal lobes (Boycott, 1961; Young, 1977).

The optic lobes are the largest lobes of the central nervous system
(Young, 1974). In each optic lobe the neuropil and somata are layered
consisting of a deep retina surrounded by outer and inner granule cell layers, and
a medulla (figure 3; see Boycott, 1961; Young, 1971 and 1974; Budelmann,
1994). The deep retinal layers are composed of distinct regions of tangential and
radial fibers and are analogous to the ganglion cell layer of the vertebrate retina
(Boycott, 1961, Young, 1974). The medulla functions as both a link to the motor
centers and a memory storage site (Boycott, 1961; Young, 1974). The optic
lobes receive direct afferent input from the eyes as well as from most of the lobes
of the supraoesophageal mass and submit efferent output to the dorsal and
medial basal lobes as wells as the subvertical and superior frontal lobes (Boycott
1961).

3.3 Role of the Supraoesophageal Lobes in Learning and Memory

As mentioned in Chapter 2, behavioral experiments with cephalopods
have indicated that the vertical lobe and associated regions play an important
part in the acquisition and retention phases of Ieaming (Boycott and Young,
1957, Wells and Wells, 1957, Young 1961, Agin et al., 2001). The vertical and
subvertical lobes were long known as "silent” areas of the cephalopod brain

because electrical stimulation to these areas yielded no Specific effect (Bert,

24

1867). However, surgical removal of these areas yields significantly impaired
acquisition and retention of visual memories in both octopuses and cuttlefish (i.e.
Sanders and Young, 1940; Boycott and Young, 1950).

Octopuses have two, distinct memory systems thatoverlap to some extent
(Young, 1991). One of these memory systems is for chemotactile Ieaming and
the other is for visual Ieaming. Each system involves specific connections
between cells in several different supraoesophageal lobes (reviewed in Young,
1991). The chemotactile Ieaming and memory system involves the higher lobes
of the supraoesophageal region, including the inferior frontal lobe, subfrontal
lobe, buccal lobe, superior -frontal, vertical and subvertical lobes (Young, 1991).
The chemotactile learning system is either not present or not well developed in
cuttlefish, which Show significantly reduced inferior frontal lobes as compared
with octopuses and rely on chemotactile input for prey handling but not hunting
(Young, 1971). Precise details of the chemotactile learning circuit in octopuses
are reviewed elsewhere (Young, 1991).

In contrast, the visual Ieaming system of cuttlefish is markedly similar to
that of octopuses (Young, 1991). Unfortunately, the complexity of the
connections of cells within the optic lobes has made the description of the
precise cellular connections and processes underlying visual memory in
cephalopods particularly difficult to ascertain (Young, 1974 and 1991). However,
a basic description of visual processing has been described. From what is
understood, retinal cells send input from the eyes to the cells within the tangential

layers of the plexiform region of the optic lobe, which may play a role as feature

25

detectors

 

are then tr‘

the optic It

 

hypothesn
stimuli wr‘.
1954).
T'r
the tent:
and ad):
circuit is
hegahv
pOttion
lobe c1
9UStat
l‘IEms

Will fig

Otto;

Shoc

detectors or "classifying cells"(Young, 1971 and 1991). These feature detectors
are then thought to send signals to the "memory cells" within the "cell islands" of
the optic lobe medulla (Young, 1965). Lending support to this particular
hypothesis is evidence from octopuses, which will team to habituate to visual
stimuli with only their suboesophageal mass and optic lobes intact (Boycott,

1 954).

The medullary "memory cells" then connect to one of two paired circuits in
the central nervous system that involve the superior frontal lobe, vertical lobe,
and adjacent subvertical lobe (Young, 1965 and 1991; Fig. 2). Each lobe in each
circuit is thought to exert opposing influences on the probability of positive or
negative responses to the visual stimuli (Young, 1991). Specifically, in the
portion of the circuit comprised of the lateral superior frontal lobe and subvertical
lobe circuit, visual signals from the optic lobes are thought to interact with
gustatory information from the buccal system, thus promoting attack of prey
items (Young, 1991). After removal of the lateral superior frontal lobe octopuses
will no longer attack prey items (Boycott and Young, 1955).

In contrast, in the vertical lobe and median superior frontal lobe portion of
the circuit, visual information is thought to interact with input from nociceptors,
and this circuit mediates learned inhibition of attack behaviors (Young, 1991).
Thus, when the vertical-median superior frontal lobe circuit is interrupted,
octopuses will continue to attack prey items even when receiving electrical

shocks, unless the shocks are given at intervals of five minutes or less (Boycott

26

 

influences
superior ii

particular

 

behveen {I
additional
one prop:

8
vertical Ic
Vertical lr
Vlsual m.-
ion-glen
AS deSC
chem0t;
duh-he t
in VlSua
in lite a
Suppor
0i filOpc
COmDOL

30m 6 0'

and Young, 1955). Additionally, visual Ieaming is significantly reduced when
more than half of the vertical lobe is removed (Boycott and Young, 1957).

It is clear that the two parallel circuits in the system above exert opposite
influences on predatory behavior (Young, 1991). The vertical, subvertical and
superior frontal regions are not connected directly with any single receptor or any
particular motor center (Boycott, 1961). The exclusivity of the connections
between these areas and the fact that they are electrically unexcitable provide
additional support for the existence of a positive feedback system, such as the
one proposed above (Young, 1938; Boycott, 1961).

Boycott and Young's (1955) study also indicated the involvement of the
vertical lobe in the transfer from short-term to long-term memory. However,
vertical lobe removal does not entirely block the retention of previously learned
visual information (Young, 1958). This is because much visual information in
long-term memory is stored within the medulla of the optic lobes (Young, 1965).
As described above, in octopuses, the vertical lobe plays an additional role in the
chemotactile memory system (Young, 1991). The effects of vertical lobe removal
during tactile discrimination Ieaming are markedly less than the effects observed
in visual discrimination tasks. However, tactile discrimination Ieaming is slower
in the absence of the vertical lobe (Wells and Wells, 1957; Wells 1965).
Supporting these findings is a study by Robertson (1994), that examined the role
of filopodial extension in octopus brains (Robertson, 1994). Cytochalasin D (a
compound that breaks down filopodia) was applied to the subfrontal lobes of

some octopuses and the vertical lobes of others. The application of cytochalasin

27

   
  
 
 

Dto the s.
applicatior
support to
involved i

Th:
its neural I

and super

 

 

 

Spatial re

hippocam
not involv
phase a n
Sanders
slia'tial le

“908533

D to the subfrontal lobes specifically caused blocking of touch Ieaming, whereas
application to the vertical lobes did not (Robertson, 1994). This study lends
support to the hypotheses that filopodial extension in the subfrontal lobes is
involved in tactile Ieaming.

Though octopus spatial learning behavior has been relatively well studied,
its neural basis has not yet been studied. It has been suggested that the vertical
and superior-frontal lobes may also play an important role in constructing internal
spatial representations (Sanders 1975). Sanders further suggests that like the
hippocampus of vertebrates, during spatial Ieaming the vertical lobe is probably
not involved in memory storage, but is involved in the memory consolidation
phase and transfer of information from short to long term memory (see review in
Sanders, 1975). To determine the anatomical and cellular networks underlying
spatial Ieaming in cephalopods, further neurobiological research remains
necessary.

Currently, there are significant gaps in our knowledge related to the
neurobiology of learning in cephalopods (Hanlon and Messenger, 1996). Extant
studies have focused primarily on electrophysiological recordings and brain
ablations in octopuses. Such results are often difficult to interpret in conjunction
with complex behaviors like Ieaming because it is often difficult to pinpoint
exactly what structure has been interfered with and whether the interference was
isolated to this structure alone. Developmental studies presented in the next

section further suggest that the circuitry underlying visual Ieaming in Sepia is

28

similar to ‘
vertical, Sl
3.4 Devel

Du

size relati

 

supraoesr
comprises
at the turn
tern. e
b€havior
vertical, :

Connect

similar to thatproposed in octopuses; visual and buccal input interact at the
vertical, sub-vertical, superior-frontal circuit in a feedback system (Young, 1991).
3.4 Development of Behavior in Sepia

During normal development, the cuttlefish vertical lobe nearly doubles in
size relative to the superior frontal lobe as well as the rest of the
supraoesophageal mass (Dickel et al., 1997). The vertical lobe of Sepia
comprises about seven percent of the entire brain volume in hatchling cuttlefish,
at the juvenile stage this increases to about twelve percent (Wu, 1959; F rosch,
1971). Behavioral experiments indicate that the development of predatory
behavior in young cuttlefish can be directly correlated with development of the
vertical, sub-vertical and superior frontal lobes as well as the development of the
connections among these lobes (Dickel et al., 1997).

Hatchling cuttlefish show some prey—capturing behaviors that are similar to
those observed in adult cuttlefish (Wells, 1958; Boletsky, 1987; Dickel et al.,
1997). Thus, a single Ieaming paradigm has been used to examine the
relationship between prey capture and Ieaming at different developmental stages
(Sanders and Young, 1940; Messenger 1971 and 1973; Dickel et al., 2001).
Specifically, cuttlefish are presented a piece of shrimp enclosed in a glass tube.
Striking the tube is painful, and the cuttlefish must team to not strike the visual
stimulus of a prey item (associative Ieaming). A hatchling cuttlefish that lacks a
well—developed vertical lobe will continue to attack the inaccessible prey, even
after many repeated trials (Wells, 1962). Adults with well-developed vertical

lobes will learn to associate attacking the shrimp with pain and cease to attack

29

the tube-e
to fatigue :
F ufi
lobe syste
vanous be
takes abo
curves of .
1973 and
Correlated
20001
Re
hunting bé
agel- Adi

 

reSFIOIISe

Young, 1 E

the tube-encased prey (Wells, 1962). The reduction in attack behavior is not due
‘0 htigue or temporary inability to strike with tentacles (Messenger, 1973).

Further evidence suggesting the importance of the vertical-superior frontal
|0be system in Ieaming stems from the timing of development of these lobes and
various behaviors during development. The vertical lobe system of S. officinalis
takes about four months to develop, and it is not until after this time that Ieaming
curves of juveniles appear comparable to those of healthy adults (Messenger
1973 and 1977). Furthermore, environmental enrichment appears positively
correlated with the development of Ieaming and memory abilities (Dickel et al.,
2000)

Related developmental evidence arises from comparisons of specific
hunting behaviors in adult and hatchlings (or juveniles less than four months of
age). Adult cuttlefish hunt prey items that move out of their visual field.
However, much like Ieamed inhibition of predatory behavior, the hunting
response is absent in cuttlefish younger than four months and is also absent in
adult cuttlefish thaat have been subjected to vertical lobe lesions (Sanders and
Young, 1940). These surgically altered cuttlefish will cease to follow prey items
once they exit the visual field (Sanders and Young, 1940). Senesoent cuttlefish
with degenerating vertical lobe systems also show a marked decrease in hunting
behaviors (Chichery and Chichery, 1992). Neurohistological analysis suggests
that age-related degeneration of the neuropil in the subvertical, precommissural,
superior frontal and inferior frontal lobes is markedly higher than degeneration

observed within the vertical lobe (Chichery and Chichery, 1992). Thus, it is the

30

 

 

degradatlc
suppresse
3.5 Propo.
Ov
are initial 5
vertical lot

the repres

 

(Boycott a
keep track
associatic
Young (15
associate
Visual me
the Vertic
long-t8m.
from seve
may play
motor fUr
levieWSG
Nu
VemCa, lo
tranSl‘er (

degradation of the entire memory system, not the vertical lobe itself that
suppresses normal hunting behavior (Chichery and Chicery, 1992).
3.5 Proposed Functions of the Vertical Lobe

Overall, it appears that the optic and the superior and inferior frontal lobes
are initial sites of association between stimuli. However, the exact role of the
vertical lobe remains unclear. The vertical lobe may ensure the persistence of
the representation of associations made in other supraoesophageal lobes
(Boycott and Young, 1955). Another proposed function of the vertical lobe is to
keep track of active optic lobe classifying cells until a positive or negative
association is made between the stimulus and the outcome (Young, 1965).
Young (1965) further proposes that after processing within the vertical lobe and
associated circuits, signals are sent back to the optic lobes, where long-term
visual memories are stored within the medulla. It has also been suggested that
the vertical lobe plays a specific role in consolidating short-tenn memories into
long-term memories (Sanders, 1975). This hypothesis is supported by findings
from several ablation studies indicating that the vertical and superior frontal lobes
may play an important role in storing memories, and not merely in controlling
motor functions during both visual and tactile Ieaming (reviewed in Young, 1964;
reviewed in Sanders, 1975).

Numerous transverse fibers from the two optic lobes that cross at the
vertical lobe, suggesting that the vertical lobe is a primary site of lnterocular
transfer (Sanders, 1975). Visual discrimination training in one part of the visual

field of only one eye is transferred to the optic lobe of the other eye via the

31

vertical lob
pathway) d
the vertlca
corpus cali
3.6 Comp.
One
suoerlor-fr
diherent Q
chemorec
thDOCam
rEileives j
memory f.
Stipenm f
infQrT'l'iatic
CampgeX
rlearly as
further in

vertical lobe, as inhibiting the optic commissures (and not the vertical lobe
pathway) does not entirely block interocular transfer (Muntz, 1961). In this way,
the vertical and superior frontal lobes may be similar in function to the vertebrate
corpus callosum (Muntz, 1961).
3.6 Comparisons with the Vertebrate Hippocampus

One salient feature of the organization of the cephalopod brain is that the
superior-frontal and vertical lobes allow mixing of different input from various
different types of receptors such as the eyes, arms, mantle, nociceptors, and
chemoreceptors. This organization is quite similar to the vertebrate
hippocampus, often considered a vertebrate learning/memory center, which
receives input from various association cortices and integrates the input during
memory formation (Kandel et al., 2000). In cephalopods, the vertical and
superior frontal lobes not only integrate various stimuli, but also enable of the
information transfer, for example from the right to the left optic lobe. The
complex connections among the various lobes of the cephalopod brain are not
nearly as well understood as connections in vertebrates, but it is possible that
further investigation of cephalopod brains may illuminate general principles

underlying the organization of complex brains.

32

Chapter 4

Calexcitin

33

4.1 Asso

A ‘
understal
instance,
condition
uncondit.
Hermissr
excites n
Contrast,
(reviewe
vestibu la
Mileage
1.974; Cl
phOlOfec
r98l8i3m

to leamll

4.1 Associative Learning in Hermissenda crassicomis

A variety of molluscan systems have been used as simple models to
understand brain function and associative Ieaming at the molecular level. For
instance, the marine snail Hermissenda crassicomis can be classically
conditioned to associate light (the conditioned stimulus) with turbulence (the
unconditioned stimulus (Alkon, 1974; Crow and Alkon, 1978). When the body of
Hermissenda is rotated, the excitation of sensory cells in the vestibular organs
excites motor neurons in the foot, causing reflexive contraction of the foot. In
contrast, stimulation of the photoreceptors promotes positive phototactic behavior
(reviewed in Alkon, et al., 1998). Repeated paired presentation of visual and
vestibular stimuli yield a significant reduction in phototactic behavior and an
increase in the number of foot contractions in response to light alone (Alkon,
1974; Crow and Alkon, 1978). Intracellular recordings from type B
photoreceptors in the eye of conditioned Hermissenda show increases in input
resistance and a reduction of voltage-dependent potassium currents in response
to Ieaming. Both of these effects are dependent on elevated levels of
intracellular calcium (Alkon et al., 1998). Thus, associative learning in
Hermissenda involves a biochemical pathway that includes the long-term
molecular regulation of potassium channels by levels of intracellular calcium
(reviewed in Alkon et al., 1998; Matzel et al., 1998).
4.2 Protein Kinase C and Associative Learning in Hermissenda

The post-association modification of potassium currents after Ieaming can

persist for weeks and appears to result from an increase in intracellular calcium

34

lilC'JCl

cell q

-4
< L)
tn

(Alta
lncre
on th

nierr

Cat
Site
The

ad

inducing the movement of the signaling enzyme protein kinase C (PKC) from the
cell cytoplasm to the cell membrane (Alkon and Nelson, 1990; Nelson and Alkon,
1995). At the cell membrane PKC reacts with a signaling protein called calexcitin
(Alkon, 1989; Alkon et al 1982, Alkon et al., 1985). In the cytoplasm, PKC acts to
increase potassium ion flow thereby decreasing membrane excitability, whereas
on the cell membrane PKC reduces potassium ion flow, which, in turn, increases
membrane excitability (Nelson et al., 1996; Ascoli et al., 1997).

Protein kinase C has also been identified in vertebrate brain tissue as an
important protein regulating memory storage processes (Bank at al., 1998).
Imaging studies have demonstrated membrane-specific changes in PKC binding
after classical conditioning of the eyeblink response in rabbits and olfactory
discrimination and water maze training of the rat (Bank at al., 1988; Olds et al.,
1989; Olds et al., 1990). After classical conditioning of the rabbit there is an
increase in membrane-associated PKC near hippocampal CA1 cell bodies one
day after training (Olds et al., 1989). After three days, PKC migrates from the
CA1 cell bodies to the CA1 cell dendrites. This translocation of PKC in CA1 cells
can be artificially induced by phorbol ester, which also causes the same
potassium ion flow reduction that takes place during associative conditioning
(Alkon, 1989). Translocation of PKC to the CA1 membrane by phorbol ester also
causes enhanced summation of excitatory post-synaptic potentials (epsps)
elicited by the activation of Schaeffer collaterals (reviewed in Golski et al., 1995).
This same enhanced epsp summation has been observed in rabbits trained with

a classical conditioning procedure (reviewed in Alkon, 1999). Thus, conservation

35

of a molecular sequence for memory storage is implied by parallel cellular events
in molluscan and mammalian memory paradigms (Alkon, 1999).
4.3 The Role of Calexcitin

Calexcitin (also called CE or cp20) is a Iow-molecular-weight (22 kDa),
low-abundance G-protein that is stimulated by elevated Calcium levels and the
presence of PKC at the cell membrane during associative Ieaming (Nelson et al.,
1996). Calexcitin is a high affinity substrate for PKC and has been found in the
central nervous system of a variety of species including snails, squid, rabbits,
and rats, and in skin fibroblasts of humans (Nelson and Alkon, 1991; Nelson et
al., 1994; Kim et al., 1995). In Hermissenda, classical conditioning produced by
associating visual and vestibular stimuli yields a 2-3 fold phosphorylation-state
increase of calexcitin by PKC at the cell membrane (Cavallaro et al., 1997).
Thus far, it is unclear whether calexcitin plays a role exclusively in Ieaming and
memory (T .J. Nelson, pers. comm., 2001).

The phosphorylation-state increase is followed by calexcitin translocation
from the cytosol to three principal types of neuronal membranes: the
endoplasmic reticulum membrane, the outer wall membrane and the nuclear
membrane (Nelson et al., 1996; Ascoli et al., 1997). At the outer wall membrane,
calexcitin inhibits voltage-gated potassium channels and causes increased
membrane excitability to further depolarizing stimuli (Nelson and Alkon, 1995).
At the nuclear membrane, calexcitin increases the rate of turnover of messenger
RNA of several specific proteins (Nelson et al., 1996). Calexcitin elicits Ca2+

release from ryanodine receptors on the membrane of the endoplasmic

36

reticulum, resulting in an overall increased amplification of Ca2+ signaling (Alkon
et al. , 1998; Nelson et al., 2001). Therefore, calexcitin appears to serve as a

signaling molecule that amplifies calcium elevation in response to Ieaming-
associated synaptic transmitters and initiates second messengers. For a
summary of the PKG/calexcitin model for associative Ieaming see Figure 4.

Microinjection of CE into Hermissenda neurons causes many of the
previously described cellular effects of Ieaming, such as increased membrane
excitability, reduction of cellular potassium currents, long-lasting depolarization,
modified dendrite arborization, and changes in vesicle transport within the axons
(Moshiah et al., 1993; Nelson et al., 1996; Alkon et al, 1998). Though a number
of other proteins have been implicated in Ieaming, CE is the only protein that
reproduces these electrophysiological and morphological effects when it is
directly injected into neurons (Ascoli et al., 1997). It has also been discovered
that calexcitin undergoes a conformational change in the presence of Ca2+,
specifically, a significant increase in presence of alpha helices and a decrease in
beta sheets (Ascoli et al., 1997). This same study further suggests that
conformational equilibrium of calexcitin may serve as a neuronal trigger for short
-term changes in activity after associative Ieaming.

Additionally, it has been discovered that CE binds to both the Ca2+ and
GTP signaling pathways, and these two pathways are known to interact (Jeng et
al., 1987; Sahyoun et al., 1991; De Matteis et al., 1993). Therefore, calexcitin

may act to combine two separate signaling pathways in response to temporal

37

association of separate inputs during the acquisition of associative memories
(Alkon and Rasmussen, 1988; Alkon, 1995).

As mentioned before, calexcitin has been found in the central nervous
system of vertebrates (rabbits and rats) and may play a role in vertebrate,
hippocampus—dependent memory processing (Sun et al, 1999). In human
Alzheimer's patients, calexcitin immunoreactivity and potassium currents in the
CNS are significantly less than in normal human subjects (Kim et al., 1995).
These data further imply the role of calexcitin in associative memory. While
calexcitin is found in both vertebrates and invertebrates, this does not indicate
that it functions identically in both groups as conservation of mechanisms or
structures involved in carrying out these mechanisms does not always imply
identity of function of the mechanism or structure (Alkon, 1995). To understand
the evolution of mechanisms involved in complex Ieaming, it is necessary to first
compare mechanisms observed in simple nervous systems to mechanisms
involved in complex nervous systems.

4.4 Comparisons with Other Models

One potential problem with generalizing the PKG/calexcitin model of
associative Ieaming to other organisms stems from research using a different
model based on classical conditioning of the gill withdrawal reflex of Aplysia
(described in Kandel, et al., 2000). The marine mollusk Aplysia can be
classically conditioned to a weak tactile stimulus to the siphon (the conditioned
stimulus) with an electric shock to the tail (the unconditioned stimulus). Paired

presentation of these stimulus results in rapid and sustained withdrawal of the

38

siphon and gill to the conditioned stimulus (Carew et al., 1981). Like the
Hemissenda model, the cellular cascade that underlies the conditioned and
unconditioned stimuli has been identified in Aplysia, and this model yields some
striking similarities to that observed in the vertebrate hippocampus (Matzel et al.,
1 998).

In Aplysia, presentation of the unconditioned stimulus immediately prior to
the conditioned stimulus induces increases in Ca2+ in the presynaptic sensory
neuron. This increase activates calmodulin, which then binds to adenylyl
cyclase, and enhances post-synaptic cAMP activity (Kandel et al., 2000). This
mechanism is similar to that found in the vertebrate hippocampal mossy fiber
which uses glutamate as a transmitter. In the mossy fiber, glutamate binds to
both NMDA and non-NMDA receptors on the CA3 pyramidal cells. It is possible
that long-term potentiation (LTP) depends on the influx of Ca2+ into the
presynaptic cell, and not activation of presynaptic NMDA receptors, or the
release of Caz" into post-synaptic cells. When Ca2+ levels increase in the
presynaptic cell, adenylyl cyclase is activated. This influx of calcium thus causes
an increase in cAMP activity which in turn activates protein kinase A (PKA)
(reviewed in Kandel et al., 2000).

Despite their differences, the hippocampal LTP model, the Aplysia model
and the Hermissenda model seem to share several interesting features (Matzel
et al., 1998). In each system, intracellular increases of calcium and activation of
G-proteins act to modify specific membrane proteins, change membrane

excitability and reduce the activation threshold, allowing for modification of action

39

potentials. When comparing the various Ieaming models in this way, the major
differences between the models are the class of protein kinase that is activated
and the site of modulation relative to the targets of action (Matzel et al., 1998).
The specific differences may reflect different solutions to the same types of
problems in different species (Matzel et al., 1998).

4.5 Calexcitin in Cephalopods

Calexcitin was previously discovered in and isolated from the optic lobe of
squid (Nelson et al., 1994; Nelson et al., 1996). Immunohistochemical labeling
demonstrated that CE is found primarily within the plexiform layer of the optic
lobe. This region is the initial location of the conjunction of fibers from the retina
and optic nerve and also contains a number of amacrine cells (Young, 1974). In
the optic lobes, calexcitin is largely localized in axon terminals, consistent with
the hypothesized synaptic function of CE affecting axonal transport and neuronal
branching (Alkon et al., 1990).

Pilot immunohistochemical and western blotting studies (Karson, 2001,
unpublished data) indicated that calexcitin is also present in cuttlefish optic lobes
and some regions of the supraoesophageal mass including the vertical and
subvertical lobes (Karson, 2001, unpublished data). It is particularly interesting
that these regions are thought to be related to associative learning in
cephalopods (see previous chapter). Thus, the relationship between calexcitin
expression and associative Ieaming in cephalopods may be of significant

interest.

Chapter 5

Objectives and Hypotheses

41

The experiments presented in the following three chapters were designed
to gain insight into the possible functions of and mechanisms underlying visual
and spatial simultaneous discrimination Ieaming in cuttlefish. For each objective,
there are a number of associated hypotheses. The specific objectives and

hypotheses that were initially generated are as follows:

1. To compare the relative rate of simultaneous discrimination Ieaming
based on visual versus spatial cues. Using a two—choice discrimination
paradigm, cuttlefish will be trained to escape a maze based on either exit
direction (spatial cue) or provided landmarks (visual cue). The following
specific hypotheses will be tested:

a. Visual and spatial cues are both used in simultaneous discrimination
Ieaming. The ability to make visual discriminations has been well
documented in cuttlefish, but recent findings suggest they are also able to
solve spatial problems. This suggests that visual and spatial cues are both
important in problem solving in nature.

b. The rate of Ieaming depends on the type of cue used. A preliminary
experiment suggests that cuttlefish use visual landmarks rather than
directional cues to find the open exit of a T-maze. This suggests that the
learning rate in visual-cue training may be faster than the Ieaming rate

observed in spatial-cue training.

42

2. To trace the expression of calexcitin (a learning-related protein) in
particular brain regions during Ieaming.

a. Calexcitin is located in Ieaming related brain regions in cuttlefish.
Specifically, I predict that trained cuttlefish will yield a greater distribution
of calexcitin-stained cell axons in the vertical lobe and optic lobes than
motor control animals in the same brain regions.

3. To correlate calexcitin expression with Ieaming events.

a. Calexcitin is involved in discrimination Ieaming. Specifically, previous
experiments suggest that trained cuttlefish will yield greater protein
expression in the vertical lobe and optic lobes than control animals in the
same brain regions because these regions are involved in associative
Ieaming.

4. To compare the expression of a Ieaming-related protein among various
brain regions of visual cue-trained, spatial cue-trained and control
cuttlefish. Comparisons between groups address the following specific
hypotheses:

a. Faster rates of Ieaming are associated with greater increases of calexcitin.

b. Cuttlefish trained to exit based on spatial cues will yield greater increases
in protein in the corresponding optic lobe. For example, cuttlefish trained
to exit the right door may show greater protein activity in the right optic
lobe than the left optic lobe.

c. Simultaneous discrimination Ieaming is associated with decreases of

protein in the cytosolic fractions and increases in protein in microsomal

43

fractions. During associative learning in other molluscs, calexcitin is
translocated within the cells. Western blots of various cellular fractions will
allow for the determination of whether similar patterns of calexcitin
translocation occur during visual or spatial simultaneous discrimination

Ieaming in cuttlefish.

Chapter 6

Behavioral Experimentation

45

6.1 Summary

The objective of the behavioral experiments was to compare the relative
rate of simultaneous discrimination Ieaming based on visual versus spatial cues.
Four experimental groups (visual, spatial, combined, and control) were tested in
a two-choice simultaneous discrimination maze and performance was compared.
These experiments supported the initial hypotheses that cuttlefish can learn to
solve simultaneous discrimination problem using both spatial and visual cues.
The experiments also indicated a significantly decreased number of trials to
criterion in the spatial-cue group compared to other groups, and a significantly
increased error rate and maze escape time in the combined-cue group when
compared to the other experimental groups. Thus, this experiment also
supported the second hypothesis that the Ieaming rate differs between
experimental groups.

6.2 Materials and Method

Experiments were conducted at the Marine Resources Center (MRC) of
the Marine Biological Laboratory in Woods Hole, MA. The tank was 1.8 m wide x
3.04 m long x 0.30 m deep and was divided into quadrants. The total volume of
the tank was about 450 gallons. The water supply was drawn from a depth of
approximately three meters below the surface of Great Harbor and gravity fed to
tanks throughout the MRC building. Salinity ranged from 31 to 33 parts per
thousand and the water temperature ranged from 17° to 21° C.

All subjects were laboratory-cultured juvenile Sepia ofi‘icinalis (mantle

length 6 -12 cm, 14 males, 10 females) ordered from the National Resource

Center for Cephalopods (MBI, Galveston). Cuttlefish were housed in one of the
four tank quadrants, each quadrant holding four to six individuals to minimize the
behavioral effects associated with crowding (Boal et al., 1999). Large pieces of
PVC pipe and artificial plants were placed in each tank to provide artificial shelter
spots. The cuttlefish were fed a mixture of live and frozen fish, squid, shrimp or
crabs twice a day, once at 07:00 and once at 17:00. A complete description of
cuttlefish mariculture can be found elsewhere (e.g. DeRusha et al., 1989; Hanley
et al., 1998).

The maze was designed based on a previous experimental design
(Karson et al., 2003). A circular testing arena was constructed from a large,
plastic barrel (71x 56 cm diameter, Figure 5). A start tube (16 cm diameter) with
doors on both ends was placed through one side of the arena 17 cm below the
top edge (12 cm below water surface). Two exit holes (16 cm diameter) were cut
on opposite sides of the arena, six cm below and perpendicular to the start tube.
Exits were fitted with movable and clear Plexiglas® doors. A clear, Plexiglas®
platform was fitted 12 cm above the bottom of arena, creating a "false bottom."
8. ofl‘icinalis tend to settle on the floor of tanks and the false bottom prohibited
them from reaching the tank floor, thus providing motivation for cuttlefish to
escape the maze. A panel of striped fabric (36x36 cm) and a panel of spotted
fabric (36x36 cm) were created and fitted with Velcro such that they could
surround, and be switched between, either exit (see figure 5).

During a pre-training period, cuttlefish were acclimated to the maze and

tested for exit preference. This was accomplished by placing the cuttleflsh in the

47

maze wit
between
and for e
success
F
groups.
on the n
and visual
soatialct
left. rega
between
Whitefish
panels 5'
control a
were ((3,
Fl
aHOWed .
Seven in
cmefisl
eSCape‘
the DUI};
at leasti

SUCCESS

maze with both doors open and switching the visual—cue panels randomly
between exits. For each individual, the frequency of exit use in each direction
and for each pattern was determined (most frequently used cue during seven
successful escapes).

Following pre-training, cuttlefish were randomly placed into one of three
groups. First, a visual-cue group was designated and was trained to exit based
on the non-preferred panel pattern, spots or stripes, regardless of exit direction
and visual-cue panels swapped randomly between doors (n=8). Second, a
spatial-cue group was trained based on the non-preferred exit direction, right or
left, regardless of panel pattern and visual-cue panels swapped randomly
between doors (n=8). Finally, a control group was trained such that each
cuttlefish was placed in the maze with both doors always open and the visual—cue
panels swapped randomly between the two doors between trials (n=8). The
control animals underwent exactly twenty experimental trials, other cuttlefish
were trained until they attained the criterion for Ieaming.

For all three groups, individual cuttlefish were placed in the start box and
allowed to swim into the testing arena. Once inside the arena, the cuttlefish had
seven minutes to escape the maze. The maze exit time was recorded. If a
cuttlefish did not escape, it was chased out of the open exit with a net and the
escape time was recorded as ten minutes (this value was arbitrary and used for
the purpose of statistical analysis). Each cuttlefish received six trials per day with
at least a 45-minute inter-trial interval. This continued until six of seven

successful, less than one minute, consecutive escapes were achieved. These

Once

to de‘

The c
that
ln=r
West

6.3 R

”amt
6Xpe
trallll
SUIIC

Hem

criteria were selected on the basis of previous experiments (Karson et al., 2003).
once the cuttlefish reached criterion, a testing trial was run with both doors open
to determine whether it continued to use the trained exit. The maze was rotated
semi-randomly in the home tank between trials and experimenter hid behind a
curtain. This insured that the cuttlefish could not rely on cues from around lab to
lmte open exit. If the cuttlefish did not use the appropriate exit during the test
trial, it was given two additional training trials, and then was re-tested for
Ieaming. This happened only one time in one individual which was in the visual
cue group. Statistical tests were performed according to methods suggested by
Zar ( 1 999) and Siegel and Castellan (1988).
Trained animals were sacrificed within five minutes of testing for Ieaming.
The Optic and vertical lobes were removed and preserved in 4%
Paraformaldehyde in 0.1 M sodium phosphate buffer for immunohistochemistry
("=1 3 , 4 animals from each group plus 1 naive animal) or frozen on dry ice for
western blots (n=13, 4 animals from each group plus 1 naive animal).
6.3 Results
The data presented here also include one additional experimental group
nan"'ecl "combined-cue." These cuttlefish were trained in a reversal Ieaming
expel‘iment (Karson et al., 2003). The cuttlefish in the combined cue group were
trained against preference in the same maze, but the striped panel always
s“Wounded the right door and a spotted panel always surrounded the left door.

HenCe, both the visual and spatial cues were relevant to the exit direction

49

('Karso

results

90“? l
cue gr:
prelere
2i prel
preterit
cuttlefis

sex juv

dilldec'
he first
betwee
subjeq
Subject
W635
hats at
blocks

l
limiter]:
declea:

mm 0‘

(Karson et al., 2003; Appendix C, data included here only refer to Reversal 0
results)

The mantle length on the last day of testing, sex, door preference and

group placement of all subjects are reported in Table 2. Including the combined-

cue group, 14 females and 23 males were tested. Of the 37 subjects, initial door

preferences were distributed randomly between left and right (16 preferred left,

21 preferred right) as well as spotted and striped (21 preferred spotted and 16
Preferred striped). In some cases an unequal number of male and female
Cuttlefish were placed in each experimental group, as it is extremely difficult to
sex juvenile cuttlefish using external characteristics.

To directly compare maze performance among animals, trials were
divided into blocks of five, and the mean maze escape time and mean error for
the fi rst, middle and last block of trials were then calculated and compared
between subjects and groups. Because each cuttlefish in the control group was
SUbiemed to only 20 trials total, analysis of maze performance for the control
SUbjeCts included all four blocks of five trials (i.e. all 20 trials). Graphs
representing overall group maze escape times across blocks of experimental
trials are shown in Figures 6 through 9. Graphs representing overall error over
blocks of trials for each experimental group are shown in figures 10 through 13.

Repeated measures analysis of variance (ANOVA) tests indicated that
cuttlefish in the visual, spatial and combined groups yielded a significant
degrease in escape time, and a significant decrease in the percentage of overall

error over the first, middle and last block of trials (results presented in Table 3,

50

Figure 6). This improvement indicates that the cuttlefishes' maze performance
improved with repeated exposure to the simultaneous discrimination problem.
Contrarily, the cuttlefish in the control group showed no significant decrease in
escape time (T able 3, Figure 6A).

The percentage of overall error was sub—divided into the percentage of
error due to not escaping the maze at all (non-escape) and the percentage of
error due to not exiting the maze within 60 seconds. exclusive of non-escapes.
This was done to determine which factor appeared to contribute more strongly
overall errors made, non—escape or non-criterion performance. There was no
significant reduction in either type of error for cuttlefish in the control group (Table
3, Figure 7A). There was a reduction in overall error for cuttlefish in the visual
group (T able 3, Figure 78). However, it appears that neither source of error
independently significantly affected overall error (Table 3). The decrease in the
percentage of errors exclusive of non-escapes more strongly contributed to
overall error in the combined cue group, while a decrease in non-escapes more
strongly contributed to overall error in the spatial group (Table 3).

The overall mean number of trials to criterion was significantly lower in the
spatial cue group than it was in either the visual cue group (one-way ANOVA,
F=3.93, P=0.067) or the combined cue group (one-way ANOVA, F=3.35,
P=0.079) (Figure 8). The overall mean escape time differed significantly
between groups (one-way ANOVA, F=7.93, P<0.0001). A Tukey Test indicated
that this significant difference resulted from differences between all pairs except

for the visual and spatial groups.

51

The mean percentage of overall error was significantly less in control
cuttlefish than the three experimental groups (one-way ANOVA, F=7.60,
P=0.0006) and a Tukey Test indicated that the trained groups did not differ
significantly from each other in mean percentage of overall error.

The overall error was again divided in to two types of error, non-escapes
and non-criterion performance (exclusive of non-escapes). ANOVA and Tukey
tests indicated that the percentage of non- escapes was significantly greater in
the combined group than any other group (one-way ANOVA, F=7.99, p<0.0005).
The remaining mean error (error not including non-escapes) was significantly
different between groups (one-way ANOVA, F=5.575, p<0.005). The Tukey Test
indicated that the difference between combined—cue and the visual-cue group
most strongly contributed to this significant difference.

Previous work with other cuttlefish has indicated that spatial information
may be salient to male cuttlefish than to females (Comer and Moore, 1980).
Thus, males and females were compared for the next analyses. Tables 4 and 5
compare the mean escape time and mean error for male and female cuttlefish
across blocks in each experimental group. The mean escape times appear to
differ between males and females on the first block of trials for the visual, spatial
and combined cue groups (Table 4). However, none of these differences were
statistically significant (Mann-Whitney U test: visual, U (4,4) =13, p>0.10, spatial U
(25) =10, p>0.10, combined, U (5, 7) =22, p>0.10). The differences between overall
mean error for males and females in all groups during the first and second block

of trials were not significant (Table 5; Mann-Whitney U Test: visual-cue group, U

52

(4,4) = 13, p>0.10, spatial-cue group, U (2,6) =10, p>0.10, combined-cue group, U
(5,7, =20, p>0.10, control U (2,5)=10, p>0.10).
6.4 Discussion

Previous experiments have established that cuttlefish can Ieam to solve
simultaneous discrimination problems (eg. Messenger, 1977; Karson et al.,
2003). This behavioral experiment is the first to demonstrate that cuttlefish are
capable of solving both spatial and visual simultaneous discriminations. Though
it is often difficult to tease apart spatial and visual memory, as internal
representations of space may rely on visual cues such as landmarks, the spatial
cues here are independent of visual landmarks (both in the maze and around the
lab). The cuttlefishes' ability to solve each type of problem is approximately
equal, as there is no significant difference in mean escape time and error, or
patterns of improvement over blocks of trials in the visual and spatial groups.
This may indicate that cuttlefish rely on both spatial and visual Ieaming in nature.
While visual information has clear utility to cuttlefish during hunting and prey
selection, perhaps spatial Ieaming has utility in evasion of predators (Karson et
al., 2003).

Control cuttlefish showed no improvement in maze performance. This
indicates that the improvement in performance seen in the cue-trained groups is
not simply due to the cuttlefish learning that there are two exits to the maze, as
the control cuttlefish would have had opportunity to Ieam this as well. Instead, it
appears that by reinforcing the selection of a particular exit, the cuttlefish Ieam

something about the particular cues associated with the exits. Furthermore, a

53

decrease in the percentage of over one-minute escapes more greatly contributed
to overall error than did a decrease in the number of non-escapes. This may
also indicate that cuttlefish did not simply learn that there were two exits, but
actually Ieamed to solve the problem using provided cues. Individuals in the
control group did not show a significant tendency to use one door or visual cue
over the other when exiting, for individuals, responses were split equally between
left and right and between spotted and striped exits.

It is interesting that the visual and spatial cue groups yielded significantly
better learning performance than the combined cue group from previous
experiments. Cuttlefish in the combined cue group were presented with two
relevant cues, door direction and pattern, whereas the spatial and visual groups
were presented a relevant and irrelevant cue with each trial (either direction
relevant, pattern irrelevant or vice versa). In previous reversal Ieaming work with
octopuses, improvement in discrimination performance over serial reversals has
been attributed not only to increased selection of relevant cue, but decreased
attention to the irrelevant cue (for instance cue position or orientation, reviewed
in Mather, 1995, and Boal, 1996). In fact, it appears that octopuses presented
with both irrelevant and relevant cues Ieam discriminations better than those
presented solely with relevant cues (Sutherland, 1959; Sutherland and
Mackintosh, 1971; Messenger and Sanders, 1972). The data presented here are
consistent with this interpretation. Such findings may provide powerful evidence
for concept-formation in cephalopods, as they not only attend to particular cues,

but choose to which cues they will attend, and are thereby sorting out relevant

54

information from irrelevant information (Mather, 1995). Despite these
hypotheses, most cephalopod discrimination experiments have employed
successive, rather than simultaneous, presentation of cues (Boal, 1996). Thus,
such data should be considered more carefully and discrimination experiments
should be redesigned.

The findings here further confirm previous observations that there is great
individual variation in cephalopod Ieaming abilities (Mather, 1995; Hanlon and
Messenger, 1996; Boal, 1996; Boal et al., 2000). Previous authors have
suggested that these differences may be due to different temperaments or prior
experiences of each individual (Mather and Anderson, 1993). It has further been
suggested that the limited behavioral repertoire of paralarval cephalopods as
compared with the large range of behaviors in the later benthic adults is a
function of increased opportunities for Ieaming in complex benthic habitats
(Boletsky, 1992). Different prior learning experiences, then, may account for
differences in behavior (Mather and Anderson, 1993). Alternatively, individual
preferences for particular door directions may be a function of laterally
asymmetrical eye-use. For instance, individual 0. vulgaris appear to have a
"dominant eye" which they use significantly more than the other eye while
examining objects in the frontal-visual field (Byme et al. 2002). The relationship
between individual preferences and the behavior of the individual merit further
investigation. Also consistent with previous work with cephalopods, is the
tendency of cuttlefish to solve the maze problem within a few trials followed by a

period of extreme behavioral fluctuation, in which the cuttlefish may not escape

55

the

isa

lot

888

spa

HOt

tral

the maze at all during several trials (J.G. Boal, pers. comm., 2002). Whether this
is a function of "boredom" or some other aspect of cephalopod behavior remains
to be seen.

Observations that male Sepia latimanus may maintain home areas during
seasonal mating aggregations (Comer and Moore, 1980) could indicate that
spatial information may be salient to male cuttlefish (Karson et al., 2003).
However, the data here do not indicate any performance differences between
trained juvenile male and female cuttlefish in any cue group. While only two
females were present in both the control and spatial cue groups, the visual and
combined cue groups contained more females and yielded no sex-specific
differences in Ieaming behavior. Thus, a larger sample size may not affect this
particular finding. It is possible, however, that differences in maze performance
could become apparent as cuttlefish become sexually mature, as males tend to
become more aggressive and active with age, whereas females do not. Further
studies examining the development of discrimination Ieaming with age could
shed light on this particular question. Further discussion of the interpretation of

the behavioral results is presented in Chapter 9.

56

Chapter 7

lmmunohistochemistry

57

7.1 Summary

The immunohistochemical experiments presented here relate to
Objectives 2 and 3: To trace the expression of calexcitin in particular brain
regions during Ieaming and to correlate expression of a calexcitin in the cuttlefish
brain with Ieaming events. Immunohistochemistry was first used to visualize
regions of wlexcitin expression in the optic lobes and supraoesophageal lobes.
Cell counting software was used to count the number of intensely calexcitin
immunopositive cells within the optic, vertical, subvertical, and dorsal basal lobes
in the visual-cue, spatial-cue and control groups. Calexcitin expression was
increased in the optic lobes of visual-cue trained cuttlefish and expression
differed between the right and left optic lobes.
7.2 Materials and Methods

Immunohistochemical methods were employed to visualize specific
regions of calexcitin immunoreactivity within the supraoesophageal and optic
lobes of Sepia ofiicinalis. Twelve Sepia brains were used for the
immunohistochemical studies, four control animals, four visual cue trained
animals, and four spatial cue trained animals (Chapter 6). After behavioral
experimentation, cuttlefish were rapidly decapitated and the brains (optic lobes
and central nervous system) were quickly removed and fixed in 4%
paraforrnaldehyde in phosphate buffered saline (PBS) solution (pH 7.0).
Following four days of fixation in paraformaldehyde, the surrounding muscle
tissue and cartilage were trimmed, and the brain tissue was immersed in 30%

sucrose solution in 0.1M PBS for cryoprotection. Tissue remained in sucrose

58

solution until it ceased to float (approximately 48 hours). Following
cry0protection in sucrose, the tissue was dipped into a beaker of 2-methylbutane
resting on dry ice for 10-20 seconds, until the tissue was frozen. Each piece of
tissue was then wrapped in foil and stored at -80 °C. Brain tissue pieces were
serially sectioned on a cryostat at 30 um, and slices were stored in a glycerol
solution in collecting wells.

Free-floating slices were rinsed in 0.1 M phosphate buffered saline (PBS;
pH 7.4) three times each for 10 minutes on a rotating platform. Then, slices were
preincubated for 60 min at room temperature in 10% normal goat serum (Sigma,
St. Louis, MO) in PBS with 0.5% Triton-X. Tissue was dip rinsed in PBS and
then incubated overnight in CE primary antibody in 0.1% triton-x solution at 4°C.
After 16—20 hours of primary antibody incubation, slices were washed in PBS four
times for 15 minutes each and then incubated in goat anti-rabbit antibody
(Sigma, St. Louis, MO) in PBS (1:300 dilution) for one hour at room temperature.
Slices were then rinsed in PBS and then placed in Avidin-Biotin Complex (ABC)
solution for 30 minutes (Vectastain® ABC kit, Vector Laboratories, Burlingame,
CA). Slices were washed in PBS twice for 10 minutes each wash and placed in
0.1M Tris buffer (pH 7.4) for 10 minutes.

Anti-calexcitin antibody (25U2) was prepared in rabbits from full-length
calexcitin protein extracted from squid optic lobes (by Dr. T.J. Nelson, Blanchette
Rockefeller Neurosciences Institute, Gaithersberg, MD).

Antibody labeling was visualized using diaminobenzadine (DAB Substrate

Kit for Peroxidase, Vector Laboratories, Burlingame, CA). Slices were then

59

mounted sequentially on silanated slides (Sigma Diagnostics, St. Louis, MO) and
placed on a slide dryer overnight. Sections were dehydrated in an ascending
ethanol series (70%, 90% and 100%), cleared with xylene, mounted with
Permount and cover-slipped. All slides were viewed under a Nikon Eclipse E400
microscope and digital images were acquired using a DXM1200 Nikon digital
camera. Optical density measurements and cell counts of these digital images
were obtained using NIH Image. Further details of cell counting procedure is
presented in sections 7.2 and 7.3.

The negative immunohistochemical control included the substitution of
normal goat serum for the primary antibody. This allowed for visualization and
comparison of background staining to staining resulting from calexcitin
immunoreactivity. All negative controls where normal goat serum had substituted
the primary antibody yielded no staining.

7.3 Overall Patterns of Labeling

The calexcitin antibody was made from squid optic lobe calexcitin injected
in rabbits, and the patterns of calexcitin staining in the squid and cuttlefish optic
lobes are nearly identical (see section 7.5 and Nelson et al.1994 and 1996).
Thus, it is reasonable to assume that staining observed here represents true
calexcitin immunoreactivity. Calexcitin immunoreactivity was then used as a
marker to observe where learning may occur in cuttlefish by correlating the
regions of immunoreactivity with previous findings related to location of the

processing of associative Ieaming and memory.

The optic lobes of S. officinalis consist of the "deep retina" (Cajal, 1917)
and the medulla (Figure 15). The deep retina is layered, and includes a striated
plexiform region composed of four layers of tangential fibers interspersed with
four layers of radial fibers surrounded by distinct inner and outer granule cell
layers (Young, 1974; Figure 15, see also Figure 3). The plexiform region is
thought to be involved in the visual analysis and classification system and the
medulla appears to be a motor center and memory store (Boycott 1961; Young,
1974). The medulla is organized as a series of radial columns that converge
toward the center of the tissue interspersed with horizontal dendrites which form
a series of layers (Cajal, 1917; Young, 1974). However, the divisions between
the layers of the medulla are not as sharp as are the divisions between layers in
the deep retina (Young, 1974).

Calexcitin immunopositive cells are evident, and include various-sized
centrifugal (approximately 10-15um diameter) and amacrine cells (approximately
1-10 um diameter) of the outer granule cell layers of the plexiform zone (Figure
9). There also appears to be staining of amacrine cell axon terminals in the
tangential layers of the plexiform region, particularly tangential layers 2 and 3
(Figure 9). The amacrine cells of the inner granule cell layer are also calexcitin
immunOpositive (Figure 9). A band of large, calexcitin-immunopositive amacrine
cells (10-15um diameter) is located between the inner granule cell layer at the
border of the plexiform region and the medulla (Figure 9).

Within the medulla, staining is seen in various sized cells (5-15pm

diameter) located within "cell islands" (Cajal, 1917) found in the "zone of radial

61

columns" (Young, 1974). The tracts of fibers (both axons and dendrites) that
separate these islands from one another do not show calexcitin immunoreactivity
(Figure 10).

The vertical-superior frontal system includes the superior frontal, vertical
and sub-vertical lobes. These lobes are thought to play an important role in the
consolidation and retrieval of memories (Boycott, 1961; Messenger, 1973,
Young, 1940; Chapter 3). The superior frontal lobe exhibits immunoreactivity
only in few, small cells (2-7 um) toward the posterior region of the lobe (Figure
11). This posterior region of the superior frontal lobe sends axons to the vertical
lobe (Young, 1979). Calexcitin immunoreactivity is far more evident in the
inferior frontal lobe, which shows staining of cells 2-10 m diameter toward the
interior region, which are all thought to send axons to the superior frontal lobe
(Figure 12; Young, 1979). There is also staining of large cells (10-15um) toward
the dorsal edge, bordering the superior frontal lobe, and the posterior edge
(Figure 12). These large cells send axons to the subvertical lobe as well as to
the brachial and buccal lobes found in the suboesophageal mass (Young, 1979).

In the vertical lobe, many of the "large cells" (10-15um, Young, 1979)
surrounding the entire edge of the lobe stain positive for calexcitin (Figure 13).
Calexcitin immunoreactivity is particularly evident at the wall of the vertical lobe
directly ventral to the subvertical lobe (Figure 14). Several small amacrine cells
found throughout the vertical lobe neuropil are also calexcitin immunopositive

(Figure 15). Finally, there is strong staining in the axons and the fibers of large

62

cells (10-15 pm diameter) associated with the vertical-subvertical tracts (Figure
14).

In the subvertical lobe, immunoreactivity is prevalent in medium to large
cells (5—15um) surrounding bundles of fibers in the large anterior region, where
the vertical and subvertical lobes connect (Figure 15). Many of the fibers
themselves also stain positive for calexcitin. Several large cells scattered
irregularly in islands throughout the posterior region of the subvertical lobe also
appear to stain positive for calexcitin (Figure 15). A number of larger cells at the
ventral portion of this same posterior region (subvertical lobe/precommissural
lobe border) also stain positive for calexcitin.

In the anterior basal lobe, a number of medium to large (5—15um)
immunopositive cells are present at the ventral and anterior walls (Figure 16).
There is also staining in the smaller neurons (1-5um) scattered through the
tangled neuropil located in both the anterior and posterior portion of the anterior
basal lobe (Figure 16).

There is staining of small cells (1-5um) scattered in irregular islands
(Young 1977) throughout the dorsal basal lobe (Figure 17). There are also a
number of larger calexcitin immunopositive cells (5-12um), grouped in clusters
within the median basal lobe (Figure 18). There is also distinct calexcitin staining
of medium and large cells (5-15um) at the posterior cell wall which spans both

the dorsal basal and medial basal lobes (Figure 18).

63

7.4

b;

{D

(a.

7.4 Comparison of Calexcitin Expression Between Experimental Groups

To examine the affect of maze training on calexcitin expression, the
number of strongly-immunopositive cells stained at similar levels was compared
in trained and control animals (Chapter 6). The regions compared included the
medulla of each optic lobe, deep retina of each optic lobe, vertical lobe,
subvertical lobe, and the dorsal basal lobes. These regions were chosen
because initial observations indicated they showed calexcitin immunoreactivity
and not prone to damage (small tears), which may have affected interpretation of
the data.

Using NIH-Image, the gray-scale values of cells stained above
background were determined for each slice. The range of values of cells stained
above background was then divided into three categories (strongly-stained cells,
medium and strongly stained cells and all stained cells). The number of cells in
each of these categories was counted in 200um by 200um (by 30 um deep)
regions. Nine of these square regions were randomly selected, each from a
different layer of tissue; thus, these counts thus reflected calexcitin
immunoreactivity throughout the tissue of interest. For the deep retinal regions of
the optic lobes, 200um by 200um squares were selected such that one
encompassed the outer granule cell layer, one the plexiform layers, and one the
inner granule cell layer. To eliminate experimenter bias, there was no indication
to which individual or group a particular slice of tissue belonged during cell

counts.

A KruskaI-Wallis one-way ANOVA was used to compare the mean
number of strongly stained cells present in each of the seven brain regions
among control, visual cue- trained and spatial cue-trained animals. The mean
number of strongly immunopositive cells for each brain region of each
experimental group is presented in Table 6. The KruskalI-Wallis test indicated
that there was no significant difference between calexcitin expression in the
vertical lobe or the posterior region of the dorsal basal lobe between the different
experimental groups (Table 7). Calexcitin expression in the subvertical lobe was
significantly higher in the control group than in the visually trained group (Tables
6 and 7). However, visually trained cuttlefish show a significantly greater mean
number of stained cells in both the medulla and deep retinal region of the right
and left optic lobes than either the spatial or control groups. (Table 7).

Calexcitin expression in the left and right optic lobes was compared for
each group of cuttlefish using a ercoxon signed ranks test (Table 8). Cuttlefish
in the control group yielded greater calexcitin expression in the left plexiform
layer than in the right plexiform layer (Table 8). Spatially trained animals yielded
the same results (Table 8). All other comparisons were statistically insignificant
(Table 8).

Because the mean number of cells expressing calexcitin appeared to
differ between the right and left optic lobes, it was necessary to determine
whether this difference was associated with the initial directional preferences of
the individuals (Table 2). Thus, the mean number of strongly stained cells in the

right and left plexiform layers and medulla was compared for all animals that

65

initially preferred the right door and for all cuttlefish that initially preferred the left
door. For left-preference cuttlefish, the mean number of stained cells did not
differ significantly between the medulla of the right optic lobe and the medulla of
the left optic lobes (erxocon signed ranks test, n=4, C=3, p>0.50). The left
plexiform layer appeared to yield greater staining than the right optic lobe
plexiform in the left preference group with a borderline-significant p-value
(VWlxocon signed ranks test, n=4, C=9, p=0.1250). The right-door preference
cuttlefish showed the same pattern of expression. Calexcitin expression in the
left plexiform layer was higher than in the right plexiform layer (erxocon signed
ranks test, n=8, C=30, p=0.0547) and there was no difference in expression
between the right and left medulla (VVrlxocon signed ranks test, n=4, C=24,
p=0.2305).
7.5 Discussion of Overall Patterns of Staining

The cephalopod visual memory system is highly distributed (and
redundant) throughout the optic lobes and central nervous system, eventually
funneling into a memory storage area (the vertical lobe and optic lobe medulla;
Young, 1991; review presented in Chapter 3). The results presented here
indicate that calexcitin is distributed within areas thought to be associated with
visual Ieaming. Specifically, calexcitin immunoreactivity is prevalent in the
tangential layers of the plexiform zone, the granule cell layers and the medulla
cell islands, as well as in the vertical, subvertical and superior frontal lobes.

The presence of calexcitin within fibers connecting the vertical and

subvertical lobes is interesting in light of studies suggesting that these

con
pie.
llaC
wit

d€l

hlg
QUE
the
till

sug

3U;
as
the

int

connections in particular play an important role in the aspects of cuttlefish
predatory behavior that rely on memory (Dickel et al., 1997). The density of fiber
tracts connecting the subvertical and vertical lobes has been shown to increase
with age, and this connective development closely correlates with the timing of
development of associative Ieaming ability (Dickel et al., 1997). Moreover,
cuttlefish at the same developmental stage raised under different temperature
conditions yield differences in the density of these tracts; warmer water yields a
higher density of tracts (Dickel et al., 1997). Different water temperatures yield
quantifiable behavioral differences between the groups including differences in
the acquisition of prey pursuit behavior as well as rates of pursuit of prey items
(Dickel et al., 1997). Related to this developmental study are data from studies
suggesting that hatchling cuttlefish do not follow prey that leaves the visual field
('hunt" it) because they lack a short-term memory system (Messenger, 1977).
The presence of calexcitin within the fibers connecting the vertical and
subvertical lobes may further suggest the importance of this particular connection
as it relates to the formation of associative memories.

Previous studies have proposed that amacrine cells in the vertical and
subfrontal lobes evolved from cells of the buccal lobes, which function primarily
as simple reflex centers (Young, 1991). The concentration of amacrine cells into
distinct lobes may have allowed for associative Ieaming by allowing for prolonged
inhibition of particular networks (Young, 1991). When the amacrine cells within
the vertical lobe are blocked with drugs, memory is significantly inhibited in

octopuses (Young, 1983). In the vertical lobe, calexcitin staining is seen in a

67

 

no
plC

lea

ino'
sys
for

in

sug

dl&

doe

231

the

W

U

number of small amacrine cells. This finding lends support to the hypothesis that
protein expression of vertical lobe amacrine cells is important in associative
Ieaming.

Calexcitin expression is visible within the inferior frontal lobe. This may
indicate that the inferior frontal lobe plays an important role in the memory
system, but not necessarily the visual Ieaming system. In octopuses, the inferior
frontal lobes are involved in a tactile memory system that functions separately
from the visual memory system (reviewed in Young, 1991). Young (1991)
suggests that because of the degree of simplicity of the inferior frontal lobe
relative to that of octopuses, decapods have only a limited capacity for tactile
discrimination Ieaming. In cuttlefish, the development of the inferior frontal lobe
does not correlate with increases in Ieaming prey-catching abilities (Dickel et al.,
2001)

Differences between octopuses and cuttlefish in regard to the importance
of the inferior frontal lobe may be explained by differences in prey capture
behavior. While octopuses rely heavily on tactile information for prey detection,
decapods detect their prey primarily by sight and use tentacles for capture
(Young, 1988). Thus, in cuttlefish, a tactile memory system would more likely be
used in prey handling rather than hunting (Young, 1988).

Calexcitin immunoreactivity is also present in the clusters of large cells in
the dorsal basal lobe and the cell wall of the dorsal basal and medial basal lobes
but it is unclear how calexcitin expression in these lobes may relate to Ieaming

and memory. The function of the dorsal basal lobe is unknown, but may be

68

in

the

VE

an

Sig

involved in escape and avoidance behavior (Young, 1971). In contrast, it may be
that the dorsal basal lobe regulates the tendency to attack visual stimuli (Parriss
1965). It has been difficult to surmise the precise role of the dorsal basal lobe, as
the behavioral effects of lesions to this lobe have been difficult to observe and
precisely describe (Boycott, 1961). Dorsal basal lobe lesions do not produce
gross motor abnormalities, but do produce slight changes in behavior in
octopuses (Boycottt, 1961). For example, lesions cause a tendency of
laboratory-dwelling octopuses to change their home position more frequently
than normal octopuses (Boycott and Young, 1955). The median basal lobe
function is also difficult to precisely describe, but is thought to be involved in
escape behaviors, as well as swimming and respiration as it controls funnel and
mantle movements (Young, 1977).

Overall, the results here are similar to those found by DiCosmo et al.
(2000), who used immunohistochemistry to examine Nitric Oxide Synthase-like
(NOS) immunoreactivity in the central nervous system of S. otficinalis. NOS has
been suggested as an important molecule underlying associative Ieaming in a
variety of organisms (reviewd in DiCosmo et al., 2000). Both calexcitin and NOS
are signaling molecules that respond to influxes of calcium. Patterns of NOS
expression are remarkably similar to patterns of calexcitin expression seen here
(DiCosmo et al., 2000). Specifically, NOS-like immunoreactivity was observed in
cells projecting from the subvertical to vertical lobes, the vertical lobe neuropil,
and the cell islands of the optic lobe medulla (DiCosmo et al, 2000). Nitric oxide

signaling may play an important role in long- term memory. However, the exact

69

‘I
ii:

in

details of this role are not well understood (Hawkins, 1996). NOS release is
activated when there is a calcium influx due to binding of glutamate to NMDA
receptors (Garthwaite et al., 1989). NOS inhibition has been demonstrated to
block visual and touch Ieaming in octopuses (Robertson et al., 1994 and 1996).
It is possible that NOS pathways and the calexcitin pathways interact, or perhaps
that different types of Ieaming invoke different molecular cascades in Sepia
oflicinalis.

7.6 Discussion of the effects of training on CE histochemistry

My results suggest that intracellular calexcitin expression differs between
trained and untrained cuttlefish as well as between spatially trained and visually
trained cuttlefish. Specifically, cuttlefish trained to make visual discriminations
yield a significantly greater number of intensely calexcitin immunopositive cells in
the medulla and deep retina of the optic lobes than in spatially trained and control
cuttlefish which had to ignore visual "noise" to solve the maze problem. Visual
cue trained cuttlefish showed an increase in intracellular calexcitin relative to the
other two groups.

In studies using the marine snail Hermissenda, experiments similarly
indicated that snails exhibiting memory of positive association yielded a higher
concentration of intracellular calexcitin than snails trained with a sensory block
(Borley et al., 2002; Kuzirian et al., 2003). Kuzirian et al (2003) hypothesized
that Ieaming induced changes should be visible in cells that modulate or control
downstream neurons, not downstream neurons themselves. For instance, in

Hermissenda, Ieaming and memory occurs and is stored in visual cells that

70

interact with a feed-back loop to affect motor neurons. Calexcitin is found in the
visual cells, but not the downstream motor cells (reviewed in Kuzirian et al.,
2003). Similarly, visual Ieaming and memory in cephalopods occurs in the optic
lobes effecting motor neurons via the vertical lobe system. This may explain why
calexcitin expression differs between visually trained cuttlefish and other groups
in the optic lobes, but not in downstream neurons in the vertical lobe complex.
Thus, immunohistochemical results from cuttlefish may further indicate that visual
memory consolidation is modulated by intracellular calexcitin concentrations
within the visual cells (Borley et al., 2002).

Moreover, the data here lend support to the hypothesis that visual
associations are consolidated and stored within the optic lobes, whereas spatial
associations are consolidated and stored in other brain regions. Young (1964)
proposed that while octopus visual memories are stored in the optic lobes, tactile
memories are stored in the posterior buccal or subfrontal region. Previous
studies have not examined sites of storage for spatial information, but it is likely
that like tactile memories, these too are stored somewhere other than the optic
lobes (J.B. Messenger, pers comm, 2003). Perhaps related to this
interpretation, is the fact that calexcitin is found primarily within invertebrate
visual systems rather than other sensory systems (reviewed in Kuzirian et al.,
2003). Calexcitin may be involved exclusively in the visual processing system.
Thus, it is quite likely that Ieaming-induced changes in calexcitin may only occur
during visual Ieaming (Kuzirian et al., 2003). This possibility is described in more

depth in chapter 9.

71

More difficult to explain is the finding that intracellular calexcitin expression
was greater for the control cuttlefish than for the spatial and visual-cue trained
cuttlefish in the subvertical lobe. It is possible, perhaps, that the control group
Ieamed some association during the twenty trials. It is also possible that the time
at which the trained cuttlefish were sacrificed was post-increase or post-release
of calexcitin, and a more suitable time period for testing may exist. To test this,
an obvious experiment would be to measure Ieaming-induced changes in
calexcitin expression at different time intervals following training (five minutes,
ten minutes, one hour, 24 hours, etc). Alternatively, calexcitin could be blocked
at various time periods pre- and post-association and the affect on Ieaming could
be observed. In the next chapter, western blotting techniques are employed to
further quantify differences in calexcitin expression between visual, spatial and
control goups. Further discussion of possible implications stemming from the
immunohistochemical (and western blotting results from Chapter 8) is presented

in Chapter 9.

72

Chapter 8

Calexcitin Expression Quantified Using Western Blots

73

8.1 Summary

Western blotting was used to further examine Objectives 3 and 4: To
correlate expression of a calexcitin in the cuttlefish brain with Ieaming events and
to compare the expression of a Ieaming-related protein among various brain
regions of visual cue-trained, spatial cue-trained and control cuttlefish.
Homogenate, cytosolic and microsomal fractions were prepared from the visual-
cue trained, spatial-cue trained and control cuttlefish. Calexcitin was quantified
in the optic lobes and supraoesophageal mass of each group. Calexcitin
expression was greater in the cytosol and microsomes of optic lobes of visual-
cue trained cuttlefish than other regions in other groups. These analyses also
indicated differences in expression between the left and right optic lobes. These
results were similar to those found in the immunohistochemical experiments
(Chapter 7).
8.2 Methods

To prepare tissue for homogenization, cuttlefish were rapidly decapitated
and the optic and supraoesophageal lobes were quickly removed and
immediately placed in separate microcentrifuge tubes on dry ice. The tubes were
then i transferred to a -80 °C freezer. Before preparation of the fractions, each
piece of frozen tissue was weighed. The weight of each piece of tissue from
each cuttlefish is shown in Table 9.
Fraction Preparation

Homogenate, cytosolic fractions and microsomal fractions for each piece

of tissue were prepared using a modification of Gray and Whittaker’s method

74

(1961; modified by T.J. Nelson). A buffer (referred to here as Buffer A) was
made from 50.0mM Tris HCI (pH7.5), 10.0mM NaCl, 1.0mM EDTA and 0.5mM
EGTA and stored on ice. 30p.l of protease inhibition cocktail (PIC) (Sigma-
Aldrich) was diluted into 1470 ul Buffer A. Each piece of brain tissue was loosely
homogenized (35 strokes) in 1 ml of Buffer A in a chilled Dounce grinder (2 ml
capacity).

50 ul of the homogenate were added to 5 ul beta mercapatoethanol in
45pl of NuPage'“ 4x LDS sample buffer (lnvitrogen Corporation), sonicated for 5
seconds and stored in a freezer at -20°C.

The remaining portion of the homogenate was placed in a 4°C centrifuge
and rotated at 10,000 g for 30 minutes. After centrifugation, the pellet was
discarded and the supernatant was placed in a 4°C ultracentrifuge and rotated at
100,000 g for 1 hour. The resulting supernatant from this spin was the cytosolic
fraction. The cytosolic fraction was added to an equal amount of 10%
betamercapetoethanol in LDS sample buffer, sonicated and stored at -20°C.

To make the microsomal fraction, the pellet leftover from the cytosolic
fraction was resuspended in 200 pl of Buffer A, sonicated for 10 seconds and
placed in a 4°C ultracentrifuge and rotated at 100,000 g for 1 hour. The
supernatant was discarded, the pellet resuspended in 100 pl of Buffer A,
sonicated for 3 seconds and an equal amount of 10% betamercapetoethanol in
LDS sample buffer was added. The microsomal fractions were also stored at -

20°C. Thus, in total, each cuttlefish yielded 9 samples; a homogenate, cytosolic

75

and microsomal fraction from each individual's left optic lobe, right optic lobe and
supraoesophageal lobes.
Protein Content Determination

The Bradford Method (Bradford, 1976) was used to determine the protein
concentration of each sample. First, a standard curve was prepared using
bovine serum albumin (BSA) and was used for calibration of a standard curve
(Figure 19).

5 ul of each brain tissue sample (homogenate, cytosolic and microsomal
fractions) were then diluted in dH20 to 1:300. 1 ml Coomassie reagent was
added to the diluted samples and readings were taken immediately in a
spectrophotometer at 595 nm. The actual protein concentration was then
determined for each sample using the slope intercept formula from the standard
curve (where (Y-A)/B = X). Then, the concentration of undiluted sample was
found by multiplying the concentration by the dilution factor (310). This number
was divided by the number of microliters of sample tested in 100 ptl of buffer.
This calculation was used to find the concentration of protein in 1 microliter. This
was then used to determine the amount of sample needed to load 10 uglul of
each sample in each gel (the desired protein concentration divided by actual
protein concentration equals the amount of sample needed). Each lane holds 19
ul of sample, thus the amount of sample needed subtracted from 15 determined
the remaining amount of sample buffer to add to the samples. See Figure 19 for
the standard curve and Table 10 for calculations of the crude protein

concentration of each sample.

76

Based on initial results from naive animals (Karson, 2001, unpublished
data), the desired protein concentration chosen was 10 uglul. The gels were
loaded such that the same type of fraction of the same brain region for different
animals was present on each gel.

Electrophoresis

Each sample was denatured by heating the microcentrifuge tubes in 85 °C
water for 8 minutes. After the samples were boiled, they were spun on a vortex
and chilled on ice for 2-3 minutes. 1 liter of running buffer was prepared using
100 ml of 10X TrislGlycine/SDS buffer (lnvitrogen) in 900 ml dH20. NuPAGE®
Novex 4-12% Bis-Tris pre—cast gels (15 wells, Invitrogen) were used and gels
were run in the Xcell SureLock" Mini-cell system (Invitrogen). The upper and
lower chambers of the mini-cell were filled with running buffer, and the gels were
loaded with samples and molecular weight markers (See-Blue Plus, Invitrogen) in
three different lanes. Two gels were run simultaneously at 125V for 90 minutes.

Once the gels were run completely, they were removed from the mini-cell
and prepared for transfer to membrane. Transfer Buffer was prepared using 50
ml of 20X NuPAGE® transfer buffer (lnvitrogen), 1 ml of NuPAGE® antioxidant
(lnvitrogen), 200ml methanol and 749 ml deionized water. Blotting pads were
soaked in700 ml of transfer buffer until saturated. Nitrocellulose membranes (45
um pore size) were cut to size and soaked in transfer buffer for five minutes. Gels
were transferred to membranes at 30V for 1 hour according to instructions

provided by Invitrogen.

77

 

 

to in
not
cale

inn

min
era
lSlg

ant

th
Sol

Me

Protein detection was performed by staining the membrane directly using
the VECTASTAIN® ABC system. To spare antibody, membranes were cut down
to include only the regions surrounding 22 kDa (approximate size of calexcitin
molecules) by using the marker lanes as cutting reference marks. Because
calexcitin was found in the optic and supraoesophageal lobes using
immunohistochemistry, it is reasonable to assume that the band on the
membrane at 22 kDa is, in fact, calexcitin. The membranes were rinsed for 15
minutes in 4 changes of 0.1% Tween 20 in Tris-buffered saline (TTBS) with
evaporated milk to block non-specific binding. 100 pl of goat normal serum
(Sigma) was added to 5 ml TTBS and 100 pl of biotinylated goat anti-rabbit
antibody (Sigma). The membranes were incubated in this solution with gentle
agitation for 30 minutes at room temperature. The membranes were then rinsed
with gentle agitation for 15 minutes in TTBS. Membranes were then transferred
to vectastain® ABC reagent, again incubated for 30 minutes at room temperature
with gentle agitation. The membranes were then transferred to a substrate
solution for 15 minutes (DAB substrate kit for peroxidase, Vector Laboratories).
Membranes were then rinsed in dH20, air dried, wrapped in cellophane and
stored at room temperature.

A flatbed scanner (Hewlett Packard Office Jet G85) was used to scan the
dyed membranes and create a series of high-resolution digital images. The
optical density (including staining intensity and width of the protein signal) was
found using NIH Image software. For each membrane, levels were calibrated

using the background as a zero value. Images of each membrane containing

78

homogenate, cytosolic and microsomal fractions from each piece of tissue from
each of the animals are shown in Figures 20-22.
8.3 Results

Because each sample was run on two different gels (one comparing the
same region and fraction for different animals, and one gel comparing different
regions in a single animal), the mean optical density for each brain region of each
individual was calculated by averaging the two optical density values.

A series of one-tailed Mann-Whitney U test were then used to compare
the mean optical density of calexcitin expression on the gels among spatially
trained, visually trained and control animals (Table 11). These analyses
indicated that calexcitin expression was significantly higher in visual cue trained
cuttlefish than spatially trained and control cuttlefish in the left optic lobe
homogenate, and the right optic lobe microsomal fractions. However, in the right
optic lobe cytosolic fraction, calexcitin expression was significantly lower in the
visual and spatial cue groups than in the control group. All other comparisons
were not statistically significant (Table 11).

Overall, the amount of calexcitin present in the homogenate of the three
brain regions did not differ (Friedman ANOVA by ranks, Fr=-107.25, p>0.50).
The amount of calexcitin present in the left and right optic lobes was significantly
greater than in the supraoesophageal lobes for both the cytosolic and
microsomal fractions (Friedman ANOVA by ranks, cytosolic Fr==19.8, p<0.01;
Fr=13.17, p<0.01).

79

To compare calexcitin expression in the right and left optic lobe within
each group, a series of Vlfilcoxon signed ranks tests were used. Results are
shown in Table 12. Calexcitin expression was significantly greater in the left
optic lobe homogenate than the right optic lobe homogenate for the visual cue
group. No other comparisons yielded significant results.

To determine whether initial door direction preference was related to
calexcitin expression in the left and right optic lobes, the data were placed in two
groups, left preference and right preference, and mean optical density was
compared between the left and right optic lobes. Among left-door preferring
cuttlefish, calexcitin expression was slightly higher in the left homogenate fraction
than in the right homogenate fraction (Wilxcoxon signs rank test, n=8, C=13,

=0.0938). However, expression did not differ in the cytosolic and microsomal
fractions (Wilxcoxon signs rank test, cytosolic, N=8, C=5, p>0.50; microsomal,
n=8, C=8, p>0.50). Among right-door preferring cuttlefish, expression did not
differ between the left and right optic lobes (Wilxcoxon signs rank test, n=5;
homogenate, C=17, p>0.50, cytosolic, C=6, p>0.50; microsomal, C=17, p>0.50).

To determine whether initial door preference was related to the weight of
the corresponding optic lobe, VVrIcoxon signs rank tests were used to compare
the weight of the right and left optic lobes among right and left-door preference
cuttlefish. This comparison was not statistically significant (right, n=7. C=12.5,

p>0.50; left, n=5, C=11, p=0.2188).

80

8.4 Discussion

The results here are similar to those found using immunohistochemistry.
Calexcitin expression appears greater in the optic lobes, specifically, the left optic
lobe homogenate and right optic lobe microsomes of visually trained cuttlefish
than in the optic lobes of spatially trained and control cuttlefish. Additionally, like
the immunohistochemical analysis, among visual cue trained cuttlefish, calexcitin
expression was greater in the left optic lobe than the right optic lobe
homogenate. Thus, the western blot results may further demonstrate the role of
the optic lobes in processing and storing of visual memories.

During associative Ieaming in Hermissenda, calexcitin is translocated from
the cytosol to the membrane of the endoplasmic reticulum (Nelson and Alkon,
1995; see review in Chapter 4). Thus, it was initially hypothesized that
associative learning would be associated with decreases in cytosolic calexcitin
and increases in microsomal calexcitin (Chapter 5). The increased levels of
microsomal calexcitin in the right optic lobes of visually trained cuttlefish may be
associated with the timing of a memory consolidation event. This interpretation
may also explain why there was more calexcitin in the cytosol of the right optic
lobe of control animals than in the trained groups.

Greater overall levels of calexcitin found in the optic lobes relative to the
amount found in the supraoesophageal lobes may indicate that calexcitin
signaling is important in Ieaming visual associations. Other associations that
involve different brain regions like spatial, tactile, or chemical, associations may

in fact rely on molecular cascades exclusive of the PKC/calexcitin system.

81

Overall, the left optic lobe yielded higher calexcitin expression than the
right optic lobe, regardless of the initial preference of the cuttlefish. In the visual-
cue-trained cuttlefish, it appears that calexcitin expression is increased in the
optic lobe opposite to the direction of the initial preference. In the control and
spatial-trained group the opposite trend is observed optic lobes, and, expression
is greater in the optic lobe associated with the initial direction preference.
Confounding the results is the immunohistochemical analysis which
demonstrated more calexcitin in the left optic lobes of all the cuttlefish, regardless
of their initial door preference. Again, it may be that cuttlefish prefer one eye
over the other, and processing in the non-preferred eye requires a greater
amount of neuronal "work." Thus, differences between the optic lobes may relate
to internal biases rather than external influences. The preference for spots or
stripes appears unrelated to calexcitin expression between optic lobes.

One other possible explanation is that differences between the eyes result
from lateralization of eye use. Octopus vulgaris uses monocular vision rather
than binocular vision to hunt and attack its prey (Muntz, 1961). Furthermore, O.
vulgaris shows a marked preference for using a particular eye for monocular
vision (Byme et al., 2002). The tendency to use one eye more than the other
may result in differential processing ability between the two eyes. In fact, door
preference may have resulted from a propensity to use one eye more than the
other. However, unlike octopuses, cuttlefish tend to use binocular vision rather

than monocular vision (Muntz, 1961). Further experimentation is necessary to

82

determine whether cuttlefish use monocular vision during visual associative
learning.

Alternatively, it is possible that different features of visual stimuli are
processed by the right and left optic lobes (Byme et al., 2002). This hypothesis
has not yet been tested. This is highly unlikely, as the structure does not appear
to differ greatly between the two optic lobes. More likely is that differences
between calexcitin in the right and left optic lobes may be associated with the
timing of transfer of information between the two lobes. Muntz (1961)
demonstrated that octopuses could be trained to make visual discriminations by
presenting objects in only one visual field (monocular training). These octopuses
then made correct responses to the same discrimination problem presented to
the untrained eye. Transfer between the lobes is interrupted when the optic
tracts are severed, or when the vertical lobe is removed. It is possible that the
training was monocular for some cuttlefish, and that complete transfer of
information between the optic lobes had not yet occurred at the time of
decapitation. No data exist regarding the timing of interocular transfer of visual
associations in cephalopods. While it is possible that greater calexcitin
expression in the left optic lobe results from the association of the stimuli first on

the left side, further experimentation remains necessary.

83

Chapter 9

General Discussion

The experiments successfully addressed the initial objectives. First, the
relative rates of simultaneous discrimination Ieaming based on visual versus
spatial cues were compared. Second, calexcitin expression in the optic and
supraoesophageal lobes was successfully traced. Third, calexcitin expression
was correlated with visual learning specifically. Finally, calexcitin expression
patterns in brain regions of different groups were successfully compared. The
results suggest several areas of further investigation, discussed in the following
sections of this chapter.

9.1 The Function of Learning in Cephalopods

Sepia officinalis is able to solve a simultaneous discrimination problem
using both visual and spatial cues, suggesting that both visual and spatial cues
are salient to cuttlefish in nature. However, it is possible that the ability to Ieam
spatial information relates more specifically to general processes underlying
flexible foraging strategies (lose-shift, win-stay; Mather, 1991; Hanlon and
Messenger, 1996). While the ability to make visual discriminations has been
documented in cuttlefish (Sanders and Young, 1940; Wells, 1962; Messenger,
1973), only recently has the ability to solve spatial problems been suggested
(Karson et al., 2003; Appendix C). In octopuses, spatial Ieaming has clear utility,
for example in locating their home dens (Mather, 1991). Unfortunately, there
have been no studies that directly relate to short and long-term movement
patterns of S. oflicinalis in nature. Such data are necessary to develop solid

hypotheses for the specific function or functions of spatial Ieaming in cuttlefish.

85

Regardless of the precise function, it seems unlikely that Sepia would
demonstrate "excess" spatial Ieaming abilities in the laboratory, that is, abilities
not used in nature (Hanlon and Messenger, 1996; Karson et al., 2003).

Many hypotheses have addressed general functions of Ieaming in
cephalopods. First, it is possible that reliance on Ieaming is imposed by the
complexity of the cephalopods' natural habitats (Packard, 1972; Hanlon and
Messenger, 1996). For example, coral reefs are highly competitive environments
with many predators and prey items. Cephalopods that live in the coral reefs
must constantly determine foraging and mating strategies while dealing with
many different predators and complex surrounds (Hanlon and Messenger, 1996).
Learning may play a key role in the ability of cephalopods to negotiate these
environmental complexities. However, it is important to note that learning may
not be restricted to those cephalopods living in coral reefs or other complex
environments (Hanlon and Messenger, 1996).

It is possible that natural selection acted on traits allowing various species
to select appropriate behaviors based on flexible use of stored memories, for
instance, complex associations previously encountered in the environment (Day
at al., 1999). In this framework, Ieaming in cephalopods may be considered a
mechanism related to phenotypic plasticity (Hanlon and Messenger, 1996).
Phenotypic plasticity is defined as an "adaptation to environmental variation in
time and space" (Dukas, 1998). In changing environments, organisms must
often change their responses to particular stimuli, and under such circumstances

these species benefit from phenotypic plasticity (Dukas, 1998).

86

As an example of the importance of plasticity, many cephalopod species
change habitats as they age (Hanlon and Messenger, 1996). These habitat
changes affect their diets as well as foraging and predator avoidance strategies
at different developmental stages and in different habitats. Learning allows
animals to change their responses to stimuli when previously experienced
associations between stimuli and events are no longer relevant (Dukas, 1998).
In cuttlefish, associative Ieaming is commonly used to discriminate between
different prey items (Messenger, 1973). This has utility bemuse the size, diet
and habitat changes radically during their lifetime (Wells, 1962). Because a
cuttlefish must alter its diet and hunting strategy as it develops, the ability to
replace old foraging tactics with newer ones would certainly be useful, and this is
greatly facilitated by Ieaming. The fact that Ieaming is widespread throughout
the coleoids, but not in nautiluses, supports fl'ris hypothesis (Wells, 1962). Hard-
shelled, slow-moving nautiluses, which do not show drastic changes in lifestyle
with growth, might be better off diverting resources elsewhere, as Ieaming would
not greatly increase these animals' abilities to forage or avoid predation (Solem,
1974)

Wells (1962) suggests that such developmental shifts in foraging behavior
are unlikely innate, as there is great individual variation in diet preference and
responses to different types of prey can be easily trained in cephalopods.
Learning is an effective strategy only when the organism's environment remains

stable to some extent. If the meaning of a particular stimulus changes too

frequently, the costs of learning and re—leaming probably exceed the very short-
lived benefits (Dukas 1998).

Further experiments comparing learning abilities among different species
of cuttlefish (as well as between octopuses and cuttlefish) may allow for
additional hypotheses related to the function of Ieaming in different cephalopod
species. Specifically, observations of S. pharaonis and S. officinalis in the
laboratory indicate striking behavioral dissimilarities (L. Dickel, pers. comm.,
2003). For instance, while S. ofiicinalis tends to remain partially buried under the
substrate, under plants or in corners of laboratory tanks, S. pharaonis constantly
swims around the tank (pers. obs.). If such differences are present in natural
populations, one might expect that these species use spatial information
differently. S. ofi‘icinalis may use spatial Ieaming to find adequate "resting spo
or S. pharaonis may use spatial information to determine their swimming routes.
Or, conceivably, in nature one of these species may not use spatial Ieaming at
all. Similarly, one might expect that cephalopods living in dark (deep) habitats
rely on visual information and thus Ieaming to a lesser extent than those in
shallower regions with more ambient light (Maddock and Young, 1987). Thus,
further comparisons among cephalopod species from different marine habitats
and comparative observations of field behavior remain necessary.

One final consideration of this topic is related to Ieaming and the short life
cycles in cephalopods. Most extant cephalopods, with the exception of Nautilus,
"live fast and die young" (Hanlon and Messenger, 1996). A short lifespan may

be viewed as an adaptation to the environment. Unlike teleosts, cephalopods do

88

not store lipid reserves. It may be advantageous for coleoids to attain maturity
and reproduce quickly, rather than risking surviving during a period when energy
demands are high but suitable prey is scarce (O'Dor and Webber, 1986).
Octopuses and decapods hunt and eat daily (Hanlon and Messenger, 1996). An
inadequate foraging period for an adult cephalopod may be a good feeding
period for a hatchling that is much smaller and feeding on different food sources.
In addition to differences in feeding behaviors in adult and juvenile cephalopods,
there are differences in the relative efficiency of locomotion. Locomotion by jet
power does not work as efficiently as locomotion by fin movements unless the
organism. is relatively small (O'Dor and Webber, 1986). Short life cycles may
allow cephalopods spend a proportionally large portion of their life in a
developmental stage during which they have adequate food resources to easily
accommodate their energetic requirements and more efficient locomotion. This
being the case, short life cycles may have evolved in a species that has complex
Ieaming ability rather than vice versa.
9.2 Relevant and Irrelevant Cues

The results indicated that the combined-cue group had greater initial
difficulty Ieaming the maze than either the visual or spatial cue groups. It would
be informative to perform experiments from the same population of cuttleflsh to
determine the nature of the difference in Ieaming ability between the combined
cue and single relevant-cue groups. It may be that Ieaming is enhanced by the
simultaneous presentation of irrelevant and relevant cues, whereas linking these

cues may "confuse" the cuttlefish (Sanders, 1975). Perhaps cuttlefish are not

89

only learning that one door is correct, but also that the other door is incorrect
(Sanders, 1975). To examine this, serial reversal experiments should compare
between the combined-cue and the single relevant cue groups (i.e visual cue or
spatial cue relevant) and determine whether these Cuttlefish learn reversals. An
additional extension of the behavioral work would be to set up a conditional
Ieaming paradigm where cuttlefish must learn to choose an exit based on a
visual cue, but the correct visual cue depends on the context (for example, light
on or off). Such experiments are currently underway (Boal and Karson,
experiments to begin 612003).
9.3 Individual Variation in Behavior

The behavioral results presented here support previous studies
demonstrating a tremendous amount of individual variation in Ieaming abilities
between individual cephalopod subjects (Mather and Anderson, 1993).
Individual variation in behavioral abilities may be a result of particular prior
experiences. To test this, one might compare discrimination Ieaming between
enriched environment raised and non-enriched environment raised cuttlefish or
between wild-caught and laboratory-reared cuttlefish. Previous authors have
suggested different "personalities," or temperaments, among individual
octopuses (Mather and Anderson, 1993). Such differences may not only result
from individual experience, but also from different levels of motivation or
agitation. Thus, it may be informative to develop a variety of related Ieaming
paradigms and determine whether faster Ieaming rates of particular individuals

remain consistent regardless of the paradigm used. Such experiments might

also demonstrate a general correlation between particular types of behavioral
abilities of cephalopods, as has been observed in several vertebrate models.
9.4 Calexcitin and Visual Associations

While the immunohistochemical data indicated that the local patterns of
staining did not appear to differ between the trained and control groups, the
intensity of calexcitin staining within particular regions did differ between the
groups, indicating higher levels of intracellular calexcitin in some groups. More
intensely calexcitin-stained cells were located within the optic lobes of visual-cue
trained cuttlefish than spatial or control groups. This finding was supported by
the western blot data, which indicated that calexcitin expression was greater in
the optic lobe homogenate and microsomal fractions of visual cue-trained
cuttlefish than in the homogenate and microsomal fractions of the spatial and
control groups.

Failure to demonstrate differences in calexcitin between spatial-cue
trained and control cuttlefish may indicate one of several things. First, it is
possible that the optic lobe is involved in Ieaming visual but not other types of
associations. In octopuses, tactile and visual associations are processed
separately, though they do interact in the vertical/superior frontal lobe system
(Young, 1991). It may be useful to compare calexcitin expression in visual-
leaming-related brain areas of octopuses that have Ieamed a visual association
to calexcitin expression in the chemotactile-learning-related brain area (the
inferior frontal lobe) of octopuses who have Ieamed chemotactile associations. I

suggest using Octopus because the circuitry underlying visual and chemotactile

91

Ieaming is better understood in octopuses than cuttlefish (see Young, 1991).
Spatial Ieaming may occur in yet another set of lobes in the cephalopod brain,
perhaps involving basal or suboesophageal regions that are more directly related
to the regulation of motor behavior. Aithough initial hypotheses suggested that
the optic lobes process spatial information from the two visual fields, there is no
reason to assume that the optic lobes are involved with spatial Ieaming at all (JB
Messenger, 2003, pers. comm).

A second possible explanation for the lack of calexcitin increase during
spatial Ieaming is that calexcitin is involved specifically in the molecular cascade
underlying visual Ieaming but not spatial Ieaming. This has been proposed in
other organisms (e.g. Hermissenda) where calexcitin appears exclusively
involved with visual systems (Kuzirian et al., 2003). Much like Hermissenda, a
great deal of the Ieaming and memory processes in cephalopods occur within in
the visual processors themselves (Alkon, 1989; Alkon et al., 1990). Kuzirian et
al. (2003) conclude that under such circumstances, Ieaming-induced changes in
memory would be found in cells that modulate the output of downstream neurons
and not the downstream neurons themselves. In other words, differences in
intracellular calexcitin levels would be detected within the visual cells rather than
the motor neurons these cells regulate. Certainly, the results here indicated
changes in visual cells and not within the motor-related brain areas. However,
baseline levels of calexcitin were found within motor activity-regulating regions of
the supraoesophageal mass and it is important to note that the dorsal and medial

basal lobes are connected to the optic tracts. It would be interesting to use

immunohistochemistry and western blotting to determine whether calexcitin is
also present in the suboesophageal regions that are motor-regulators and are not
connected to the visual system. The presence of calexcitin in these regions may
suggest that calexcitin does not strictly play a role in visual Ieaming and memory.
The presence of calexcitin in non-visual processing brain areas may
indicate that the role is not limited to visual associations or visual Ieaming
(Kuzirian et al., 2003). Even when found in brain regions associated with visual
Ieaming, in cuttlefish, calexcitin may simply be involved in visual processing and
not visual associative Ieaming. In fact, calexcitin may be present in any cells
associated with forms of Ieaming and memory or even other functions that are
calcium regulated or modulated (Kuzirian et al., 2003). To test whether calexcitin
in the optic or supraoesophageal lobes is associated with visual Ieaming
specifically, it would be interesting to block calexcitin in these regions using lead
or perhaps anti-sense RNA or antisense peptides. By doing so, one could
determine whether blocking calexcitin in the manner also blocks visual learning.
This would more clearly demonstrate a direct relationship between calexcitin and
visual Ieaming, rather than merely between calexcitin expression and neuronal
activity. One also might use immunohistochemical studies of calexcitin to
observe patterns of expression in the motor-regulatory suboesophageal lobes.
Kuzirian et al. (2003) further surmise that the baseline levels of calexcitin
present in naive snails and in brain regions unaffected by Ieaming suggest that
calexcitin remains at a particular homeostatic level and/or is regulated by other

sensory inputs. A similar phenomenon may be observed in cuttlefish. To test

93

this hypothesis, one might observe the effects of different types of sensory input
exclusive of visual input on intracellular calexcitin levels. This could be
accomplished using blinded cephalopods, or cephalopods kept in the dark
(Kuzirian, pers 00mm, 2003). Again, observing tactile Ieaming and calexcitin
expression the inferior frontal lobe of octopuses may also help get at this
question.
9.5 Time Course of Learning and Calexcitin Expression

It would be informative to observe the time course of visual discrimination
learning in this paradigm and compare it to time courses of calexcitin expression
in the optic lobes. It may be particulariy useful to observe the relationship
between calexcitin expression and the visual Ieaming paradigm where cuttlefish
are trained to associate the visual stimulus of a prey item with pain (Messenger,
1973). This paradigm is recommended, as it has been particularly well studied
and time course has been determined (Messenger, 1973b; Agin et al., 1998).
Post-Ieaming calexcitin increases return to baseline levels 24 hours after
associative Ieaming in Hermissenda (Kuzirian et al., 2003). It would be useful to
determine whether such results are found during associative Ieaming in
cephalopods. Additionally, it would be interesting to block calexcitin at particular
time intervals to determine whether this has any effect on learning. Such
experiments are currently being attempted with Hermissenda (Kuzirian et al.,
2003). One way to block learning may be through the introduction of lead, which
has been shown to block Ieaming in Hermissenda by blocking calcium channels,

thereby inhibiting calexcitin (Kuzirian et al., 1996 and 1998). In these

experiments, blocking with lead did not have toxic effects on Hermissenda or
block normal activity patterns (Kuzirian et al., 1998). This may be an equally
useful technique to block visual associative Ieaming in cephalopod molluscs.
Antisense RNA and peptides could also be used to block calexcitin and thus
perhaps associative Ieaming.
9.6 lnterocular Transfer

Perhaps one of the most confusing results of these experiments was the
difference in calexcitin expression observed between the left and right optic lobes
of the various groups. Not only did calexcitin expression differ between the right
and left optic lobes, visual Ieaming yielded greater increases in one of the two
optic lobes. Such results may warrant further studies examining the extent to
which the left and right optic lobes (or other brain regions) are functionally
lateralized. Greater calexcitin expression observed in one optic lobe, may simply
imply that a greater number of connections must be established to make
associative connections in one optic lobe relative to the other. For instance,
because the non-preferred eye is used less frequently it may require a greater
threshold of stimulation to initiate activity because there are fewer pre—existing
associative connections between cells. It is possible, then, that biases in eye use
do directly impact calexcitin expression and result in differences between the
right and left optic lobes.

Other lobes may also exhibit lateralization of function that is specific to
particular sources of sensory input, or even from the right and left visual fields.

However, Wells (1978) asserts that octopus brains are probably functionally

95

symmetrical in terms of the right and left sides of the body and that either side of
the brain could control the whole octopus. Because the structure of the right and
left optic lobes does not appear to differ, and there is no reason to assume that
the right and left optic lobes process different features of the visual field (J.B.
Messenger, pers. comm, 2003). Furthermore, hemispheric lateralization is not
found in other invertebrates. Further details related to the type of information that
is transferred between the optic lobes of cephalopods and the timing of this
transfer are yet to be ascertained.

Regardless of whether there is an equivalent to "hemispheric
lateralization" in cephalopods, as is observed in vertebrate brains, interocular
transfer is an important aspect of visual processing in cephalopods (Muntz,
1962). lnterocular transfer in cephalopods occurs via the vertical lobe (Muntz,
1962). Octopuses use monocular vision, thus interocular transfer is quite
important. Octopuses trained to use on eye to make a simple visual
discrimination show complete transfer to the other side (Muntz, 1961). Transfer
is blocked when connections between the vertical and superior frontal lobes are
severed (Muntz, 1961). Similarly, touch experiments with octopuses show
transfer of information between arms on the left and right side of the body via the
supraoesophageal lobes (Wells and Young, 1966). Tactie transfer experiments
indicate that if the supraoesophageal lobes are split before training, tactile
learning occurs only in the trained side, whereas the untrained and trained sides
perform equally if the split is made after training (Wells and Young, 1966). In

light of evidence here suggesting molecular differences between the right and left

optic lobes in cuttlefish, and other studies suggesting biases in eye use among
individual octopuses, such inquiries warrant further exploration (Byme et al.,
2002)
9.7 Future Directions

It would be interesting to compare the vertebrate-like Ieaming capabilities
and the underlying neural correlates of the abilites of cephalopods to that of
vertebrates. It would be equally interesting to compare the neural correlates
underlying Ieaming in cephalopods to those observed in simple molluscan
systems. Comparisons between cephalopods and vertebrates may further
illuminate general principles underlying complex forms of Ieaming. Whereas
comparisons between cephalopods and simple molluscan systems could inform
us, for example, to what extent cephalopods use complex Ieaming and to what
extent they solve problems using more ancient molluscan strategies. Eventually,
comparisons of complex Ieaming in vertebrates and cephalopods will aid in our
understanding of the principles underlying the organization and evolution of

complex brains and behaviors.

APPENDIX A

Tables

98

Table 1: Brain lobes associated with the supraoesophageal mass Sepia
ofi‘icinalis and presumed functions as identified in the squid Loligo (Young, 1974,
and 1977).

 

 

f Region Function Controlled
Perioesophageal Regulation of jet propulsion and
mass escape behavior
Subpendunculate Regulation of ocular pressure
lobe

Precommisural lobe Unknown

Anterior basal lobes Steering (dipping/rolling)
Posterior basal lobes Steering (yawing), jet
(median and dorsal) propulsion, tentacle movement
and chromatophores

I nferior frontal lobe Tactile Ieaming and memory

Superior frontal lobe Visual Ieaming and memory,
control of attack behavior,

 

 

 

 

Vertical lobe Tactile and visual Ieaming and
memory
subvertical lobe Learning and Memory
@ic lobes Vision
99

Table 2: Initial descriptions of the cuttlefish trained in the behavioral
experiments, including name, sex, group, mantle length at beginning of
experiment, door direction preference and visual-cue preference.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Name Sex Group Mantle Length Directional Visual-cue

(cm) preference preference
Bruiser male control 10.5 left spot
Freckle female control 6.3 right spot
Isis male control 7.4 right spot
Pokey female control 8 right spot
Nougat male control 7.5 right stripe
Raoul male control 8.8 right stripe
Willow male control 7.3 left spot
Stohlio male control 7.2 @th stripe
Howard male spatial 9.7 left stripe
Kidd male spatial 7.2 right spot
Satine female spatial 7.6 right spot
Snookums male spatial 9 left spot
Bindi female spatial 8.1 right spot
Marilyn male spatial 8.2 right spot
Prince male spatial 7.1 left stripe
Nestle male spatial 8.5 right spot
Guy male visual 9.1 fght spot
lpanema female visual 7.6 left stripe
Mustarg female visual 7.6 left spot
Orion male visual 6.2 right stripe
Scratch male visual 7 Eff stripe
7&7 female visual 8.4 left stripe
Stone female visual 7.9 right spot
Bilbo male visual 6.4 Teft spot
Lola female visual 7 right stripe
Matey male combined 6.8 right stripe
Chip male combined 7.3 79m stripe
WW female combined 7.5 right stripe
Fleck male combined 7.4 (an spot
Dottie male combined 7.8 left spot
Mrs. Dash female combined 7.7 right stripe
Pan male combined 9.7 left spot
Maximus male combined 10.1 rig ht stripe
Biggie male combined 9.6 right stripe
Little Suzy female combined 6.4 left spot
Fubar female combined 6.2 left spot
Lena female combined 7.1 left spot

100

 

Table 3: Repeated measures ANOVA results of cuttlefish maze performance
across the first, middle and last block of experimental trials (or all four blocks of
trials for control cuttlefish) for all groups. Analyses include decrease in mean
maze escape time, decrease in mean overall error, decrease in percentage of
error exclusive of non-escapes and decrease in non-escapes. Significant
decreases demonstrate Ieaming and are indicated by a p-value listed in boldface.

 

 

 

 

 

 

 

 

 

 

[ Analysis n Group F p-value
. Decrease in escape 8 Control 2.094(321) p>0.25
time between the first,
middle and last block 8 Visual 9.075 (2,14, p<0.005
of trials
8 Spatial 8.929 (2,14) p<0.005
12 Combined 14.311 (2, 22) p<0.0005
Decrease in overall 8 Control 2.000 (3, 21) p>0.10
error between the first,
middle and last block 8 Visual 11.53 (2,14) p<0.001
8 Spatial 12.631 (2, 14) p<0.001
12 Combined 28.774 (2,, 22) p<0.0005
Decrease in 8 Control 1.000 (3, 21) p>0.25
percentage of error
(exclusive of non- 8 Visual 1.663 (2,14) p >0.25
escapes)
8 Spatial 5.810 (2, 14) p<0.025
12 Combined 26.143 (2, 22) p<0.0005
fiecrease in 8 Control 2.492 (3, 21) p>0.25
percentage of non-
escapes 8 Visual 2.726 (2,14) p>0.1
8 Spatial 14.212 (2, 14) p<0.0005
12 Combined 4.224 (2, 22, p<0.05
\

 

101

 

Table 4: Mean 1- standard error of maze escape times (seconds) for the first,
middle and last block of five trials during training for male and female cuttlefish in
the control, visual, spatial and combined cue groups. Control group includes the
2"“ block of five trials as "middle block." None of these differences were
statistically significant (Mann-Whitney U test: visual, U (4,4) =13, p>0.10, spatial U
(2,5) =10, p>0.10, combined, U (5,7) =22, p>0.10).

 

 

 

 

 

 

First Block Middle Block Last Block
Male Female Male Female Male Female
109.23 114.80 81.37 120.40 32.13 26.70
Control 3 1- : t i 2
38.69 55.40 61.64 77.20 14.57 1.10
338.70 190.00 235.55 144.70 36.55 31.00
Visual t z 2 t z 2
90.90 44.77 57.26 67.30 15.94 7.57
238.96 381.80 135.03 141.70 27.76 15.00
Spatial : z 1- : z 1-
79.18 38.2 34.02 76.7 6.65 4.6
406.26 514.72 384.31 387.12 49.43 46.8
Combined i: z z 1 i 1
60.63 37.28 59.43 60.70 17.52 13.07

 

 

 

 

 

 

 

102

 

Table 5: Mean 2: standard error of overall percentage of error for the first, middle
and last block of five trials during training for male and female cuttlefish in the
control, visual, spatial and combined cue groups. Control group includes the 2"d
block of five trials as "middle block." The differences between males and females
were insignificant (Mann-Whitney U Test: visual-cue group, U (4,4) = 13, p>0.10,
spatial-cue group, U (2,5) =10, p>0.10, combined-cue group, U (5,7) =20, p>0.10,
COl'ltTOI U (2,5)=10, p>0.10).

 

 

 

 

 

 

 

 

 

 

 

 

First Block Middle Block Last Block
Male Female Male Female Male Female
30.00% 40.00% 13.33% 50.00% 10.00% 20.00%
Control :9: a: z z z 2
8.56% 0.00% 13.33% 10.00% 6.83% 0.00%
85.00% 60.00% 55.00% 50.00% 10.00% 15.00%
Visual : z z 1 :L- 1-
9.57% 11.55% 15.00% 20.82% 5.77% 5.00%
57.67% 100% 43.33% 50.00% 6.67% 0%
Spatial :t z z z z 2».
14.06% 0% 12.02% 10.00% 4.21% 0%
80.00% 92.00% 88.57% 84.00% 0% 8.00%
Combined 2 4.» 2 2 t 2
11.55% 4.90% 23.04% 7.48% 0% 8.00%

 

103

 

Table 6: Mean number of intensely stained calexcitin cells in various 200um by
200p.m by 30pm cuttlefish brain regions post-training. Kruskall-Wallis one-way
ANOVA was then used to determine whether there were significant differences
between the groups. Regions printed in italic indicate statistically significant
differences between the groups (Refer to Table 7). Counts displayed in boldface
indicate values significantly larger than the other groups.

 

 

 

 

Region Control Visual Spatial
Vertical Lobe 2.8752025 2.82017 2.6:0.12
Subvertical Lobe 141511.02 5.130.833 6.5552064
Dorsal Basal Lobe 7.1352039 6.92:0.09 7.03:0.9

 

Right Optic Lobe (Deep Retina) 3.67:0.25 6.18:1.09 3.762029

 

Right Optic Lobe (Medulla) 3.21-0.46 5.49:0.30 4.75:0.25

 

Left Optic Lobe (Deep Retina) 4.78:0.19 6.610.56 5.00:0.41

 

 

Lefi Optic Lobe (Medulla) 377520.81 6.36:0.81 5.00:0.41

 

 

 

 

 

104

Table 7: Kruskall-Wallis one way ANOVA results for differences in the mean
number of strongly calexcitin stained calls in various brain regions (refer also to
Table 6). Significant differences between the groups are indicated by a boldface
p-value.

 

 

 

 

Region p-value
Vertical Lobe N=12 (4,4,4), KW=3. 16 p>0.10
Subvertical Lobe N=12 (4,4,4), KW=6.31 p<0.05
Dorsal Basal Lobe N=12 (4,4,4), KW=0.77 p>0.10

 

Right Optic Lobe (Deep Retina) N=12 (4,4,4), KW=10.86 p<0.001

 

Right Optic Lobe (Medulla) N=12 (4,4,4), KW=7.30 P<0.05

 

Left Optic Lobe (Deep Retina) N=12 (4,4,4), KW=7.431 P<0.05

 

 

Left Optic Lobe (Medulla) N=12 (4,4,4), KW=4.813 P<0.10

 

 

 

 

105

Table 8: Wilcoxon signed ranks test results, comparing the mean number of
strongly calexcitin stained cells in the left and right optic lobe for each
experimental group. Significant results indicated in boldface (note, small sample

size precluded p-values less than 0.0625).

 

[7 Left optic lobe Left optic lobe
plexiform versus medulla versus
right optic lobe right optic lobe

plexiform medulla

 

Control (nfi) T“=10, p=0.0625 T '=6, p=0.4375

 

Visual (n=4) 'l"=5.5, p=.5000 T=7, p=0.3125

 

Spatial (n=4) 1*=1o, p=0.0625 1*=7, p=0.3125

 

 

 

 

 

106

Table 9: Data on the initial tissue weight of the pieces of tissue used for western
blot analysis. Initial direction preference and group data also included. See
Table 2 for additional information.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Ft Cuttlefish Group Brain region Weight Direction
(1) preference
f Stone Visual Left Optic Lobe 0.29 Right
Right Optic Lobe 0.21
Supraeosophageal Lobes 0.67
r Lola Visual Left Optic Lobe 0.17 Right
Right Optic Lobe 0.2
Supraeosophageal Lobes 0.67
Guy Visual Left Optic Lobe 0.22 Right
Right Optic Lobe 0.24
Supraeosophageal Lobes 0.67
Scratch Visual Left Optic Lobe 0.25 Left
Right Optic Lobe 0.22
i Sgimeosophggeal Lobes 0.44
Howard Spatial Left Optic Lobe 0.33 Left
Right Optic Lobe 0.36
Supraeosophageal Lobes 0.54
Snookums Spatial Left Optic Lobe 0.23 Left
Right Optic Lobe 0.24
Supraeosophageal Lobes 0.44
Marilyn Spatial Left Optic Lobe 0.14 Right
Right Optic Lobe 0.18
X Supraeosophageal Lobes 0.44
Isis Control Left Optic Lobe 0.53 Right
Right Optic Lobe 0.44
t Supraeosophageal Lobes 1 .1 5
Nougat Control Left Optic Lobe 0.27 Right
Right Optic Lobe 0.22
\ Supraeosophageal Lobes 0.65
Willow Control Left Optic Lobe 0.28 Left
Right Optic Lobe 0.24
\ Supraeosophageal Lobes 0.72
Bilbo Visual Left Optic Lobe 0.38 Left
Right Optic Lobe 0.35
Supraeosophageal Lobes 0.86
Freckle Control Left Optic Lobe 0.35 Right
Right Optic Lobe 0.44
Supraeosophgeal Lobes 1.24

 

 

 

 

 

107

 

Table 10: Bradford method results used to determine the protein concentration
for each fraction (homogenate, cytosolic and microsomal fractions of the left optic
lobe (LOL), right optic lobe (ROL), and supraoesophageal lobes (CNS) of each
cuttlefish. Columns A and B indicate the identity of each cuttlefish, the group they
belonged to and the fraction of interest. Column C shows the absorption reading
of each fraction read in a spectrophotometer at 595 nm. Column D shows the
amount of protein in each diluted fraction, found using the slope intercept
equation Y=-0.005631 + 0.00386119X (see Figure 26). Column E shows the
amount of protein in undiluted samples (Column D multiplied by final dilution
factor of 310). Column F shows the amount of sample loaded to yield a protein
concentration of 10 ug/pl. Column G shows the amount of additional buffer

required to load the sample (19 pl total) into each lane.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

108

 

 

 

A B C D E F G
Fraction Absorption Diluted final sample Buffer
Value Protein protein loaded loadedl
(Y) uglml uglul ul ul
Willow LOL HOM 0.001 1.7174 0.53239 18.78 0.22
ROL HOM 0.006 3.0123 0.93382 10.71 8.29
control CNS HOM 0.12 32.537 10.0864 0.99 18.01
LOL CYT 0.047 1 3.631 4.22555 2.37 16. 63
ROL CYT 0.056 15.962 4.94812 2.02 16.98
CNS CYT 0.112 30.465 9.44414 1.06 17.94
LOL MlC 0.08 22.177 6.87498 1.45 17.55
ROL MIC 0.129 34.868 10.809 0.93 18.07
P CNS MIC 0.152 40.824 12.6556 0.79 18.21
Lola LOL HOM 0.368 96.766 29.9974 0.33 18.67
ROL HOM 0.001 1.7174 0.53239 18.78 0.22
visual CNS HOM 0.073 20.364 6.31298 1 .58 17.42
LOL CYT 0.047 13.631 4.2555 2.37 16.63
ROL CYT 0.21 55.846 17.312 0.58 18.42
CNS CYT 0.205 54.551 16.9107 0.59 18.41
LOL MIC 0.024 7.6741 2.37897 4.20 14.80
ROL MIC 0.016 5.6022 1.73668 5.76 13.24
CNS MIC 0.058 16.48 5.10869 1.96 17.04
Howard LOL HOM 0.097 26.58 8.23985 1 .21 17. 79
ROL HOM 0.316 83.298 25.825 0.39 18.61
spatial CNS HOM 0.182 48.594 15.0642 0.66 18.34
LOL CYT 0.048 13.89 4.30583 2.32 16.68

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 10 ROL CYT 0.057 16.21 5.02841 1.99 17.01
Cont'd CNS CYT 0.342 90.032 27.9099 0.36 18. 64
LOL MIC 0.032 9.746 3.02125 3.31 15.69
ROL MIC 0.024 7.6741 2.37897 4.20 14.80
CNS MIC 0.151 40.565 12.5753 0.80 18.20

Snookums LOL HOM 0.005 2.7533 0.85353 11 .72 7.28
ROL HOM 0.054 15.444 4.78755 2.09 16.91

spatial CNS HOM 0.052 14.926 4.62698 2.16 16.84
H LOL CYT 0.214 56.882 17.6333 0.57 18.43
ROL CYT 0.201 53.515 16.5896 0.60 18.40

CNS CYT 0.068 19.07 5.91155 1 .69 17.31
LOL MIC 0.269 71.126 22.049 0.45 18.55
ROL MIC 0.229 60.766 1 8.8376 0.53 18.47
CNS MIC 0.024 7.6741 2.37897 4.20 14.80

Guy LOL HOM 0.152 40.824 12.6556 0.79 18. 21
ROL HOM 0.158 42.378 13.1373 0.76 18.24
visual CNS HOM 0.087 23.99 7.43699 1 .34 17. 66
LOL CYT 0.096 26.321 8.15956 1 .23 1 7. 77
ROL CYT 0.1 14 30.983 9.60471 1 .04 1 7. 96
CNS CYT 0.753 196.48 60.9075 0.1 6 18.84
LOL MIC 0.037 11.041 3.4268 2.92 16.08
ROL MIC 0.116 31.501 9.76528 1 .02 17. 98

CNS MIC 0.068 19.07 5.91155 1.69 17.31
Scratch LOL HOM 0.101 27.616 8.56099 1 .1 7 1 7. 83
ROL HOM 0.211 56.105 17.3925 0.57 18. 43
visual CNS HOM 0.03 9.228 2.86068 3.50 15.50
LOL CYT 0.073 20.364 6. 31298 1 .58 1 7. 42
ROL CYT 0.378 99.356 30.8002 0.32 18.68

CNS CYT 0.175 46.781 14.5022 0.69 18.31
LOL MIC 0.203 54.033 16.7502 0.60 18.40
ROL MIC 0.014 5.0842 1.57611 6.34 12.66
CNS MIC 0.08 22.177 6.87498 1 .45 1 7. 55
Isis LOL HOM 0.066 18.552 5.75098 1 .74 17.26
ROL HOM 0.00001 1.461 0.4529 22.08 -3.08
control CNS HOM 0.051 14.667 4.54669 2.20 16.80
LOL CYT 0.069 19.329 5.99184 1 .67 1 7. 33

 

109

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 10 ROL CYT 0.139 37.458 11.6119 0.86 18.14
Cont'd CNS CYT 0.068 19.07 5.91155 1 .69 17.31
LOL MIC 0.103 28.134 8.72156 1.15 17.85

ROL MIC 0.112 30.465 9.44414 1 .06 17.94

CNS MIC 0.17 45.486 14.1007 0.71 18.29

Marilyn LOL HOM 0.025 7.9331 2.45925 4.07 14.93
ROL HOM 0.048 13.89 4.30583 2.32 16.68

spatial CNS HOM 0.082 22.695 7.03556 1.42 17.58
LOL CYT 0.001 1.7174 0.53239 18.78 0.22

ROL CYT 0.001 1.7174 0.53239 18.78 0.22

CNS CYT 0.052 14.926 4.62698 2.16 16.84

LOL MIC 0.002 1.9764 0.61267 16.32 2.68

ROL MIC 0.001 1.7174 0.53239 18.78 0.22

CNS MIC 0.09 24.767 7.67785 1.30 17. 70

Nougat LOL HOM 0.172 46.004 14.2613 0.70 18.30
ROL HOM 0.051 14.667 4.54669 2.20 16.80

control CNS HOM 0.001 1.7174 0.53239 18.78 0.22
LOL CYT 0.05 14.408 4.4664 2.24 16.76
ROL CYT 0.059 16.739 5.18898 1 .93 1 7. 07

CNS CYT 0.03 9.228 2.86068 3.50 15.50

LOL MIC 0.001 1.7174 0.53239 18.78 0.22

ROL MIC 0.133 35.904 11.1301 0.90 18.10

CNS MIC 0.057 16.21 5.02841 1.99 17.01

Stone LOL HOM 0.075 20.882 6.47355 1 .54 17.46
ROL HOM 0.062 17.516 5.42984 1.84 17.16
visual CNS HOM 0.072 20.105 6.2327 1.60 17.40
LOL CYT 0.059 16.739 5.18898 1 .93 1 7.07

ROL CYT 0.031 9.487 2.94097 3.40 15.60

CNS CYT 0.053 15.185 4.70726 2.12 16.88

LOL MIC 0.051 14.667 4.54669 2.20 16.80

ROL MIC 0.237 62.838 19.4799 0.51 18.49

CNS MIC 0.022 7.1561 2.21839 4.51 14.49

Freckle LOL HOM 0.001 1.7174 0.53239 18.78 0. 22
ROL HOM 0.034 10.264 3.18183 3.14 15.86
control CNS HOM 0.081 22.436 6.95527 1.44 17.56
LOL CYT 0.052 14.926 4.62698 2.16 16.84

 

 

 

 

 

 

 

110

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 10 ROL CYT 0.06 16.998 5.26926 1.90 17.10
Cont'd CNS CYT 0.346 91.068 28.2311 0.35 18.65
LOL MIC 0.068 19.07 5.91155 1 .69 17.31

ROL MIC 0.184 49.112 15.247 0.66 18. 34

CNS MIC 0.172 46.004 14.2613 0.70 18.30

Bilbo LOL HOM 0.102 27.875 8.64128 1 .16 17.84
ROL HOM 0.037 11.041 3.4268 2.92 16.08

visual CNS HOM 0.211 56.105 17.3925 0.57 18. 43
LOL CYT 0.077 21.4 6.63413 1 .51 17.49

ROL CYT 0.124 33.573 10.4076 0.96 18.04

CNS CYT 0.147 39.53 12.2541 0.82 18.18

LOL MIC 0.147 39.53 12.2541 0.82 18.18

ROL MIC 0.121 32.796 10.1667 0.98 18.02

CNS MIC 0.189 50.407 15.6262 0.64 18.36

 

 

 

 

 

 

 

Ill

 

 

Table 11: Mann Whitney U test results comparing mean calexcitin expression on
western blot membranes among the homogenate, cytosolic and microsomal
fractions of the left optic lobe, right optic lobe and supraoesophageal lobes
between each experimental group (visual-cue, spatial-cue and motor control). A
boldfaced p-value indicates significant differences between groups, and the
significantly higher group is also denoted in boldface.

 

 

 

 

 

 

 

 

 

 

 

Region Comparison U P
Left Optic Lobe Homogenate Spatial vs. Visual 14 3,5 =0.10
Control vs. Visual 16 4,5 =0.1 0
Spatial vs. Control 4 3,4 >0.10
Right Optic Lobe Spatial vs. Visual 12 3,5 >0.10
Homogenate Control vs. Visual 15 4,5 >0.10
Spatial vs. Control 7 3,4 >0.10
Supraoesophageal Lobe Spatial vs. Visual 3 3,5 >0.10
Homogenate Control vs. Visual 8 4,5 >0.10
Spatial vs. Control 3 3 .4 >010
Left Optic Lobe Cytosolic Spatial vs. Visual 9 3,5 >0.10
Control vs. Visual 12 4,5 >0.10
Spatial vs. Control 7 3,4 >0.10
Right Optic Lobe Cytosolic Spatial vs. Visual 12 3,5 >0.10
Control vs. Visual 20 4,5 =0.001
Sgatial vs. Control 12 3,4 =0.05
Supraoesophageal Lobe Spatial vs. Visual 7 3,5 >0.10
Cytosolic Control vs. Visual 8 4,5 >0.10
Spatial vs. Control 9 3,4 >O.10
Left Optic Lobe Microsomal Spatial vs. Visual 9 3,5 <0.05
Control vs. Visual 8 4,5 >0.10
Spatial vs. Control 9 3,4 >0.10
Right Optic Lobe Microsomal Spatial vs. Visual 13 3,5 =0.10
Control vs. Visual 18 4,5 =0.05
L Spatial vs. Control 8 3,4 >0.10
i Supraoesophageal Lobe Spatial vs. Visual 10 3,5 >0.10
I Microsomal Control vs. Visual 5 4.5 >0.10
L Spatial vs. Control 5 3,4 >0.10

 

 

 

 

112

 

 

Table 12: Wilcoxon signed ranks test comparing mean calexcitin expression on
western blot membranes in the left and right optic lobe for each experimental
group- Significant differences between the lobes are indicated in boldface.

 

 

Left optic lobe
homogenate

versus right optic

Left optic lobe
cytosolic versus

right optic lobe

Left optic lobe
microsomal

versus right optic

 

 

 

L lobe homogenate cytosolic lobe microsomal
< Control (N=4) T'=7, p=0.3125 T +=7, p=0.3125 T+=6, p=0.4375
IL Visual (N=5) 1*=15, p=0.0313 1*:12.,=o.1563 1*=7, p=0.500

3 Spatial (N=3) 1*:3, p=0.6250 T*=3, P=0.6250 T=5, p=0.2500

 

 

 

 

113

 

APPENDIX B

Figures

114

Optic I be Bucca ganglia

 

 

‘ ’fr’

Su oeso ha eal
Optic tract Esophagus p mas: g

t" )4
I' .

Figure 1: A diagram of the brain of an octopus as seen from above showing
the basic organization of the major components of the cephalopod brain
(after Young, 1971; Sanders, 1975). Anterior end of the octopus located at

top of diagram, posterior at bottom.

115

 

”99'. 2: Saggital Section of the supraoesophageal mass of Sepia officinalis.
Regions are labelled as follows: A) vertical lobe, B) subvertical lobe, C) superior

' lobe, D) inferior frontal lobe, E) anterior basal lobe, F) medial basal lobe,
G) dorsai basal lobe. Anterior of cuttlefish=right, dorsal=up.

116

  
  
  

Inner Granule Cell Layer

Plexifonn Region

 

Outer Gr nule Cell Layer

”Gun 3: Transverse section of the deep retinal region of Sepia oflicinalis.
Showing the inner and outer granule cell regions. layered plexiform region and
medulla (Magnification=20X). Black bar indicates distance of 0.1mm.

117

Ca2+ K+

Receptor Channel Channel
- - -
' I ‘ A) Induction

d ) B) Consolidation
/
e C. .

 

 

 

 

 

 

I. 1
[p3 Ryanodin
Receptor
) C) Storage
Nucleus and
Ribosomes J

 

 

 

 

Figure 4 : Diagram illustrating the major cellular components underlying
associative memory in Hermissenda (adapted from Alkon, 1998). A) Induction
involves an influx of calcium, phosphorylation of PKC and subsequent
translocation from cytoplasm to the cell membrane, phosphorylation of Calexcitin
(CE) and translocation to the cell membrane and inhibition of voltage-gated
potassium channels. B) Consolidation, CE is translocated to the ER membrane
and nuclear membrane thus activating the ryanodine receptor and further
inhibiting potassium. C) Storage, involving re-uptake of excess calcium as well
as DNA and protein synthesis.

118

Removable 9E Exit Exit Diameter

 

 

 

 

 

Door V “M ’ 56 cm
A
_/
Start Tube

 

Figure 5: Schematic drawing of the simultaneous discrimination maze as viewed
from above. Important maze features include the start tube and two exit holes
cut through the walls of the arena, patterned panels (visual cues) that an be
switched randomly between doors, clear plexiglas doors, and a clear, plexiglas
"false bottom" fitted in the maze. An example escape route is depicted with
dashed arrows.

119

 

 

 

 

 

 

    

 

 

 

 

 

 

 

500 ' 500—
.. ,_ A - IlMiddlm
§ 40° 7 mi §400 LEiLast
E 350 4 [32 3; 350 - —m
5, 300 3 93 g 300 «
8256 El" 8250 -
8 200 4 200 _
i 150 - g 150 4
0
5 10° 7 2 100 —
5: E u a flu 5° ‘
(3' Block D 0 ‘
SOOI
4 . soc , ,
A433 . ,__ A 450.7 1 IDFirst
§ IFIrst - I [IMiddle
, 400 ,
. ‘ x 350 4 :ElLarsti .
“ _ S 3007
8 a 250 3
§ 200 3
2 1505
$8 100
o ’2
Block Block

Figure 6: Mean escape time 1 standard error over the first, middle and last block
of five trials for cuttlefish in each experimental group. A) Control group, which
only received 20 experimental trials, shows results for all four blocks of five trials,
the apparent decrease was insignificant (repeated measures ANOVA, n=8,
F=2.094(3,21), p>0.25). B) \fisual-cue group showed a statistically significant
decrease (Repeated Measures ANOVA, F=9.075 (2,14), p<0.005). C) The spatial-
cue group also showed a statistically significant decrease (Repeated Measures
ANOVA, F=8.929 (2,14), p<0.005). Finally, the combined-cue group showed a
statistically significant decrease as well (Repeated Measures ANOVA, F=14.311
(2, 22), p<0.0005).

120

 

 
 

  

 

 

 

 

   

 

 

 

 

90% 90%
I 80%
80% ""~'
-. -.- ,, 70% lFrrst
70% El First A I Middle
3% 60%< 'M‘dd'e «1360701 mail
‘I: 91-3?!” '5 50%
g 50% 5 t
a 40% ~ c 40%
g 30% g 30%
20% , 20% ~
m E @ ‘°°’° ‘
0% . - 0% _
Block
C. I).
90%
80% ‘ iFirst
70% — 4,, lMiddle
§ 60% _ QLESI.
E 50% ) 37?
g 40% ~ g
2° am- i
20% - 2
10% 1
0% «

 

 

Block

 

Figure 7: Mean overall error time :t standard error over the first, middle and last
block of five trials for cuttlefish in each experimental group. A) Control group did not
show a signifcant decrease in error (repeated measures ANOVA, n=8, F=2.000(3,21),
p>0.10). B) The visual-cue group showed a statistically significant decrease in error
(Repeated Measures ANOVA, F=11.53 (2,14). p<0.001). C) The spatial-cue group also
showed a statistically significant decrease (Repeated Measures ANOVA, F=12.631
(2,14), p<0.001). Finally, the combined-cue group showed a statistically significant
decrease as well (Repeated Measures ANOVA, F =28.774 (2, 22), p<0.0005).

121

 

 

45 — lI:Icontro|
40 _ I )Dvisual
Ilspatial
q, _ .
E 35 I I! combined
31' 30 —
O
'0’) 25 -
.Q
g 20 ,
C
§ 15 I
2 10 -
5 z
o «L L,

 

 

 

 

 

 

 

Group

Figure 8: Comparison of mean number of trials to criterion level performance
between control, visual cue-trained, spatial cue-trained and combined cue-
trained cuttlefish. The overall mean number of trials to criterion was significantly
lower in the spatial cue group than it was in either the visual cue group (One-Way
ANOVA, F=3.93, P=0.067) or the combined cue group (One-Way ANOVA,
F=3.35, P=0.079).

I22

 

Figure 9: Horizontal sections of the plexiform layer of the optic lobe of S.
ofi'icinalis. Black arrows demonstrate regions of intense staining. Black bar
indicates distance of 0.1 mm. A: Calexcitin immunopositive cells of different sizes
are evident throughout the granule cell layers, and tangential layers of the
plexiform region. B: Same region with calexcitin staining absent because the
primary antibody is replaced with goat normal serum.

123

.. ,4. My

. a.»

gnaw”, .. ,...

h...

by .

 

Horizontal

officinalis.

sections ofthe medulla oftheopticlobeofS.

Black arrows demonstrate regions of intense staining. Black bar indicates
distance of 0.1 mm. A: Calexcitin immunopositive cells are present throughout

the "cell islands” of the medulla. B: Same region with staining absent because

primary antibody replaced with goat normal serum.

Figure 10

 

Figure 11: Horizontal sections of the superior frontal lobe of S. officinalis . Black
arrows demonstrate regions of intense staining. Black bar indicates distance of
0.1 mm. A: Calexcitin immunopositive cells are present in the most posterior
portion of the lobe. B: Same region with staining absent because primary
antibody is replaced with goat normal serum.

 

Flgure12: Horizontal sections of the posterior inferior frontal lobe of S.
oflicinalis. Black arrows demonstrate regions of intense staining. Black bar
indicates distance of 0.1 mm. A: Large calexcitin immunopositive cells are
located toward the dorsal/posterior edges of the inferior frontal lobe. B: Same
region with staining absent because primary antibody is replaced with goat
normal serum.

 

Figure 13: Horizontal sections of the vertical lobe of S. officinalis. Black arrows
demonstrate regions of intense staining. Black bar indicates distance of 0.1 mm.
A: Calexcitin immunopositive amacrine cells are evident throughout the neuropil.
B: Same region with staining absent because primary antibody is replaced with
goat normal senim.

 

 

.f 1‘ ;. .1. ,5
I 24.54. . .- -~x...,a

c ,, .

Figure 14: Horizontal sections of the vertical/subvertical lobes of S. oflicinalis.
Black arrows demonstrate regions of intense staining. Black bar indicates
distance of 0.1 mm. A: Calexcitin immunopositive tracts connect the vertical and
subvertical lobes. B: Same region with staining absent because primary antibody
is replaced with goat normal serum.

 

r ' 'r 3,? 7
i. ,u -,’1 r “a > ‘
a 4.

I

 

- -.« ' .1 i ' 1'“ n’ .
Figure 15: Horizontal sections of the subvertical lobe of S. officinalis. Black
arrows demonstrate regions of intense staining. Black bar indicates distance of
0.1 mm. A: Calexcitin immunopositive cells are located throughout islands
scattered in the neuropil. B: Same region with staining absent because primary
antibody is replaced with goat normal serum.

129

 

Figure 16: Horizontal sections of the anterior basal lobe of S. ofi‘icinalis. Black
bar indicates distance of 0.1 mm. A: Calexcitin immunopositive cells of different
size (white arrows) in the cell layer of the lateral region of the lobe. B: Same
region with staining absent because primary antibody is replaced with goat
normal serum.

130

 

Figure 17: Horizontal sections of the median basal lobe of S. ofl‘icinalis. Black
arrows demonstrate regions of intense staining. Black bar indicates distance of
0.1 mm. A: Calexcitin immunopositive cells of different size throughout clusters of
cells in the neuropil. B: Same region with staining absent because primary
antibody is replaced with goat normal serum.

131

 

    

Figure 18: Horizontal sections of the posterior wall of the dorsal basal lobe of S.
ofiicinalis. Black arrows demonstrate regions of intense staining. Black bar
indicates distance of 0.1 mm. A: calexcitin immunopositive cells at the posterior
wall. B: Same region with staining absent because primary antibody is replaced
with goat normal serum.

132

0.35 ~

.0
co
J

Y=-0.005631 4» 0.003861 19X

.0 .o

—\ .o N

01 N 01
l l 1

Net Absorption (595 nm)
.0

0.05 -

 

o T l I T I
O 20 40 60 80 1 00

Protein Concentration in uglml

 

Figure 19: Protein concentration curve of BSA at different concentrations
read at 595 nanometers used to determine the protein concentration of
samples for western blotting.

133

 

 

 
 

 

b_ AB CD E F G Hl J K L

“.4 a...“ _:n.___’__._-'

  

   

C.

ABCDEFG H I JKL

 

ABCDEF G IJKL

  

 

Figure 20: Nitrocellulose membranes loaded with homogenate and
immunostained for calexcitin. Arrow indicates approximate lomtion of 23 kDa on
the membrane. Cuttlefish identities are as follows: A-erlow, B-Lola, C-Howard,
D-Snookums, E-Guy, F—Scratch, G-lsis, H—Man’lyn, l-Nougat, J-Stone, K—Freckle,
and L-Bilbo. Blank spaces are location of the molecular weight marker on the
gels. a.) left optic lobe homogenate, b.) right optic lobe homogenate, and c.)
supraoesophageal region homogenate.

134

 

 

 

-‘-g¢z~£‘_:v- ' ,,. em a ' Iv 'r' .,.~£-«.;._ ‘4- Anson?! '7 .. .. ~' mum

 

 

 

ABM CDEFGHIJKL

b.

_ "H's?“ :m.‘ which -t‘aMm.z-:mrulnfifif* ’
h
Basia: 25‘». ‘ i“? W7 E‘fi'r " -.-".5"fi-‘ Amp-mas , fig: “‘5’? 5'?" .

ABCDEF GHIJKL

 

 

 

 

C.

. . ‘3' _ J' ' - '. .I «' . ' . «. , . z . -" . ' ..-' -. ' . -. .-
. . * .'-_ a 9 .. . '. - \_' - ' ' - q'-._ r-- . ' - ' . . _ - ' ' -- .- ' ._! padre". ~-- I-.- ‘ ' .. _. ‘1 .v:-- .~ir,;.:é.'_".a'r "-
. . . ‘ .. . . ._ _ . . ,l
, L . . , , -, ,

A BCDEFGHIJKL

 

Figure 21: Nitrocellulose membranes loaded with microsomal fractions and
immunostained for calexcitin. Arrow indicates approximate location of 23 R

De on the membrane. Cuttlefish identities are as follows: A-erlow, B-Lola, C-
Howard, D-Snookums, E-Guy, F-Scratch, G-lsis, H-Marilyn, l-Nougat, J—Stone,
K-Freckle, and L-Bilbo. Blank spaces are the location of the molecular weight
marker on the gels. a.) left optic lobe cytosolic fraction, b.) right optic lobe
cytosolic fraction, and c.) supraoesophageal region cytosolic fraction.

135

 

 

a.

i t I, 2;

AWBCDEFCHl

 

    

 

 

'I
""-" '-.'_'-‘.‘_-{' ""- xi”: 0.5. -xslmflr .

.J

m‘.»..
-:--- a
. .2 J
,F'

 

aw.
A B C D E F G H I J K L

C.

_ _ ' . .' 1'. '3 . 5’ kt"’é'{UIC?{IL". .-.v _ _ —.

AB CDEFGHIJKL

 

 

Figure 22: Nitrocellulose membranes loaded with cytosolic fractions and
immunostained for calexcitin. Arrow indicates approximate lowtion of 23
kDa on the membrane. Cuttlefish identities are as follows: A-Willow, B-
Lola, C-Howard, D—Snookums, E-Guy, F-Scratch, G-lsis, H-Marilyn, l-
Nougat, J-Stone, K-Freckle, and L—Bilbo. Blank spaces are the location of
the molecular weight marker on the gels. a.) left optic lobe microsomal
fraction, b.) right optic lobe microsomal fraction, and c.) supraoesophageal
region microsomal fraction.

136

 

 

 

APPENDIX C

Reprint

137

 

Karson, MA, Boal, JG and RT Hanlon. 2003. Experimental Evidence for Spatial
Learning in Cuttlefish (Sepia ofi'icinalis). Journal of Comparative Psychology, in

press.

138

Laboratory mazes were used to study spatial-Ieaming capabilities in
cuttlefish (Sepia officinalis), using escape for reinforcement. In preliminary
observations, cuttlefish in an artificial pond moved actively around the
environment and appeared to Ieam about features of their environment In
laboratory experiments, cuttlefish exited a simple alley maze more quickly with
experience and retained the Ieamed information. Similar improvement was not
found in open-field mazes or T-mazes, perhaps because of motor problems.
Cuttlefish Ieamed to exit a maze that required them to find openings in a vertical
wall. The wall maze was modified to an arena, and simultaneous discrimination
Ieaming and reversal Ieaming were demonstrated. These experiments indicate
that cuttlefish improve performance over serial reversals of a simultaneous,
visual—spatial discrimination problem.

Cephalopod mollusks have complex nervous systems and highly diverse
behaviors. Sensitization, habituation, associative learning, and spatial Ieaming
have all been demonstrated (for reviews, see Bittennan, 1975; Hanlon &
Messenger, 1996; Mather, 1995; G. D. Sanders, 1975). Because cephalopods’
behavioral abilities have been compared with those of lower vertebrates and
because they are evolutionarily distant from species more commonly used in
Ieaming experiments, cephalopods are worthwhile test species for functional
explanations of the evolution of complex nervous systems (Budelmann, Bullock,

& VVIlliamson, 1997; Packard, 1972).

139

Many octopus species forage away from a home den that they return to
repeatedly for shelter (e.g.. Ambrose, 1982; Boyle, 1983, 1988; Forsythe &
Hanlon, 1997; Hartwick, Ambrose, & Robinson, 1984; Mather 1991). Studies of
detours and maze Ieaming have supported field observations suggesting spatial
learning (Boal, Dunham, Williams, & Hanlon, 2000; Mather, 1991; Moriyama &
Gunji, 1997; Wells, 1964). Other complex learning behaviors observed in
octopuses include avoidance Ieaming (Boycott, 1954), discrimination Ieaming
(Sutherland & Muntz, 1959), and reversal Ieaming Mackintosh & Mackintosh,
1964). Octopuses learn relatively quickly and retain the Ieamed information. It is
clear that Ieaming is an important aspect of the natural history of octopuses.

Neurobiological evidence suggests that cuttlefish, like octopuses, could be
capable of complex Ieaming. Cuttlefish have a larger overall brain to body size
ratio than do octopuses (Maddock & Young, 1987). Furthermore, the vertical
lobe in the cuttlefish, a brain area thought to be involved in Ieaming and memory
(Young 1960 and 1965), occupies roughly 24% of the total brain volume,
whereas the vertical lobe of the octopus occupies about 13% of the total brain
volume (Maddock & Young, 1987; Wrrz, 1959). In cuttlefish, an improvement in
Ieaming has been correlated with increases in the volume of the vertical lobe
during post-embryological development (Dickel, Chichery, & Chichery, 1998;
Messenger, 1973). This finding suggests that vertical lobe size could be
associated with Ieaming. Thus, learning and memory could play an important

role in cuttlefish natural history.

140

 

Cuttlefish are clearly capable of associative learning. Through negative
reinforcement, cuttlefish can be trained to not strike at prey items (e.g., Agin, et
al,, 1998; Chichery & Chichery, 1992; Dickel et al., 1998; Messenger, 1968 and
1977).

There are no published studies of spatial learning in cuttlefish; however,
the natural history of cuttlefish provides little evidence for good spatial-Ieaming
ability (Boletsky, 1983; Hanlon & Messenger, 1988). Unlike octopuses, cuttlefish
do not take shelter in home dens but instead rely primarily on crypsis for defense
(Boletsky, 1983). Seasonal onshore—offshore migrations have been documented
in cuttlefish (Boletsky, 1983), and there is evidence that Australian cuttlefish
return seasonally to a specific breeding site (Hall & Hanlon, 2002). However,
because of their short lifespan (1-2 years) and typical semelparity, it is likely that
learning is not involved in these migrations. Cuttlefish in nature frequently swim
in and around vertical barriers, and tagging experiments indicate that Australian
cuttlefish forage away from and then return to particular rocks (O’Dor, personal
communication, 2002). Spatial Ieaming may be useful in negotiating these
obstacles (F. K. Sanders & Young, 1940). There is still much to be discovered
about the role of spatial Ieaming in the natural history of cuttlefish.

Preliminary observations of 7 cuttlefish placed individually in a large,
outdoor artificial pond (13.4 m by 13.7 m; see Figure 1A) indicated that cuttlefish
moved around a new environment in a manner consistent with an interpretation
of exploration. There was a barrier within the experimental pond with a small

opening that allowed the cuttlefish to travel from one side of the pond to the other

141

 

(see Figure 1A). Cuttlefish appeared to remember the location of this hole and
returned to the site of the hole when it was blocked off (the open hole was
approached an average of 14.3 times/hr; the closed hole was approached an
average of 25.8 times/hr). The cuttlefish’s repeated use of the hole in the barrier
and frequent return to this hole suggest that the cuttlefish Ieamed about features
of the artificial pond and retained the information from one day to another. Thus,
it seemed that cuttlefish could be used in laboratory experiments designed to
evaluate spatial Ieaming.

We attempted to provide cuttlefish with maze problems that were
comparable with a natural spatial-Ieaming problem, using escape as motivation.
In Experiment 1, we asked whether cuttlefish would learn the simple task of
exiting a straight alley when escape was the sole motivation. Experiments 2, 3,
and 4 were initial attempts to design appropriate mazes that allowed for choice
and to assess spatial Ieaming in cuttlefish. Finally, in Experiment 5, we asked
whether cuttlefish would show improvement over serial reversals of a two-choice,
spatial—visual discrimination problem.

General Method

Experiments were conducted at the Marine Biomedical Institute (MBl) of
the University of Texas Medical Branch, Galveston, Texas, and the Marine
Resources Center (MRC) of the Marine Biological Laboratory, Woods Hole,
Massachusetts. At the MBl, all experimental tanks were interconnected on the
same 13,000-L recirculating sea water system, dedicated to holding cephalopods

and their live food. Water was a mixture of natural seawater from the Gulf of

142

Mexico and artificial seawater made from Instant Ocean brand salts (Aquarium
Systems, Mentor, OH). Salinity ranged from 31 parts per thousand (ppt) to 35
ppt, and water temperature ranged from 16 °C to 18 °C. In this closed system,
water exiting each tank passed through mechanical, chemical, and biological
filters and was treated with ultraviolet light to kill pathogens. Water flow was
continuous at all times, including during experimental trials. At the MRC, the
water supply was drawn from a depth of approximately 3 meters below the
surface of Great Harbor and gravity fed to tanks throughout the MRC building.
Salinity ranged from 31 ppt to 33 ppt, and the water temperature ranged from 15
°C to 21 °C.

All subjects were laboratory-cultured cuttlefish (Sepia officinalis) from the
National Resource Center for Cephalopods (MBI, Galveston, TX). Cuttlefish
were housed in small groups (4—7 individuals) in large experimental tanks (MBI,
122 cm long by 183 cm wide by 80 cm deep; MRC, 366 cm diameter by 90 cm
deep). Gravel, large pieces of polyvinyl chloride pipe, and artificial plants were
placed in each tank. Unless otherwise specified, experiments were conducted in
the home tanks. The cuttlefish were fed a mixture of live and frozen fish, shrimp,
and crabs twice a day in the morning and early evening. A complete description
of cuttlefish mariculture can be found elsewhere (e.g., DeRusha, Forsythe,
DiMarco, & Hanlon, 1989; Hanley et al., 1998).

Data were log transformed where noted and analyzed using parametric
and nonparametric statistics as specified (Siegel & Castellan, 1988).

Experiment 1

Method

This experiment tested the feasibility of using a simple escape maze with
cuttlefish. Seven sub-adult cuttlefish (9—15-cm mantle length [ML]) were used.
An alley (150 cm long by 20 cm wide by 20 cm deep) was constructed with
wooden sides and a clear, Plexiglas bottom (see Figure 1A). The alley was
suspended over and submerged in a home tank, such that the water depth in the
alley was 10 cm. The last 20 cm of the alley had no bottom so the cuttlefish
could escape by swimming down into the tank below.

Cuttlefish were placed at the closed end of the alley and allowed to swim
to the open end to escape into the home tank below. Cuttlefish are highly visual
and like to settle (and bury) on the bottom in a large, open space. They did not
settle on the clear bottom of the alley but swam until they found their way out.

There were four stages to the experiment. In Trials 1—15, the open end of
the maze faced north. In Trials 16—20, the open end faced south. ln Trials
21—25, the open end again faced north. During Trials 26-30, the open end was
reversed on each trial.

Results

The performances of all cuttlefish are plotted in Figure 2. The cuttlefish
showed significant overall reduction in escape time from a mean of 123 s in Trial
1 to a mean of 20 s in Trial 30: repeated measures analysis of variance
(ANOVA), six blocks of five trials, F(5, 30) = 6.74, p <0.001. Differences between

individuals were not significant: repeated measures ANOVA, six blocks of five

144

 

 

trials, F(5, 30) = 1.10, p > 0.25. There was no detectable effect on maze-escape
performances from reversing the maze orientation.

Discussion

The cuttlefish quickly learned the simple task of exiting a straight alley with
escape as the only motivation. Resufts were similar to those obtained by Walker,
Longo, and Bittennan (1970) using octopuses in an alley with a food reward at
the end of the maze. Whereas octopuses required extensive shaping before
readily pursuing the food, cuttlefish required no pre-training in this escape maze.
The increase in escape time observed between Trials 21 and 25 was probably
due to a water quality problem resulting from a blower failure.

Overall, this experiment suggests that learning in cuttlefish can be
evaluated using a simple escape maze. Additionally, the lack of effect of maze
reversal (Trial 25) suggests that the cuttlefish did not rely on visual features
around the laboratory to solve this simple maze problem.

Experiment 2

Method

An open-field maze was constructed of a round tank (100 cmdiameter by
7 cm deep) with a clear Plexiglas bottom and opaque, blue sides (see Figure
1B). A hole (20 cm diameter) was cut into the bottom of the maze 11 cm from
one side. This hole allowed the cuttlefish to swim out of the bottom of the maze
into the home tank below. For Group 1, a landmark was provided (6—cm by 6-cm
by 2-cm piece of Styrofoam tethered and floating directly above the maze). No

specific landmarks were provided for Group 2.

145

 

Individual sub-adult cuttlefish (N = 18; 7-9—cm ML) were placed in the

maze and allowed to swim until escaping into the home tank (no time limit).

 

Trials were repeated twice daily for 20 days. For the first 25 trials, training
proceeded as described above. For Trials 26—30, the location of the opening
was rotated 180°. In Trials 31-40, the location of the maze exit was assigned
randomly to one of the four cardinal directions, with the constraint that each

direction was presented at least twice.

 

Results

No significant improvement in maze escape time was found: repeated
measures ANOVA, F(4, 52) =2.35, p = 0.67.

Discussion

In this experiment, problems arose from the peculiar way that cuttlefish
swim (see Figure 3). Cuttlefish swim slowly forward and backward by using their
fins and swim quickly backward by jetting with their funnels. Fine directional
control is not possible when jetting, and jetting cuttlefish move haphazardly
around the mazes. A number of forward-swimming cuttlefish turned around once
they found the maze exit and attempted to jet backward through the exit hole. At
times, they missed the hole and had to turn around and again search for the exit.
Other times, the cuttlefish would find the escape hole on a forward approach, but
on contacting the hole, they became excited or agitated and jetted backwards.
Some of the cuttlefish repeatedly forward approached, touched, and jetted
backward, appearing “frustrated” with their inability to get through the maze.

These various motor constraints probably interfered with maze learning.

I46

Experiment 3

Method

A T maze (100 cm long by 20 cm wide by 10 cm deep) was constructed
with black Plexiglas sides and a clear Plexiglas bottom (see Figure 1C). Exits
(20 cm by 20 cm) were cut through the bottom at the end of each arm of the T
and fitted with removable clear Plexiglas doors. These holes allowed the
cuttlefish to swim down into the home tank below. White panels on the walls of
the right arm of the T and green panels on the left arm were provided as
directional cues (i.e., bright vs. dark).

Sub-adult cuttlefish (N = 15; 8—13cm ML) were first pre-trained with 12
trials in an alley maze (the central alley of the T maze; see also Experiment 1).
After runway training, the cuttlefish were then pre-trained with 12 trials with the
open arm of the maze varied (forced turn; six right and six left turns in semi-
random order; Fellows, 1967). The cuttlefish were then divided into three groups
and given 24 training trials as follows: For Group 1, the exit was on the right
(white) arm of the T maze. For Group 2, the exit was on the left (green) arm of
the T maze. For Group 3, the exit was located on either the right (white) or left
(green) arm of the maze, varied semirandomly from trial to trial.

Trials were conducted twice a day with a 6-hr inter-trial interval.
Immediately after training, learning was tested in two ways. First, cuttlefish
received one trial with both exits open (Test 1). This tested the possibility that
cuttlefish could detect the location of the open exit using an unintended cue (e.g.,

water current). Second, the cuttlefish received one trial with the exits to both

147

maze arms open and the wall panels reversed (right arm green, left arm white;
Test 2). This tested the relative importance of turning direction versus wall panel
cues.

Results

No improvement in exit time was found: arcsine transformed proportions,
repeated measure ANOVA, for group, F(2, 11) = 0.55, p > 0.50; for trial, F(3, 33)
= 0.28, p >0 .80; for interaction, F(6, 33) = 1.33, p > 0.25. Individual
performances were examined for evidence of Ieaming. Three cuttlefish from
Group 2 that chose the left—green arm in their initial choice and in Test 1 chose
the right-green arm in Test 2.

Discussion

Motor constraints may have again interfered with learning. The cuttlefish
often swam down the runway portion of the maze, but when they got to the top of
the T, they often rested on the bottom and did not move much thereafter. At this
point, they needed to make a 90° turn right or left in a small space, interrupting
momentum toward finding the exit in the maze. A Y-shape maze could work
better.

The 3 cuttlefish that chose the green arm in both Test 1 and Test 2 may
have chosen the maze arm on the basis of visual landmarks rather than
directional cues. However, this result is uncertain because no evidence for
Ieaming was found.

Experiment 4

Method

148

 

 

 

Observations from the pond suggested that cuttlefish could learn to use
openings in vertical barriers. Consequently, a wall maze was constructed by
placing a vertical, opaque Plexiglas wall 50 cm from the short end of a
rectangular experimental tank, creating a small testing arena and a larger home
tank (see Figure 1D). The side of the wall facing the testing arena was covered
with material: artificial sea grass on the left and camouflage mesh on the right.
Two holes (20 cm diameter) were cut into the wall: one on the left and one on the
right. The hole on the left side was 10 cm above the bottom of the tank, and the
hole on the right side was 60 cm above the bottom of the tank. Both exit holes
remained open throughout the experiment. A piece of opaque, plastic sheeting
the exact width of the tank was angled at 45° from the base of the wall to the
water surface on the opposite side of the testing arena (see Figure 10). Thus,
within the testing arena, there was no horizontal surface on which the cuttlefish
could settle, providing increased motivation for escape. A gravel substrate was
provided in the home tank.

Individual subadult cuttlefish (N = 18; 9—13-cm ML) were placed in the
testing arena midway between the openings and facing the back wall. The time
to exit the arena (maximum 10 min) and the exit used were recorded. If a
cuttlefish did not escape, the escape time was recorded as 1,000 5.

Results

Cuttlefish demonstrated a significant decrease in exit time within 10 trials:
repeated measures ANOVA, F (9, 135) = 7.47, p < .001 (see Figure 4). The

mean escape time in the 1st trial was 15.6 min (only 3 of 18 cuttlefish escaped),

149

 

 

whereas the mean escape time in the 10th trial was 4.0 min (15 of 16 escaped).
This was not a result of an improvement in maze escape time (mean first escape
time = 154.0 3, SEM = 33.2; mean last escape time = 154.0 s, SEM =40.3).
Thirteen of 18 cuttlefish preferred the left—lower hole, and cuttlefish exited
through their preferred hole an average of 78.00% of the time (range =
50.00%-100.00%, SD = 16.17).

Discussion

Cuttlefish learned to escape from this two-choice maze. They did not
escape more quickly; time to escape the small arena was simply the time it took
the cuttlefish to turn around and swim out. Initially, those that did not exit swam
rapidly around the testing arena, repeatedly approaching the corners. At the
end, those that did not escape wedged themselves between the plastic sheeting
and the wall and remained still for the full 10 min.

Results from this experiment indicate that the problem with the cpen field
and T-mazes lies not with the cuttlefish Ieaming abilities but instead with the
maze design and cuttlefish motor behavior. In both the wall maze and the alley
maze, cuttlefish exited while swimming slowly forward, and thus, motor
constraints did not interfere with learning.

Cuttlefish are generally benthic; thus, the greater preference for the
left-lower hole may have resulted because cuttlefish remained near the bottom
of the testing arena and the lower hole was easier to locate.

Experiment 5

Method

150

 

The purpose of this experiment was to further evaluate Ieaming of a
simultaneous, visual—spatial discrimination task. Subsequently, reversal Ieaming
of this problem was tested. Twenty cuttlefish were used in this experiment: 13
adults at the MRC (15—25-cm ML) and 8 sub-adults at the MBI (10—15-cm ML).
A circular testing arena was constructed from a large, plastic barrel (71 cm deep_
56 cm diameter; see Figure 1E). A start tube (16 cm diameter) with doors on
both ends was placed through one side of the arena, 17 cm below the top edge
(12 cm below the water surface). Two exit holes (16 cm diameter) were cut on
opposite sides of the arena, 6 cm below and perpendicular to the start tube. A
panel of striped fabric (36 cm by 36 cm) surrounded the exit located on the right
side of the start tube, and a spotted panel of fabric (36 cm by 36 cm) surrounded
the exit located on the left side of the start tube. Exits were fitted with movable,
clear Plexiglas doors. The bottom of the arena was fitted with a plastic mesh
cone (point up, 12 cm high) such that, like the wall maze, there was no horizontal
surface on which the cuttlefish could settle. The maze was placed within the
home tank at the beginning of each set of trials.

In each trial, 1 cuttlefish was herded into the start tube. After 1 min, the
door to the testing arena was opened. The cuttlefish was given 3 min to exit the
tube and enter the arena. If the cuttlefish did not enter the arena, the
experimenter chased it into the maze using a net. Once the cuttlefish was inside
the arena, the start tube was blocked off. The cuttlefish was given 7 min to
escape the arena before the experimenter chased it out of the open exit with a

net. The escape time for each trial was recorded. Any trial in which the cuttlefish

151

 

failed to escape was assigned an escape time of 10 min. In each trial, the maze
was positioned in a random direction within the home tank to control for the
possibility that cuttlefish used cues around the laboratory to locate the open exit.

There were four stages to this experiment: pre-training, preference testing,
training, and reversal. Individual cuttlefish (N = 21) were pre-trained with both exit
doors open. Trials were repeated until the cuttlefish independently exited the
maze once through each exit. After pre-training, exit preference was determined
by observing the exit most frequently used in five further escapes. After pre-
training and preference testing, experimental trials began. In the first set of
training trials (Reversal 0), the cuttlefish’s preferred exit remained closed. Each
cuttlefish received six trials per day (at least 45-min intertrial interval) until six of
seven consecutive escapes were achieved in less than 1 min. Once this criterion
had been reached, the cuttlefish were given a probe trial with both exits open.
This probe tested the possibility that cuttlefish could detect the location of the
open exit using an unintended cue (e.g., water current). If the cuttlefish did not
escape through the trained exit, they were given two more training trials and then
another probe trial; this occurred only three times, in 3 different cuttlefish. Once
the cuttlefish was trained, the open door was closed, the opposite door was
opened, and the training procedure repeated. Training and door reversal
continued as time permitted.

Results

All cuttlefish completed pre-training successfully (range = 5—16 trials).

There were no notable performance differences between cuttlefish tested at MBL

152

 

 

and MBI or between female and male cuttlefish. Nine cuttlefish preferred the
striped (right) exit and 12 cuttlefish preferred the spotted (left) exit. Fourteen
cuttlefish had strong preferences (4 or 5 out of 5 trials). These strong
preferences were split evenly between the two exits (8 cuttlefish preferred the
striped exit, and 6 preferred the spotted exit).

For the 20 cuttlefish that completed Reversal 0 (training against
preference), the mean number of trials was 36 (range = 13-72), the mean
escape time was 5 min 8 3 (range = 2 min 50 s—6 min 24 s), and the mean
percentage of escapes in less than 1 min was 36.2% (range = 13.9%—64.7%).
One cuttlefish died during the course of training for Reversal 0.

There was a marked improvement in performance for cuttlefish that
completed subsequent reversals. For cuttlefish completing Reversal 2, the mean
number of trials during Reversal 2 was 17 (range = 6—28), the mean escape time
was 2 min 48 8 (range = 1 min 4 s-4 min 6 s), and the mean percentage of
escapes in less than 1 min was 61.9% (range = 33.3%—76.9%).

Four cuttlefish completed six reversals and showed a significant
improvement in maze performance. Improvement was indicated by a significant
decrease in the number of trials per reversal, a significant decrease in the
percentage of errors per reversal, and a significant increase in the number of
escapes in less than 1 min (see Table 1 and Figures 5 and 6). The apparent
decrease in average escape time across all reversals was not significant (see

Table 1).

153

 

 

 

For further analyses, performances on Reversals 0, 2, 4, and 6 (against
original preference) were considered separately from performances on Reversals
1, 3, and 5 (consistent with original preference). For all variables, there was
significant improvement when trained against original preference but not when
trained with original preference (see Table 1).

Two of the cuttlefish that completed Reversal 6 showed performances
suggestive of one- to two-trial Ieaming in later reversals: One cuttlefish
completed Reversals 3 and 6 in just eight trials per reversal, and the other
cuttlefish completed Reversals 5, 6, and 9 in eight trials per reversal. Because
criterion level performance was six of seven consecutive escapes in less than 1
min, the absolute minimum possible number of trials in a reversal was seven.

Discussion

Previous authors have established that cuttlefish can learn a simultaneous
discrimination task (Messenger, 1977). This experiment is the first to establish
that cuttlefish improve over serial reversals of a simultaneous discrimination
problem. Results are consistent with those found in octopuses (Mackintosh &
Mackintosh, 1964; Young, 1962) but not successively (Mackintosh, 1962).
Improvement over serial reversals may indicate that cuttlefish are “Ieaming to
learn” (Harlow, 1949). This ability allows animals to update or releam solutions
to problems when previous solutions are no longer relevant.

Significant improvement in escape performance was observed across
reversals in which the cuttlefish were trained against their original door

preference; only minor (statistically insignificant) improvement was observed

154

 

across reversals in which the cuttlefish were trained consistent with their original
door preference. This is similar to findings from octopus discrimination-Ieaming
experiments (Boal, 1996).

The performances of 2 cuttlefish were indicative of two-trial learning.
Mackintosh and Mackintosh (1964) also observed an octopus that showed one-
trial learning. One-trial Ieaming has been attributed only to vertebrates (see
reviews in Bitterman, 1989; Mackintosh, Wilson, & Boakes, 1985). The few
published studies on reversal Ieaming in marine invertebrates (e.g., crabs: Datta,
Milstein, & Bittennan, 1960; isopods: Thompson, 1957) did not show one-trial
learning, but anecdotes with octopuses suggest that very rapid Ieaming of some
tasks should be investigated more thoroughly.

It remains unclear which cue the cuttlefish used to locate the open exit:
direction relative to the start box or panel pattern surrounding the exit. Further
experimentation to distinguish between these alternatives is currently in
progress.

General Discussion

This is the first set of experiments to examine spatial and reversal Ieaming
in cuttlefish. Specifically, this set of experiments demonstrated that cuttlefish (a)
can solve an experimental maze problem with escape as the sole motivation, (b)
have particular motor needs that can confound the results of spatial-learning
experiments, and (c) show spatial Ieaming and improvement across serial

reversals when these confounders are overcome.

155

 

 

It is unclear how cuttlefish may use such Ieamed spatial information in the
natural environment. Unlike many octopuses, which probably use learning to
relocate home shelters, cuttlefish rely primarily on crypsis for defense. Cuttlefish
constantly move in and around vertical obstacles in the natural environment, and
spatial Ieaming may help them negotiate these obstacles (Sanders & Young,
1940). Alternatively, perhaps cuttlefish use spatial Ieaming to relocate good
foraging patches or return to a prominent landmark between forages (O’Dor,
personal communication, 2002). Unfortunately, we have no field data that _
address these particular hypotheses.

Spatial learning is widespread among animals. Spatial-Ieaming studies
have focused on mammals and birds but have also been successful in reptiles,
amphibians, and fish as well as in invertebrate groups such as arthropods and
octopuses (for reviews, see Capaldi, Robinson, & Fahrbach, 1999; Golledge,
1999; Sherry, 1998; Wehner, 1981). Spatial problems typically encountered by
different animals, such as returning to home or finding food sources, are
remarkably constant across different species. Locating food resources, shelter,
and escape routes are each facilitated by spatial Ieaming. It remains unclear
which aspects of cephalopod life history might have caused their great cognitive
divergence from other mollusks. These mazes could provide a tool for future

studies addressing this question more precisely.

156

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 1. Illustrations of the mazes tested with cuttlefish as seen from above
(except for E). Starting points are indicated with an S, and exit locations are
shaded with diagonal lines. A: Artificial pond used for preliminary observations.
B: Alley maze used in Experiment 1. C: Open-field maze used in Experiment 2.
D: T maze used in Experiment 3. E: Wall maze used in Experiment 4 (left: side
view) with detailed view of the wall and two exit doors (right). F: Two-choice

discrimination maze used in Experiment 5.

157

 

1601

 

 

 

 

 

 

 

1401
@1201"
Q) 2
g 100
ll— 1
g 801 v .—-<
8 6011
m
“J 40+ ..
20- ' i
0 +T 1 T 111j 1 1111 11f11 1 5111 11111111
13 5 7 911131517192123252729
Trial

Figure 2. Mean escape time for cuttlefish in the runway maze (Experiment 1; N =
7). The maze orientation was reversed at Trial 16, Trial 21, and Trials 26—30.

Error bars represent standard errors

158

 

 

 

 

 

Figure 3. Cuttlefish use fins for slow, forward movement (small arrow) and jet

propulsion for rapid, backward swimming (large arrow). Adapted from Tompsett,

1939.

159

 

 

O
—+
__
———+—
-—-—4)——
as.
.21.
“I”

12 3 4 5 67 8 910

Figure 4. Mean escape time for cuttlefish in the wall maze (Experiment 4; N =
18). Cuttlefish demonstrated a significant decrease in exit time within eight trials.

Error bars represent standard errors.

160

.h
C
1

l

09
O
1

l

N
O
1

Number of Trials

 

I

 

 

 

R0 R1

  

 

I Against Preference
[I With Preference

 

 

 

f-—-—l

   

 

 

 

 

 

 

I 1

R2 R3 R4 R5 R6 R7 R8

 

 

(N =2o)(N =17)(N =14XN =12)(N =6) (N =5) (N =4) (N =1) (N =1)

Reversal

Figure 5. Mean number of trials until reversal (criterion-level performance;

Experiment 5; N = 20). Cuttlefish showed a significant decrease in the number of

trials per reversal when trained against original preference (solid bars) but not

when trained in the direction consistent with original preference (open bars).

Error bars represent standard errors. R = Reversal.

l6]

 

 

50% -
45% -
40% d
35% 4 I Against Preference
30% . [3 With Preference

25% -l
20% o
15% -
10% s
5% s
0%

   
   

 

 

 

 

Nonescapes

 

 

 

 

 

 

 

R0 R1 R2 R3 R4 R5 R6 R7 R8
(N =20)(N =17)(N =14)(N =12)(N =6) (N =5) (N =4) (N =1) (N =1)
Reversal

Figure 6. Mean percentage of errors (nonescapes) per reversal (Experiment 5; N
= 20). Cuttlefish showed a significant decrease in the percentage of errors per
reversal when trained against original preference (solid bars) but not when
trained in the direction consistent with original preference (open bars). Error bars

represent standard errors. R = Reversal.

162

Table 1. Results of analyses of the effects of serial reversals on Ieaming (Page
Test for Ordered Alternatives). Significant results indicated in boldface.

 

 

 

 

 

 

 

 

Variable N k L p
Number of trials per reversal
-overall 4 7 505 <0.01
-against preference 4 4 115 <0.01
-with preference 4 3 50 >0.05
Percentage of error
-overall 4 7 521 <0.001
-against preference 4 4 114 =0.01
-with preference 4 3 53 >0.05
Number of escapes < 1 min
-overall 4 7 497 <0.025
-against preference 4 4 112 <0.05
-with preference 4 3 46 >0.05
Mean escape time
-overall 4 7 480 >010

 

163

 

 

LITERATURE CITED

164

 

Agin, V., R. Chichery and M.-P. Chichery (2001). "Effects of Ieaming on
cytochrome oxidase activity in cuttlefish brain." Neuroreport 12: 113-116.

Agin, V., L. Dickel, R. Chichery and M.-P. Chichery (1998). "Evidence for a
specific short-term memory in the cuttlefish Sepia." Behavioural Processes 43:
329-334.

Alkon, D. L. (1974). "Associative training of Hermissenda." Journal of General
Physiology 64: 70-84.

Alkon, D. L. (1989). "Learning in a marine snail." Scientific American 249: 70-84.
Alkon, D. L. (1999). "Molecular principles of associative memory that are
conserved during the evolution of species." Malacalogia 41: 321-329.

Alkon, D. L., H. Ikeno, J. Dworkin, D. L. Mcphie, J. L. Olds, I. Lederhendler, L.
Metzel, B. Schreurs, A. Kuzirian and C. Collin (1990). "Focusing of neuronal
branches: An automatic correlate of associative memory." Proceedings of the
National Academy of Sciences of the USA 87: 1611-1614.

Alkon, D. L., I. I. Lederhendler and J. J. Shoukimas (1982). "Primary changes of
membrane currents during retention of associative learning." Science 215: 693-
695.

Alkon, D. L. and T. J. Nelson (1990). "Specificity of molecular changes in
neurons involved in memory storage." Federation of American Societies for
Experimental Biology Journal 4: 1567-1576.

Alkon, D. L., T. J. Nelson, W. Zhao and S. Cavallaro (1998). "Molecular cascades
and associative memory." Trends in Neurosciences 21: 529-537.

Alkon, D. L. and H. Rasmussen (1988). "A spatial-temporal model of cell
activation." Science 239: 998—1005.

Alkon, D. L., M. Sakakibara, R. Forrnan, J. Harrigan, I. Lederhendler and J.
Farley (1985). "Reduction of two voltage-dependent K+ currents mediates
retention of a learned association." Behavioral Neural Biology 44: 278-300.

Allen, A., J. Michels and J. 2. Young (1985). "Memory and visual discrimination
by squids." Marine Behaviour and Physiology 12: 81-97.

Ambrose, R. F. (1982). "Shelter utilization by the molluscan cephalopod Octopus
bimaculatus." Marine Ecology-Progress Series 7: 67-73.

Aronson, R. B. (1991). "Ecology, paleobiology and evolutionary constraint in the
octopus." Bulletin of Marine Science 49: 245-255.

165

Ascoli, G. A., K. Z. Luu, J. L. Olds, T. J. Nelson, P. A. Gusev, C. Bertucci, E.
Bramanti, A. Raffaelli, P. Salvadori and D. L. Alkon (1997). "Secondary structure
and Ca2+-induced conformational change of calexcitin, a Ieaming-associated
protein." Journal of Biological Chemistry 272: 24771-24779.

Bank, 3., A. DeWeer, A. M. Kuzirian, Rasmussen and D. L. Alkon (1988).
"Classical conditioning induces Iong-terrn translocation of protein kinase C in
rabbit hippocampal cells." Proceedings of the National Academy of Sciences 85:
1988-1 992.

Bert, P. (1867). "Mémoire sur la physiologie de la seiche (Sepia ofiicinalis, Linn)"
Mémoires de la Société des Sciences Physques et Naturelles de Bourdeaux 5:
1 1 5-138.

Bitterman, M. E. (1975). Critical Commentary. Invertebrate Learning. J. D. a. A.
W. WC Corning. New York, Plenum Press. 3: Cephalopods and Echinoderms:
139-245.

Boal, J. G. (1996). "A review of simultaneous visual discrimination as a method of
training octopuses." Biological Reviews of the Cambridge Philosophical Society
71: 157-190.

Boal, J. G., A. W. Dunham, K. T. Williams and R. T. Hanlon (2000).
"Experimental evidence for spatial Ieaming in octopuses (Octopus
bimaculoides)." Journal of Comparative Psychology 114: 246-252.

Boal, J. G., R. A. Hylton, S. A. Gonzalez and R. T. Hanlon (1999). "Effects of
crowding on the social behavior of cuttlefish (Sepia ofiicinalis)." Contemporary
Topics 38: 49-55.

Boletsky, S. v. (1977). "Post-hatchling behaviour and mode of life in
cephalopods." Symposiat of the Zoological Society of London 38: 557-567.

Boletsky, S. v. (1987). Juvenile Behaviour. Cephalopod Life Cycles. P. Boyle.
London, Academic Press. II. Comparative Reviews: 4560.

Borley, K. A., H. T. Epstein and A. M. Kuzirian (2002). "Effects of sensory block
on calexcitin levels in the photoreceptors of Hermissenda crassicomis."
Biological Bulletin 203: 197-198.

Boucaud-Camou, E. and B. J. (1991). The migrations of the cuttlefish (Sepia
officinalis L) in the English Channel. The Cuttlefish. First lntemational
Symposium on the Cuttlefish Sepia. E. Boucaud-Camou. Caen, Centre de
Publications de l'Université de Caen: 179-189.

166

 

 

Boycott, B. B. (1954). "Learning in Octopus vulgaris and other cephalopods."
Publicazioni della Stazione Zoologica di Napoli 25: 67-93.

Boycott, B. B. (1961). "The functional organization of the brain of the cuttlefish
Sepia ofiicinalis." Proceedings of the Royal Society of London B 152: 503-534.

Boycott, B. B. and J. Z. Young (1950). "The comparative study of learning."
Symposia of the Society of Experimental Biology 4: 432-453.

Boycott, B. B. and J. Z. Young (1955). "A memory system in Octopus vulgaris
Lamark." Proceedings of the Royal Society of London B 143: 449-480.

Boycott, B. B. and J. 2. Young (1956). The subpedunculate body and nerve and
other organs associated with the optic tract of cephalopods. Bertil Hanstrtim:
Zoological papers in honour of his 65th birthday, Nov. 20 1956. K. G.
Vanstrand. Lund, Zoological Institute.

Boycott, B. B. and J. Z. Young (1957). "Effects of interference with the vertical
lobe on visual discriminations in Octopus vulgaris Lamarck." Proceedings of the
Royal Society of London B 146: 439-459.

Boyle, PR (1986). Neural control of cephalopod behavior. The Mollusca. A. O. D.
VWIlows. Orlando, Academic Press. 9, Neurobiology and Behavior: 1-99.

Boyle, P. R., Ed. (1983). Cephalopod Life Cycles. Vol.1. Species Accounts.
London, Academic Press.

Boyle, P. R., Ed. (1988). Cephalopod Life Cycles. Vol. 2. Comparative Reviews.
London, Academic Press.

Bradford, M. M. (1976). "A rapid and sensitive method for the quantification of
microgram quantities of protein utilizing the principle of protein-dye binding."
Analytical Biochemistry 72: 248-254.

Budelmann, B. U. (1994). "Cephalopod sense organs, nerves and the brain:
adaptations for high performance and life style." Marine and Freshwater
Behaviour and Physiology 25: 13-33.

Budelmann, B. U., T. H. Bullock and R. Williamson (1997). Cephalopod brains:
Promising preparations for brain physiology. Cephalopod Neurobiology. R. W. a.
L. M. NJ Abbott. Oxford, Oxford University Press: 399-413.

Budelmann, B. U., U. Riese and H. Bleckmann (1991). Structure, function,
biological significance of the cuttlefish 'lateral lines'. The Cuttlefish. First
International Symposium on the Cuttlefish Sepia. E. Boucaud-Camou. Caen,
Centre de Publcations de l'Université de Caen: 201-209.

167

 

 

Bullock, T. H. (1965). Mollusca:Cephalopoda. Structure and Function in the
Nervous System of Invertebrates. T. B. a. G. Horridge. San Fransisco, W.H.
Freeman and Company: 1433-1515.

Byme, R. A., M. Kuba and U. Griebel (2002). "Lateral asymmetry of eye use in
Octopus vulgaris." Animal Behaviour 64: 461-468.

Cajal, S. R. (1917). "Contribucion al conocimiento de la retina y centros opticos
de los cefalopodos." Trabajos del Laboratorie de lnvestigaciones Biologicas de la
Universidad de Madrid 15: 1-82.

Capaldi, E. A., G. E. Robinson and S. E. Fahrbach (1999). "Neuroethology of
spatial Ieaming: The birds and the bees." Annual Reviews in Psychology 50: 651-
682.

Carew, T. J., E. T. Walters and E. R. Kandel (1981). "Classical conditioning in a
simple withdrawal reflex in Aplysia califomica." Journal of Neuroscience 1(12):
1426-1437.

Cavallaro, S., N. Meiri, C.-L. Yi, S. Musco, W. Ma, J. Goldberg and D. L. Alkon
(1997). "Late memory-related genes in the hippocampus revealed by RNA
fingerprinting." Proceedings of the National Academy of Sciences (USA) 94:
9669-9673.

Chase, R. and M. J. Wells (1986). "Chemotactic behaviour in Octopus." Journal
of Comparative Physiology A 158: 375-381.

Chichery, R. and M. P. Chichery (1992). "Learning performances and aging in
cuttlefish." Experimental Gerontology 27: 233-239.

Corner, B. D. and M. H.T. (1980). "Field observations on the reproductive
behavior of Sepia latimanus." Micronesica 16: 235-260.

Crow, T. J. and D. L. Alkon (1978). "Retention of an associative behavioral
change in Hermissenda." Science 201: 1239-1241.

Datta, L. G., S. Milstein and M. E. Bitterman (1960). "Habit reversal of the crab."
Journal of Comparative and Physiological Psychology 53: 275-278.

Day, L. B., D. Crews and W. Wilczynski (1998). "Spatial and reversal Ieaming in

congeneric lizards with different foraging strategies." Animal Behaviour 57: 393-
407.

168

 

 

‘ DeRusha, R. H., J. W. Forsythe, F. P. Dimarco and R. T. Hanlon (1989).
"Alternative diets for maintaining and rearing cephalopods in captivity."
Laboratory Animal Sciences 39: 306-312.

Dickel, L., J. G. Boal and B. U. Budelmann (2000). "The effect of early
experience on Ieaming and memory in cuttlefish." Developmental Psychobiology
36: 101-1 10.

Dickel, L., M. P. Chichery and R. Chichery (1997). "Postembryonic maturation of
the vertical lobe complex and early development of predatory behavior in the
cuttlefish (Sepia oli‘icinalis)." Neurobiology of Learning and Memory 67: 150-160.

Dickel, L., M. P. Chichery and R. Chichery (1998). "Time differences in the
emergence of short- and long-term memory during post-embryonic development
in the cuttlefish Sepia." Behavioural Processes 44: 81-86.

Dickel, L., M. P. Chichery and R. Chichery (2001 ). "Increase of Ieaming abilities
and maturation of the vertical lobe complex during postembryonic development in
the cuttlefish, Sepia." Developmental Psychobiology 40: 92-98.

DiCosmo, A., C. Cristo, A. Palumbo, M. D'lschia and J. B. Messenger (2000).
"Nitric oxide synthase (NOS) in the brain of the cephalopod Sepia oflicinalis."
Journal of Comparative Neurology 428: 411427.

Dukas, R., Ed. (1998). Cognitive Ecology: The Evolutionary Ecology of
Information Processing. Chicago, University of Chicago Press.

Fellows, B. J. (1967). "Chance stimulus sequences for discrimination tasks."
Psychological Bulletin 67: 87-92.

Fiorito, G. and P. Scotto (1992). "Observational Ieaming in Octopus vulgaris."
Science 256: 545-547.

Forsythe, J. W. and R. T. Hanlon (1997). "Foraging and associated behavior by
Octopus cyanea Gray 1849 on a coral atoll." Journal of Experimental Marine
Biology and Ecology 209: 15-31.

Garthwaite, J., G. Garthwaite, R. M. J. Palmer and S. Moncada (1989). "NMDA
receptor activation induces nitric oxide synthesis from arginine in rat brain slices."
Nature 336: 385-388.

Goldsmith, M. (1917). "Acquisition d'une habitude chez le Poulpe." CR Acaad Sci
Paris 164: 448.

Golledge, R. 6., Ed. (1999). Wayfinding Behavior: Cognitive Mapping and Other
Spatial Processes. Baltimore, Johns Hopkins University Press.

169

 

 

Golski, S., J. L. Olds, M. Mishkin, D. S. Olton and D. L. Alkon (1995). "Protein
kinase C in the hippocampus is altered by spatial but not cued discriminations: a
component task analysis." Brain Research 676: 5362.

Gray, E. G. and V. P. Whittaker (1962). "The isolation of nerve endings from
brain: an electron-microscope study of cell fragments derived by homogenization
and centrifugation." Journal of Anatomy? 96: 79-87.

Hall, K. C. and R. T. Hanlon (2002). "Principal features of the mating system of a
large spawning aggregation of the giant Australian cuttlefish Sepia apama
(Mollusca: Cephalopoda)." Marine Biology 140: 533-545.

Hanley, J. S., N. Shashar, R. Smolowitz, R. Bullis, W. Mebane, H. R. Gabr and
R. T. Hanlon (1998). "Modified laboratory culture techniques for the European
cuttlefish Sepia ofi‘icinalis." Biological Bulletin of the Marine Biological Laboratory,
Woods Hole 195: 223-225.

Hanlon, R. T. and J. B. Messenger (1988). "Adaptive coloration in young
cuttlefish (Sepia officinalis L.): The morphology and development of body
patterns and their relation to behavior." Philosophical Transactions of the Royal
Society of London B 320: 437-487.

Hanlon, R. T. and J. B. Messenger (1996). Cephalopod Behaviour. Cambridge,
Cambridge University Press.

Harlow, H. F. (1949). "The formation of Ieaming sets." Psychological Review 56:
51-65.

Hartwick, E. B., R. F. Ambrose and S. M. C. Robinson (1984). "Den utilization
and the movements of tagged Octopus dolfleini." Marine Behaviour and
Physiology 11: 95-1 10.

Kandel, E. R., J. H. Schwartz and T. M. Jessell, Eds. (2000). Principles of Neural
Science. New York, McGraw-Hill.

Karson, M. A., J. G. Boal and R. T. Hanlon (2003). "Experimental evidence for
spatial Ieaming in the cuttlefish Sepia ofl‘icinalis." Journal of Comparative
Psychology.

Kim, C. S., Y. F. Han, R. Etcheberrigaray, T. J. Nelson, J. Olds, T. Yoshioka and
D. L. Alkon (1995). "Alzheimer and beta-amyloid-treated fibroblasts demonstrate
a decrease in a memory-associated GTP-binding protein, Cp20." Proceedings of
the National Academy of Sciences (USA) 92: 3060-3064.

170

 

 

Kuzirian, A. M., F. M. Child, H. T. Epstein, P. J. S. Smith and C. T. Tamse
(1996). "Lead affects Ieaming by Hermissenda crassicomus. " Biological Bulletin
191: 260-261.

Kuzirian, A. M., H. T. Epstein, D. Buck, F. M. Child, T. J. Nelson and D. L. Alkon
(2003). "Pavlovian conditioning-specific increases of caZ+-and GTP-binding
protein, calexcitin in identified Hermissenda visual cells." Journal of

Neurocytology In press.

Kuzirian, A. M., H. T. Epstein, T. J. Nelson, N. S. Rafferty and D. L. Alkon (1998).
"Lead learning and calexcitin in Hermissenda." Biological Bulletin 195: 198-201.

Long, T. M., R. T. Hanlon, A. Ter Maat and H. M. Pinsker (1989). "Non-
associative Ieaming in the squid Lolliguncula brevis (Mollusca, Cephalopoda)."
Marine Behaviour and Physiology 16: 1-9.

Mackintosh, J. (1962). "An investigation of reversal learning in Octopus vulgaris
Lamarck." Quarterly Journal of Experimental Psychology 14: 15-22.

Mackintosh, N. J. and J. Mackintosh (1963). "Reversal Ieaming in Octopus, with
and without irrelevant cues." Quarterly Journal of Experimental Psychology 15:
236-242.

Mackintosh, N. J. and J. Mackintosh (1964). "Performance of Octopus over a
series of reversals of a simultaneous discrimination." Animal Behaviour 12: 321-
324.

Mackintosh, N. J., B. Wilson and R. A. Boakes (1985). "Differences in
mechanisms of intelligence among vertebrates." Philosophical Transactions of
the Royal Society of London B 308: 53-63.

Maddock, L. and J. Z. Young (1987). "Quantitative differences among the brains
of cephalopods." Journal of Zoology (London) 212: 739-767.

Mather, J. A. (1991). "Navigation by spatial memory and use of visual landmarks
in octopuses." Journal of Comparative Physiology A 168: 491-497.

Mather, J. A. (1995). "Cognition in Cephalopods." Advances in the Study of
Behavior 24: 317-353.

Mather, J. A. and R. C. Anderson (1993). "Personalities of octopuses (Octopus
rubescens)." Journal of Comparative Psychology 107: 336-340.

Matzel, L. D., A. C. Talk, I. A. Muzzio and R. F. Rogers (1998). "Ubiquitous

molecular substrates for associative learning and activity-dependent neuronal
facilitation." Reviews in the Neurosciences 9: 129-167.

171

 

 

Messenger, J. (1977). "Prey-capture and Ieaming in the cuttlefish, Sepia."
Symposia of the Zoological Society of London 38: 387-395.

Messenger, J. a. G. S. (1972). "\fisual preference and two-cue discrimination
learning in Octopus." Animal Behaviour 20: 580-585.

Messenger, J. B. (1968). "The visual attack of the cuttlefish Sepia ofl‘icinalis."
Animal Behaviour 16: 342-357.

Messenger, J. B. (1971). The optic tract lobes. The anatomy of the nervous
system of Octopus vulgaris. J. Young. Oxford, Clarendon Press: 481-506.

Messenger, J. B. (1973). "Learning in the cuttlefish, Sepia." Animal Behaviour 21 :
801 -826.

Messenger, J. B. (1973). "Learning performance and brain structure: a study in
development." Brain Research 58: 519-523.

Moriyama, T. and Y. P. Gunji (1991). "Autonomous Ieaming in maze solution by
Octopus." Ethology 103: 499-513.

Muntz, W. R. A. (1961). "lnterocular transfer in Octopus vulgaris." Journal of
Comparative Physiological Psychology 54: 49-55.

Nelson, T. J. and D. L. Alkon (1991). "GTP-binding proteins and potassium
channels involved in synaptic plasticity and Ieaming." Molecular Neurobiology 5:
31 5328.

Nelson, T. J. and D. L. Alkon (1995). "Phosphorylation of the conditioning-
associated GTP-binding protein cp20 by protein kinase C." Journal of
Neurochemistry 65: 2350-2357.

Nelson, T. J., S. Cavallaro, C.-L. Yl, D. McPhie, B. G. Schreurs, P. Gusev, A.
Favit, O. Sohar, J. Kim, S. Beushausen, G. Ascolie, J. L. Olds, R. Neve and D. L.
Alkon (1996). "Calexcitin: a signaling protein that binds calcium and GTP, inhibits
potassium channels and enhances membrane excitability." Proceedings of the
National Academy of Sciences 93: 13808-13813.

Nelson, T. J., T. Yoshioka, S. Toyoshima, Y. F. Han and D. L. Alkon (1994).
"Characterization of a GRP-binding protein implicated in both memory storage
and interorganelle vesicle transport." Proceedings of the National Academy of
Sciences of the USA 91: 9287-9291.

Norman, M. (2000). Cephalopods: A World Guide. Hackenheim, ConchBooks.

172

 

O'Dor, R. K. and D. M. Webber (1986). "The constraints on cephalopods: why
squid aren't fish." Canadian Journal of Zoology 64: 1591-1605.

Olds, J. L. and D. L. Alkon (1990). "A role for protein kinase c in associative
Ieaming." The New Biologist 3: 27-35.

Olds, J. L., M. L. Anderson, D. L. McPhie, L. D. Staten and D. L. Alkon (1989).
"Imaging of memory-specific changes in the distribution of protein kinase C in the
hippocampus." Science 245: 866-869.

Olds, J. L., S. Golski, D. I. McPhie, D. L. Olton, M. Mishkin and D. L. Alkon
(1990). "Discrimination Ieaming alters the distribution of protein kinase C in the
hippocampus of rats." Journal of Neuroscience 10: 3707-3713.

Packard, A. (1972). "Cephalopods and fish: the limits of convergence." Biological
Reviews 47: 241-307.

Packard, A. (1988). Visual tacties and evolutionary strategies. Cephalopods:
Present and Past. J. W. a. J. Kullmann. Stuttgart, Schweizerbart'sche
Verlagsbuchhandlung: 89-103.

Parriss, J. R. (1965). "Learning and the dorsal basal lobe in the octopus." Journal
of Comparative Neurology 125: 1-8.

Robertson, J. D. (1994). "Cytochalisin D blocks touch Ieaming in Octopus
vulgaris." Proceedings of the Royal Society of London B 258: 61-66.

Robertson, J. D., J. Bonaventura and A. P. Kohm (1994). "Nitric oxide is required
for tactile Ieaming in Octopus vulgaris." Proceedings of the Royal Society of
London B 256: 269-273.

Robertson, J. D., J. Bonaventura, A. P. Kohm and M. Hiseat (1996). "Nitric oxide
is necessary for visual Ieaming in Octopus vulgaris." Proceedings of the Royal
Society of London B 263: 1739-1743.

Sanders, F. K. and J. Z. Young (1940). "Learning and other functions of the
higher nervous centres of Sepia." Journal of Neurophysiology 3: 501-526.

Sanders, G. D. (1975). The cephalopods. Invertebrate Learning. J. D. a. A. W.
WC Corning. New York, Plenum Press. 3: Cephalopods and Echinoderms: 1-
101.

Sherry, D. F. (1998). Ecology and neurobiology of spatial memory. Cognitive
ecology: The evolutionary ecology of information processing and decision
making. R. Dukas. Chicago, University of Chicago Press: 261-296.

173

 

 

 

Siegel, S. and N. J. Castellan, Eds. (1988). Nonparametric Statistics for the
Behavioral Sciences. New York, McGraw Hill.

Solem, A. (1974). The Shell Makers: Introducing Mollusks. New York, John Wiley
and Sons.

Sutherland, N. S. and N. J. Mackintosh (1971). Mechanisms of Animal
Discrimination Learning. New York, Academic Press.

Sutherland, N. S. and W. R. A. Muntz (1959). "Simultaneous discrimination
training and preferred direction of motion in visual discrimination of shape in
Octopus vulgaris." Pubblicazioni della Stazione Zoologica di Napoli 31: 109-126.

Teichert, C. (1988). Main features of cephalopod evolution. In: Paleontology and
Neontology of Cephalopods (M.R. Clarke and. ER. Truemann, eds). San Diego,
Academic Press. 12: 11-79.

Thompson, R. (1957). "Successive reversals of a position habit in an
invertebrate." Science 126: 163-164.

Thorpe, W. H. (1963). Learning and Instinct in Animals. London, Methuen.

Tompsett, D. H. (1939). "Sepia." Liverpool Marine Biology Committee Memoirs
32: 1-184.

Uexkiill, J. v. (1905). Leiftaden in das Stadium der Experimentellen Biologie der
Wassertiere. Wiesbaden, Bergmann.

Walker, J. J., N. Longo and M. E. Bittennan (1970). "The octopus in the
laboratory: handling maintenance, training." Behavior Research Methods and
Instrumentation 2: 15-18.

Wehner, R. (1981). Spatial vision in arthropods. Handbook of Sensory
Physiology. H. Atrium. New York, Springer-Verlag. VII: 287-617.

Wells, M. J. (1958). "Factors affecting reactions to Mysis by newly hatched
Sepia." Behaviour 13: 96—1 11.

Wells, M. J. (1962). "Early Ieaming in Sepia." Symposia of the Zoological Society
of London 8: 149-169.

Wells, M. J. (1964). "Detour experiments with octopus." Journal of Experimental
Biology 40: 187-193.

Wells, M. J. (1965). "The vertical lobe and touch Ieaming in the octopus." Journal
of Experimental Biology 42: 233-255.

174

 

 

Wells, M. J. (1978). Octopus: Physiology and Behaviour of an Advanced
Invertebrate. London, Chapman and Hall.

Wells, M. J. and J. Wells (1956). "Tactile discrimination and the behaviour of
blind Octopus." Pubblicazioni della Stazione Zoologica di Napoli 28: 94-126.

Wells, M. J. and J. Wells (1957). "The effect of lesions to the vertical and optic
lobes on tactile discrimination in Octopus." Journal of Experimental Biology 34:
378-393.

Wells, M. J. and J. Wells (1958). "The effect of vertical lobe removal on the
performance of octopuses in retention tests." Journal of Experimental Biology 35:
337-348.

Wells, M. J., J. Wells and R. K. O'Dor (1992). "Life at low oxygen tensions: the
behaviour and physiology of Nautilus pompilius and the biology of extinct forms."
Journal of the Marine Biological Association of the UK 72: 313-328.

Wells, M. J. and J. Z. Young (1966). "Lateral interaction and transfer in the tactile
memory of Octopus." Journal of Experimental Biology 45: 383-400.

Wiiz, K. (1954). "Etudes quantatives sur le systéme nerveux des Céphalopodes."
Comptes Rendus de I'Académie des Sciences 238: 1353-1355.

Young, J. Z. (1938). The evolution of the nervous system and of the relationship
of organism and environment. Evolution: essays presented to ES. Goodrich.
Oxford, Clarendon Press.

Young, J. Z. (1939). "Fused neurons and synaptic contacts in the giant nerve
fibres of cephalopods." Philosophical Transactions of the Royal Society of
London B 229: 465-503.

Young, J. 2. (1958). "Effect of removal of various amounts of the vertical lobes on
visual discrimination by Octopus." Proceedings of the Royal Society of London B
149: 441-462.

Young, J. Z. (1960). "The failures of discrimination Ieaming following the removal
of vertical lobes in Octopus." Proceedings of the Royal Society of London B 153:
1 8-46.

Young, J. Z. (1961). "Learning and discrimination in the octopus." Biological
Reviews 36: 32-96.

Young, J. Z. (1962). "Repeated reversal of training in Octopus." Quarterly Journal
of Experimental Psychology 14: 206-222.

175

Young, J. Z. (1963). "Light— and dark-adaptation in the eyes of some
cephalopods." Proceedings of the Zoological Society of London 140: 255-272.

Young, J. Z. (1964). A Model of the Brain. Oxford, Clarendon.

Young, J. Z. (1965). "The organization of a memory system." Proceedings of the
Royal Society of London B 163: 285-320.

Young, J. Z. (1971). The anatomy of the nervous system of Octopus vulgaris.
Oxford, Clarendon Press.

Young, J. Z. (1974). "The central nervous system of Loligo. I. The optic lobe."
Philosophical Transactions of the Royal Society of London B 267(263-302).

 

Young, J. Z. (1977). "The nervous system of Loligo. Ill. Higher motor centers:
the basal supraoesophageal lobes." Philosophical Transactions of the Royal
Society of London B 276(351-398).

Young, J. Z. (1979). "The nervous system of Loligo. V. The vertical lobe
complex." Philosophical Transactions of the Royal Society of London B 285: 311-
354.

Young, J. Z. (1983). "The distributed tactile memory system of Octopus."
Proceedings of the Royal Society of London B 218: 135-176.

Young, J. Z., Ed. (1988). Evolution of the Cephalopod Brain. The Mollusca. San
Diego, Academic Press.

Young, J. Z. (1991). "Computation in the Ieaming system of cephalopods."
Biological Bulletin 180: 200—208.

Zar, J. H. (1999). Biostatistical Analysis. Upper Saddle River, Prentice Hall.

 

176