Burish Wang03 BBE Preprint

7/30/2019 Burish Wang03 BBE Preprint

http://slidepdf.com/reader/full/burish-wang03-bbe-preprint 1/31

Brain architecture and social complexity in birds and dinosaurs

Mark J. Burish†, Hao Yuan Kueh‡, and Samuel S.-H. Wang*

Department of Molecular Biology, Princeton University, Princeton, NJ 08544.

*Corresponding author:Dr. Samuel WangPrinceton University Molecular BiologyLTL Bldg., Washington RoadPrinceton, NJ 08544 USATelephone: (609) 258-0388FAX: (609) 258-1035E-mail: [email protected]

†Current address: 340 Light Hall, Vanderbilt University, Nashville, TN [email protected]

‡Current address: Graduate Program in Biophysics, Harvard University, Cambridge, MA [email protected]

Running title: Avian brain architecture and social complexity

Key words: brain architecture, cognitive, social, Machiavellian intelligence, multivariate, birds, Archaeopteryx

6 figures, no tables



Burish et al. - Avian brain architecture and social complexity 12/3/02 12:31 PM

2

Abstract. Vertebrate brains vary tremendously in size, but differences in form are

more subtle. To bring out functional contrasts that are independent of absolute

size, we have normalized brain component sizes to whole brain volume. The set

of such volume fractions is the cerebrotype of a species. Using this approach in

mammals we previously identified specific associations between cerebrotype and

behavioral specializations. Among primates, cerebrotypes shift in a progressive

manner, are linked principally to enlargement of the cerebral cortex, and are

associated with increases in the complexity of social structure. Here we extend

this analysis to include a second major vertebrate group, the birds. In birds, the

telencephalic volume fraction is strongly correlated with social complexity. This

correlation can explain nearly three-fourths of the observed variation in relative

telencephalic size and is not found for other avian behavioral specializations.

Interpolating this correlation for Archaeopteryx , an ancient bird, suggests that its

social complexity was likely to have been on a par with modern chickens. The

relative size of forebrain structures may be an anatomical substrate for social

complexity, and perhaps cognitive ability, that can be generalized across a range

of vertebrate brains, including dinosaurs.

Introduction

How can the architectures of brains be compared to bring out functional contrasts

independent of absolute size? For instance, among mammals, from the desman to the sperm

whale, brain size varies by over five orders of magnitude. At the same time, across phylogeny




3

mammalian brains also show broad similarities in structure, internal connectivity, and anatomical

composition [Niewenhuys et al., 1998]. Thus, examining brain architecture to obtain functional

principles presents a great challenge to comparative neuroanatomy.

One approach to this problem is to make quantitative comparisons [Jerison, 1973;

Stephan et al., 1981; Armstrong, 1983; Jerison, 1991; Finlay and Darlington, 1995; Barton and

Dunbar, 1997]. Investigations have been directed at describing statistical trends and exceptions,

and also finding explanations for these observations. The result has been a variety of approaches

to comparing brains, each with its own strengths and weaknesses.

In one early approach, brain size was examined using body size as a reference measure

[Jerison, 1973; Jerison, 1991]. This approach revealed relationships that appear to follow power

laws; that is, a trend that is approximately fitted by a line in log-log coordinates. These are called

allometric scaling relationships [Thompson, 1942] and have been found repeatedly throughout

the animal kingdom for many anatomical components. Such scaling relationships, even if they

do not yet have a functional explanation, might be used as a baseline from which to compare

individual species. Such comparison has in turn been used to identify outliers – individual

species that depart from the scaling trend. This approach succeeds in identifying general

deviations, such as the extremes of human and dolphin brain size compared with other mammals

of similar body mass. These deviations, also known as residuals, are usually expressed as an

“encephalization quotient” [Jerison, 1973].

Residuals analysis has also been attempted using individual brain components. In this

case the sizes of individual brain components are fitted against body size [Armstrong, 1983;

Harvey and Krebs, 1990], whole brain [Douglas and Marcellus, 1975], or brain subdivisions




4

[Barton et al., 1995; Barton and Harvey, 2000] as reference measures using data for an entire set

of animals, and specializations in brain components are identified. In all cases, this approach has

several drawbacks. First, fit lines vary widely depending on how a data set is analyzed; both the

slope and intercept depend on the taxonomic level at which the data are grouped [Gould, 1975;

Martin and Harvey, 1985; Nealen and Ricklefs, 2001; Wang et al., 2002]. Therefore a linear fit

to pooled data contains the intrinsic ambiguity that no species is associated with a unique scaling

trend. Scaling trends derived this way are unlikely to capture the overall statistical properties of

the data set. Thus, residuals calculated relative to some fit may contain artifacts that are very

difficult to interpret [Wang et al., 2002]. Second, individual brain components each have their

own allometries [Fox and Wilczynski, 1986]; therefore, the combination of any two of them will

hardly ever be expected to follow a strict power law [Zhang and Sejnowski, 2000]. Finally, even

if taxonomic groups (clades) and brain components were subdivided to the point that basic size

trends could be identified clearly, few functional rationales have yet been found to account for

such relationships (though see [Harvey and Krebs, 1990; Stevens, 2001] and the extensive work

in circulatory systems [West et al., 1999]).

In a second type of approach, the use of allometric relationships has been bypassed by

analyzing absolute brain component volumes [Finlay and Darlington, 1995; Finlay et al., 2001].

Such an approach reveals that differences in component sizes across species tend to be correlated

to each other, consistent with a developmental principle of concerted brain growth. In this model

brain growth is regulated in part by a shared set of instructions that govern overall developmental

rates. This work has since been extended to include multivariate analysis as a means of

searching for patterns in growth spanning multiple brain regions [Finlay et al., 2001]. However,




5

the results of such an analysis have been dominated by the effects of absolute size differences

among species. Also, most analyses of this type so far still begin by identifying common trends,

an approach that once again depends on initial examination of the data set as a whole.

Recently, as a means of eliminating the effects of absolute size we have normalized brain

component data on a species-by-species basis. We divide a component’s volume by whole brain

volume to obtain its volume fraction F [Jerison, 1991; Clark et al., 2001]. Although

normalization is straightforward, it differs from prior analyses. Most importantly, this definition

of relative component size does not depend on any cross-species fits. Volume fractions of all

components of the brain for a given species can be defined as the “cerebrotype” for that species,

and it is possible to quantify the dissimilarity in brain structure between two species by simply

calculating a Euclidean distance between the two cerebrotypes. These distances can be analyzed

easily using multivariate methods such as multidimensional scaling or principal component

analysis (also known as singular value decomposition).

Of the possible reference standards for size normalization, we have chosen the whole

brain as a baseline because generally, brain components connect principally with other brain

structures. Since larger targets are likely to require more neural projections, the combined size of

all targets (i.e. the whole brain) is a natural size standard. Another method for normalization has

been to divide the component size by brainstem size [Lefebvre et al., 1997; de Winter and

Oxnard, 2001]; however, that approach creates new interdependencies among components that

bias the outcome of multivariate analysis. Also, different brain components develop under

shared developmental growth mechanisms, suggesting that one way to identify size differences




7

The resulting graphs facilitate the identification of finer taxonomic distinctions among

taxa. For instance, Scandentia (tree shrews), once categorized as primates [Stephan et al., 1981],

are quite well separated from them in cerebrotype (Figure 2a) and are in fact nearer to

insectivores. The insectivores themselves also can be subdivided into distinct groups of points

representing taxonomic subdivisions (Figure 2b), with the exception of the overlapping

cerebrotype distributions of Soricidae (shrews) and Tenrecidae (tenrecs). This exception is

instructive because it arose entirely from the cerebrotypes of three tenrecs that live and/or hunt in

the water. These tenrecs have reduced olfactory bulbs and piriform cortex, specializations that

may reflect differential needs of aquatic versus terrestrial lifestyles in tenrecs. Thus, cerebrotype

analysis provides a means of taxonomically grouping mammals by their brain structure.

In the case of primates, cerebrotype-based measures have helped identify coordinated

changes in brain architecture (Figure 2c). According to the social intelligence hypothesis of

primate brain evolution [Whiten and Byrne, 1988; Dunbar, 1995; Barton and Dunbar, 1997], a

larger cerebral cortex may confer selection advantages by providing increased cognitive capacity

for social dynamics. This social selection pressure would lead to progressive relative

enlargement of the cerebral cortex over time.

Fossil and molecular evidence and morphological evidence from non-brain characters

shows that successively more derived primate taxa have arisen over time: lemurs/lorises,

followed by tarsiers, New World monkeys, Old World monkeys, and then hominoids. The

arrangement of the corresponding cerebrotypes is collinear with this order, with the exception of

some overlap between New World and Old World monkeys. Interestingly, this overlap arises

from the four New World monkeys in this database – Ateles, Lagothrix, Cebus, and Saimiri –




8

with larger group sizes and complex social structures resembling those of Old World monkeys

[Kinzey, 1997]. The cerebrotypes of large group-size New World monkeys differ from other

New World monkeys principally in F neo (large group-size 69.3±0.5%, other 61.8±1.7%) but are

similar to Old World monkeys (70.4±2.4%). Taken together, these results suggest that increased

derivation was indeed accompanied by progressive changes in the cerebrotype.

Birds

Like mammals, birds are highly diverse and have spread to fill a variety of aerial,

terrestrial, and marine niches throughout the planet. However, birds and mammals have diverged

for more than 300 million years since their common tetrapod ancestor, yielding brain plans that

differ in fundamental ways. Each lineage has distinctive brain structural specializations

[Niewenhuys et al., 1998], such as the retention of nuclear organization in birds and the

appearance of a prominent laminated cerebral cortex in mammals.

Birds and mammals also share a number of convergent features. Both groups are warm-

blooded and have a higher basal metabolic rate than most other vertebrates. Compared to other

vertebrates, both groups have large brains relative to body weight; this relationship is nearly the

same for these two groups but is significantly offset compared with reptiles and amphibians

[Jerison, 1991]. Bird body sizes span most of the range for mammals (though mammals can be

larger). Finally, certain bird and primate lineages (songbirds and primates) have been suggested

to have evolved with great rapidity [Wyles et al., 1983]. Given these similarities, birds present

an attractive comparison group relative to mammals.




9

Materials and Methods

We used brain measurements from 33 bird species [Nottebohm, 1981; Rehkämper et al.,

1991; DeVoogd et al., 1993; Boire and Baron, 1994]. Body weights were taken from the same

literature sources except for the canary [Harper and Turner, 2000]. Measurements of brain

parameters in theropod dinosaurs and Archaopteryx came from reports using CT scanning of

endocasts for three-dimensional reconstruction [Rogers, 1998; Larsson et al., 2000].

Birds were sorted into categories of social structure using flock size and interactions

between individuals [del Hoyo et al., 1992; Cramp, 1994; Zann, 1996]. Solitary birds (group

size: 1 to 3 birds) are often territorial, with the majority of intraspecies interactions occurring

during the mating season. Covey birds (group size: 5 to 50 birds) commonly show dominance

structures such as pecking orders and leks. Colonial birds (group size: hundreds to thousands of

conspecifics) exhibit communal structures such as gregarious roosts and large foraging flocks.

Transactional birds (group size: variable, often found in large flocks) show a high level of

communal structure and strong one-on-one interactions outside the mating season. In this data

set, the transactional birds were the mallard ( Anas platyrhynchos), a highly gregarious species

with socially coordinated loafing and other comfort activities [Desforges and Wood-Gush, 1975;

Brodsky et al., 1988; Cramp, 1994]; two songbirds (Corvus c. corone, carrion crow; Garrulus

glandarius, jay), which exhibit social learning and unusual spring social gatherings [Goodwin,

1986; Cramp, 1994]; the magellanic penguin (Spheniscus magellanicus), which is highly

communal and altruistically defends the nests of unrelated neighbors [Boersma, 1988; Capurro et

al., 1988; del Hoyo et al., 1992]; and two parrots ( Melopsittacus undulatus, budgerigar; Pionus

menstruus, blue-headed parrot), which display a hierarchy based on age, association, and




10

experience that has been compared to primate fission-fusion social structures [Ingels, 1978;

Lowry, 1991; del Hoyo et al., 1992].

For multivariate analysis, Euclidean cerebrotype distances were calculated as previously

described [Clark et al., 2001] among 28 species [Boire and Baron, 1994] using telencephalon,

diencephalon, mesencephalic tegmentum, optic tectum, cerebellum, and myelencephalon. For

the display of these distances, multidimensional scaling was used as a means of representing the

variation in brain structure in two dimensions. Multidimensional scaling was performed as

previously described [Clark et al., 2001] using MATLAB (The Mathworks, Inc., Natick, MA).

Regression tree analysis was performed to analyze the relationship between telencephalic

volume fraction and behavioral categories. Regression tree analysis [Venables and Ripley, 1994]

is a form of multivariate analysis in which a grouping (in this case, by behavioral category) is

found that accounts for the most variance in a set of observations; the data are split according to

the grouping; and the procedure is repeated. Regression tree analysis was implemented using the

R language (The R Project, http://www.r-project.org/).

Results

We used brain measurements from 33 species [Nottebohm, 1981; Rehkämper et al., 1991;

DeVoogd et al., 1993; Boire and Baron, 1994; Clark et al., 2001]. The variation in volume

fractions among species ranged from 0.9% to 8.3% (standard deviation) of total brain volume.

The greatest variation occurred in the volume fraction of telencephalon ( F tel ), 58.4 ± 8.3% (mean

± s.d.; n=33 species). The overall range for F tel , 44% to 78%, is quite similar to the range found

in mammals, 50% to 85% [Clark et al., 2001]. Likewise, for F cbl , the overall range (7 to 21%)




11

and variation (14.2 ± 3.4%, 28 species) are similar to mammals (range 6 to 23% by individual

species, variation 13.5 ± 2.4% across 19 mammalian taxa; see Figure 2 of Clark et al., 2001).

The variability in telencephalon and non-telencephalic brain components can also be

quantified relative to body weight. Using a power law model (that is, a linear fit in log-log

coordinates), the correlation of body size with brain size (r 2=0.822, n=33) leaves 17.8% of the

variance unexplained. The amount of unexplained variance (Figure 3) is greater for

telencephalon (23.8%) than for non-telencephalic regions (11.1%; P =0.05, one-tailed z-test).

The relative strictness of scaling in non-telencephalic brain components is consistent with the

idea that a given body size requires a certain amount of “basal” brain [Portmann, 1947a,b; Boire

and Baron, 1994]. Conversely, the observed variations in telencephalic volume suggest that this

structure may reflect functional variations that are not dependent on body size.

Because of the correspondence between telencephalic size and cognitive complexity in

primates [Reader and Laland, 2002], we looked for quantifiable analogous parameters of bird

behavior. As one measure, we assigned species to categories based on social interaction. We

focused on the qualities of group size and inter-individual interaction because they are relatively

straightforward to observe and can be directly compared to other behaviors such as eating habits

and mating type. Other measures relating to cognitive complexity have previously been used,

such as foraging and nesting innovations [Lefebvre et al., 1997; Lefebvre et al., 1998;

Nicolakakis and Lefebvre, 2000] and tool-making [Hunt et al., 2001; Weir et al., 2002]. These

qualities and social complexity are likely to share common neural substrates. In primates,

various measures of cognitive ability are strongly correlated with each other and with indices of

cerebral cortical size [Reader and Laland, 2002]. Furthermore, group size has been used in




12

primates as an indicator of social complexity [Dunbar, 1995]. We therefore considered social

complexity as a rapid means of sorting species that does not require an exhaustive survey of

known behaviors [Lefebvre et al., 1997; Lefebvre et al., 1998; Nicolakakis and Lefebvre, 2000;

Reader and Laland, 2002].

Sorting the birds into categories of social structure revealed a strong correspondence

between telencephalic volume fraction ( F tel ) and level of social complexity (Figure 4a).

Progressively increasing ranges of F tel were seen for solitary (49.9 ± 2.2%; mean ± s.d., n=4),

covey (53.9 ± 4.6%, n=13), colonial (59.8 ± 5.1%, n=10), and transactional birds (71.5 ± 4.0%,

n=6) (ANOVA: P<0.001). Pairwise comparison showed significant differences (two-tailed t-

test, P<0.02) between all groupings except for solitary vs. covey birds ( P=0.12). Thus,

increasing degrees of social complexity are accompanied by an increased relative volume of the

telencephalon.

To test for possible links between F tel and traits not explicitly cognitive in nature, we also

divided the species in five other ways (Figure 4a): 1) Eating habits, into herbivores, omnivores,

and carnivores (specifically, fish-, insect-, and/or marine invertebrate-eating birds). 2) Migration

pattern, into sedentary (permanent residents), migratory (seasonal residents), and nomadic (long-

distance travelers that establish new home territories) [Faaborg, 1988; Zann, 1996]. 3) Flight

capacity, into flightless/terrestrial (no flight or short and heavy flight) and aerial. 4) Mating type,

into monogamous or polygynous [Faaborg, 1988; del Hoyo et al., 1992]. 5) Learned

vocalization: no mimicry (no known capacity for learned vocalizations) or vocal learning

(capable of intraspecies or interspecies learned vocalizations).




13

One difficulty in quantitatively determining the contributions of these traits to F tel is the

possibility that memberships in different categories may co-vary with one another. This problem

can be addressed with regression tree analysis [Venables and Ripley, 1994], a form of

multivariate analysis that is particularly appropriate for small data sets. For this data set, the

largest contributions to variation in F tel came from social structure (Figure 4b): the first and

second branch points separated birds into transactional , colonial , and covey/solitary and

accounted for a combined total of 71.2% of the variance in F tel . Learned vocalization and flight

accounted for a small amount of the remaining variance in F tel (3.2% and 4.4% of total variance,

respectively; rightmost branch points, Figure 4b). The low contribution of learned vocalization

is consistent with the observation that the song system comprises less than 10% of the

telencephalon [DeVoogd et al., 1993]. Therefore, the bulk of the apparent contribution of these

traits to variation in F tel can most parsimoniously be accounted for by a single general factor,

social complexity as defined by our groupings.

We also searched for correlates of behavioral categories in other patterns of brain

architecture. Analysis of non-telencephalic volume fractions revealed no significant correlation

between behavioral groupings and any single brain component (ANOVA and t -test; data not

shown). In search of more subtle relationships, we then performed multidimensional scaling on

the cerebrotypes. The resulting relational maps showed little or no clustering by phylogenetic

grouping (Figure 5a). The dependence of F tel on phylogeny was tested specifically by arranging

species according to a published phylogeny [Sibley and Ahlquist, 1991] and calculating, at each

node, differences in DNA hybridization temperature [Sibley and Ahlquist, 1991] as a measure of

evolutionary distance. This quantity and differences in F tel were uncorrelated (r S = –0.14),




14

indicating that variations in telencephalic size cannot be explained by random evolutionary

divergence [Harvey and Pagel, 1991].

Species were separable by social structure (Figure 5b) in a pattern that matched the

regression tree analysis. Re-analysis done after exclusion of the telencephalon [Clark et al.,

2001] eliminated this relationship (Figure 5c), indicating that interspecies variations related to

social complexity are confined to the telencephalon and not other brain divisions. No other

behavioral trait was associated with distinct groupings of either whole-brain or telencephalon-

excluded cerebrotypes (data not shown).

Dinosaurs

Our findings for birds and mammals indicate that the telencephalon's role in guiding

social interactions has either evolved independently multiple times or, more likely, has persisted

throughout hundreds of millions of years of evolution. In the case of birds, the relationships we

found are so strong that we were motivated to examine their predecessors, the dinosaurs. This

case is of particular interest because behavioral data for these animals is speculative, based on

fossil footprints or physical features [Hopson, 1977]. Thus brain structure may provide a major

source of information [Edinger, 1961].

Dinosaurs (Saurischia and Ornithischia) arose in the late Triassic era approximately 230

million years ago. Until their extinction approximately 65 million years ago, along with other

reptiles ( Pterosauria and Crocodylia) they were the dominant terrestrial vertebrates [Hopson,

1977; Sereno, 1999]. They spanned a large range of body sizes, from 45 kg to nearly 100,000 kg;

the lower end of this range overlaps with the upper ranges of living reptiles and birds [Hopson,




15

1977]. Approximately 150 million years ago, during the Late Jurassic, they gave rise to an

ancient bird, Archaeopteryx.

Some dinosaur skulls have been preserved with sufficient integrity that endocasts reveal

the size and external form of the brain [Edinger, 1961; Edinger, 1964]. The general form of

dinosaur brains is reptilian [Jerison, 1969; Hopson, 1977; Larsson et al., 2000]; for instance, the

features of Allosaurus brains are very similar to those of modern crocodilians [Rogers, 1998].

The sizes of dinosaur brains are also consistent with predictions made by extrapolating values

from modern reptiles along allometric straight-line fits [Jerison, 1969; Hopson, 1977]. An

exception to these patterns in shape and size is found in the case of Archaeopteryx [Edinger,

1926; Hopson, 1977], whose brain has more prominent hemispheres and follows the external

contours of the skull, similar to modern birds. Archaeopteryx’s body was an estimated 300-500 g

[Jerison, 1969; Hopson, 1977] and its brain was approximately 1.1 cm3 [Larsson et al., 2000],

placing it much closer to the brain-body allometric relationship for living birds than for living

reptiles.

Because impressions found in endocasts can reveal surface features such as the

boundaries between brain components, gross aspects of dinosaur brain architecture can be

compared with living forms (Figure 6). One feature that can be distinguished is the

telencephalon [Rogers, 1998; Larsson et al., 2000]; this allows estimation of F tel . The

telencephalons of theropods have been measured using three-dimensional reconstruction by CT

scanning of endocasts [Rogers, 1998; Larsson et al., 2000]. This method has found F tel values of

24% (Carcharodontosaurus), 28% ( Allosaurus), and 33% (Tyrannosaurus) [Larsson et al.,

2000]. These values are at or below the range of living reptiles (33% to 56%) [Platel, 1976] and




16

considerably smaller than any bird. In contrast, an Archaeopteryx endocast shows a value for F tel

of 45%. This value is within the range of solitary and covey birds. In particular, the brain

proportions of Archaeopteryx are similar to chickens, which have the smallest F tel (44%)

considered here. Thus, extrapolation from the known behavior and brain structure of living

forms suggests that brains of dinosaurs (and thus social complexity, and perhaps other cognitive

traits) were comparable to those of reptiles living today, and that Archaeopteryx is likely to have

been a solitary or covey bird, on a par with chickens.

Discussion

Previous analysis has suggested a central role for the cerebral cortex in the rise of social

or “Machiavellian” intelligence in primates [Whiten and Byrne, 1988; Dunbar, 1995; Barton and

Dunbar, 1997]. Our recent results show that in birds as well, the degree of social complexity

correlates strongly with relative telencephalic size. The telencephalon therefore plays a central

role in guiding social interactions in two highly divergent vertebrate taxa.

Variation in the architecture of bird brains has been considered previously by Portmann

[Portmann, 1946; Portmann, 1947a; Portmann, 1947b]. He found that taxa with relatively large

adult brains (scaled to body size) tend at hatching to be altricial, reaching physical independence

late and thus requiring intensive parental care. The adult brains of altricial birds are also those

with large hemispheres, that is, with a large F tel . Conversely, precocial birds tend to have smaller

adult brains and smaller values of F tel . Similarly, in primates, late postnatal maturation is

correlated with relative enlargement of the neocortex [Allman and Hasenstaub, 1999]. Taken




17

together, these results suggest that accelerated telencephalic growth, altriciality, and social

complexity are generally linked phenomena in warm-blooded vertebrates.

Our findings support the view that a major functional role of the vertebrate telencephalon

is the regulation of social interaction. They also suggest that a layered cerebral cortex is not

necessary for such telencephalic functions. This may be true for other cognitive skills besides

social intelligence: a number of the birds described in our study as “transactional” are known for

their reasoning abilities. For instance, crows can modify objects into tools without prior

experience [Weir et al., 2002] and can manufacture functionally lateralized (“handed”) tools

using skilled, multiple-step crafting procedures [Hunt, 2000; Hunt et al., 2001]. Parrots are

capable of impressive feats of communication and reasoning and can master concepts such as

number, relative comparison, and object permanence [Pepperberg and Brezinsky, 1991;

Pepperberg, 2002]. These birds have large relative telencephalic volume, and their cognitive

capabilities include planning, reasoning, and prediction functions generally considered the

province of the mammalian cerebral cortex.

We have applied our analysis of living birds to an extinct bird, Archaeopteryx, and have

placed putative limits on its degree of social complexity based on extrapolation from its related

living successors. Although until now published descriptions of fossil braincases [Molnar, 1985;

Wittmer, 1990; Chatterjee, 1991; Elzanowski, 1991; Elzanowski and Galton, 1991; Currie and

Zhao, 1993] have only rarely included volume estimates of divisions of the brain, our work

indicates that such measurements may be of interest.

In the future, quantitative approaches to neuroanatomy and behavior can be improved in a

number of ways. From an ethological standpoint, it would be useful to arrive at a measure of




18

behavioral complexity that can be applied across species [Bullock, 1993]. An ideal solution to

this problem would allow comparisons not only within large clades (e.g. birds, mammals) but

also across all animals. However, even though problem-solving ability and social complexity

have been observed in animals as divergent as crows, octopuses, dolphins, and humans,

comparing them with each other is, at present, quite difficult.

Another unresolved issue is the detailed functional interpretation of volume fraction. In

other words, what is the relationship between a structure’s relative or absolute size and its

functional role in the entire nervous system? Underlying the simple measures of macroscopic

volume is a wide array of changes that also vary systematically across species (see Harrison et al.,

in press). For instance, the cellular components of cerebral cortex change in number [Changizi,

2001; Elston et al., 2001], size [Elston et al., 2001] and connectivity [Braitenberg, 2001]; the

relationship between these parameters to function is unknown. With sufficient understanding of

the underlying cellular and network processes, it may be possible to rationalize size in terms of

neural processing.

Acknowledgements. We thank P.P. Mitra and D.A. Clark for collaborative efforts; R. Jornsten for advice on

statistical analysis; M. Hau, P.R. Hof, S. Tavazoie, and M. Wikelski for stimulating discussions; and T.A. Barney for

excellent secretarial aid. M.J.B. is supported by the Howard Hughes Medical Institute. S.S.-H.W. is supported by

the Alfred P. Sloan Foundation.




19

Figure legends

Figure 1 Constancy of cerebellar volume fraction and variability of cerebral cortical volume

fraction across mammalian taxa. Cerebellar volume fractions ( F cbl ) are plotted at order level

except for the primates, which are divided into Strepsirhini (lemurs and lorises), Platyrrhini (New

World monkeys), Cercopithecoidea (Old World monkeys), and Hominoidea (apes and human).

F cbl values are indicated by solid symbols; cerebral cortical volume fractions ( F neo), by open

symbols. Error bars indicate standard deviation and the shaded bar indicates mean +/- s.d. for the

pooled data. Statistically significant differences between an individual taxon and the entire

group are identified in bold (asterisk, P<0.02; double asterisk, P<0.01). Modified from [Clark et

al., 2001].

Figure 2 Clustering of cerebrotypes by taxon, plotted using multidimensional scaling. (a)

Variation in the 11-component cerebrotype for insectivores, tree shrews, and primates. (b)

Insectivores divided into taxa. (c) Primates divided into taxa. Modified from [Clark et al., 2001].

Figure 3 Relative variation in brain components compared with body size. Volumes of

telencephalon (filled symbols) and non-telencephalic brain components (open symbols) as a

function of brain size. Brain and body measurements came from published data on 33 species

[Nottebohm, 1981; Rehkämper et al., 1991; DeVoogd et al., 1993; Boire and Baron, 1994;

Harper and Turner, 2000]. The telencephalon is significantly more variable than non-

telencephalic brain components.




20

Figure 4 Analysis of telencephalic volume fraction and social complexity in birds. (a)

Relationships between avian behavioral characteristics and telencephalic volume fraction ( F tel ).

Boxplots of F tel values indicate median (midline), 25% and 75% quartiles (box edges), minimum

and maximum values (whisker ends), and outliers residing more than 1.5 times the box length

beyond the upper or lower quartile (dots). The numbers below boxes indicate number of species

in each category. (b) Regression tree showing optimal separation of F tel . For partitioning, all six

behavioral characteristics from (a) were used except nomadic migration, for which the data set

contained only two parrots, M. undulatus and P. menstruus. The regression tree gave splits by

social structure at the first and second nodes that accounted for major components of variance in

F tel . Values at the nodes in bold indicate the fraction of total variance in F tel accounted for by the

node. The branch length is proportional to the amount of variance reduction. For each leaf,

percentages give mean F tel and (number of species in the leaf). The minimum deviance

[Venables and Ripley, 1994] was set at 0.01 and the minimum leaf size was set at 3 species.

Figure 5 Multidimensional scaling of avian brain architecture. Multidimensional scaling of

avian cerebrotypes was performed using an iterative bootstrapping procedure [Clark et al., 2001];

the figure shows the mapping with lowest root-mean-squared (rms) error in 10 or more trials.

The color-coded regions represent minimum convex polygons for each group. Scale bars

represent 10% Euclidean distance. a, Taxonomic clustering of bird cerebrotypes. Ga.,

Galliformes: pheasants/guineafowl/turkeys; Ap., Apodiformes: swifts/hummingbirds; Co.,

Columbiformes: pigeons/doves; Ch., Charadriiformes: shorebirds; Ps., Psittaciformes: parrots.

The rms fractional error is 7.2%. b, Social clustering of same bird cerebrotypes. c, Social




21

clustering when telencephalon was excluded from cerebrotype data and the remaining brain

regions renormalized. The rms fractional error for the renormalized scaling is 11.1%.

Figure 6 Comparison of dinosaur telencephalic volume fractions with living forms. Dinosaur

telencephalic volume fractions F tel were estimated from CT scan reconstructions of endocasts of

four dinosaurs [Larsson et al., 2000]. Boxplots are calculated from bird data as cited in the text,

from 35 reptiles [Platel, 1976], and 70 fishes [Ridet and Bauchot, 1990].




22

REFERENCES

Allman, J. and A. Hasenstaub (1999) Brains, maturation times, and parenting. Neurobiol. Aging, 20:

447-454.

Armstrong, E. (1983) Relative brain size and metabolism in mammals. Science, 220: 1302-1304.Barton, R.A., A. Purvis and P.H. Harvey (1995) Evolutionary radiation of visual and olfactory brainsystems in primates, bats and insectivores. Philos. Trans. R. Soc. Lond. B Biol. Sci., 348: 381-392.

Barton, R.A. and R.I.M. Dunbar (1997) Evolution of the social brain. In Machiavellian intelligence II:evaluations and extensions(ed. by A.W. Whiten and R.W. Byrne), Cambridge University Press, NewYork, pp. 240-263.

Barton, R.A. and P.H. Harvey (2000) Mosaic evolution of brain structure in mammals. Nature, 405:

1055-1058.

Boersma, P.D. (1988) Magellanic penguins of Patagonia. Sphenisicid Penguin Newsletter, 1: 2-3.

Boire, D. and G. Baron (1994) Allometric comparison of brain and main brain subdivisions in birds. J.

Brain Res., 35: 49-66.

Braitenberg, V. (2001) Brain size and number of neurons: an exercise in synthetic neuroanatomy. J.Comput. Neurosci., 10: 71-77.

Brodsky, L.M., C.D. Ankney and D.G. Dennis (1988) The influence of male dominance on socialinteractions in black ducks and mallards. Animal Behav., 36: 1371-1378.

Bullock, T.H. (1993) How are more complex brains different? One view and an agenda for comparativeneurobiology. Brain Behav. Evol., 41: 88-96.

Capurro, A., E. Frere, M. Gandini, T. Holik, V. Lichtschein and P.D. Boersma (1988) Nest density and population size of Magellanic penguins (Spheniscus magellanicus) in Cabo Dos Bahías, Argentina. Auk,105: 585-588.

Changizi, M.A. (2001) Principles underlying mammalian neocortical scaling. Biol. Cybern., 84: 207-215.

Chatterjee, S. (1991) Cranial anatomy of a new Triassic bird from Texas. Philos. Trans. R. Soc. Lond. BBiol. Sci., 332: 277-346.

Clark, D.A., P.P. Mitra and S.S.-H. Wang (2001) Scalable architecture in mammalian brains. Nature,411: 189-193.

Cramp, S. (Ed.) (1994) Handbook of the birds of Europe, the Middle East and North Africa : the birds of

the Western Palearctic, Oxford University Press, Oxford.

Currie, P.J. and X.-J. Zhao (1993) A new troodontid (Dinosauria, Theropoda) braincase from the

Dinosaur Park Formation (Campanian) of Alberta. Can. J. Earth Sci., 30: 2231-2247.de Winter, W. and C.E. Oxnard (2001) Evolutionary radiations and convergences in the structuralorganization of mammalian brains. Nature, 409: 710-714.

del Hoyo, J., J. Elliott and J. Sargatal (Eds.) (1992) Handbook of the birds of the world, Lynx Edicions,Barcelona.

Desforges, M.F. and D.G.M. Wood-Gush (1975) A behavioural comparison of domestic and Mallardducks. Spatial relationships in small flocks. Animal Behav., 23: 698-705.




23

DeVoogd, T.J., J.R. Krebs, S.D. Healy and A. Purvis (1993) Relations between song repertoire size andthe volume of brain nuclei related to song: comparative evolutionary analyses amongst oscine birds.Proc. R. Soc. Lond. B Biol. Sci., 254: 75-82.

Douglas, R.J. and D. Marcellus (1975) The ascent of man: deductions based on a multivariate analysis of

the brain. Brain Behav. Evol., 11: 179-213.Dunbar, R.I.M. (1995) Neocortex size and group size in primates: a test of the hypothesis. J. Hum. Evol.,28: 287-296.

Edinger, T. (1926) The brain of Archaeopteryx. Annu. Magazine Nat. Hist., 18: 151-156.

Edinger, T. (1961) Fossil brains reflect specialized behaviour. World Neurology, 2: 934-941.

Edinger, T. (1964) Recent advances in paleoneurology. Prog. Brain Res., 6: 147-160.

Elston, G.N., R. Benavides-Piccione and J. DeFelipe (2001) The pyramidal cell in cognition: acomparative study in human and monkey. J. Neurosci., 21: RC163.

Elzanowski, A. (1991) New observations on the skull of Hesperornis with reconstructions of the bony

palate and otic region. Postilla, 207: 1-20.Elzanowski, A. and P.M. Galton (1991) Braincase of Enaliornis, an Early Cretaceous bird from England.J. Vert. Paleontology, 11: 90-107.

Faaborg, J. (1988) Ornithology: an ecological approach, Prentice-Hall, Englewood Cliffs, N.J.

Finlay, B.L. and R.B. Darlington (1995) Linked regularities in the development and evolution of mammalian brains. Science, 268: 1578-1584.

Finlay, B.L., R.B. Darlington and N. Nicastro (2001) Developmental structure in brain evolution. Behav.Brain Sci., 24: 263-278.

Fox, J.H. and W. Wilczynski (1986) Allometry of major CNS divisions: towards a reevaluation of

somatic brain-body scaling. Brain Behav. Evol., 28: 157-169.Goodwin, D. (1986) Crows of the world, Comstock Publishing Associates, London.

Gould, S.J. (1975) Allometry in primates, with emphasis on scaling and the evolution of the brain. In Approaches to primate paleobiology, Vol. 5 (ed. by F.S. Szalay), S. Karger, Basel, pp. 244-292.

Harper, E.J. and C.L. Turner (2000) Nutrition and energetics of the canary (Serinus canarius). Comp.Biochem. Physiol. B, 126: 271-281.

Harvey, P.H. and J.R. Krebs (1990) Comparing brains. Science, 249: 140-146.

Harvey, P.H. and M. Pagel (1991) The comparative method in evolutionary biology, Oxford UniversityPress, Oxford.

Hopson, J.A. (1977) Relative brain size and behavior in Archosaurian reptiles. Annu. Rev. Ecol. Syst.,8: 429-448.

Hunt, G.R. (2000) Human-like, population-level specialization in the manufacture of pandanus tools by New Caledonian crows Corvus moneduloides. Proc. R. Soc. Lond. B, 267: 403-413.

Hunt, G.R., M.C. Corballis and R.D. Gray (2001) Animal behaviour: laterality in tool manufacture bycrows. Nature, 414: 707.

Ingels, J. (1978) Notes on the Pionus parrots. Avicult. Magazine, 84: 196-198.




24

Jerison, H.J. (1969) Brain evolution and dinosaur brains. Am. Nat., 103: 575-588.

Jerison, H.J. (1973) Evolution of the brain and intelligence, Academic Press, New York.

Jerison, H.J. (1991) Brain size and the evolution of mind, American Museum of Natural History, New

York.Kinzey, W.G. (1997) New World primates: ecology, evolution, and behavior, Aldine de Gruyter, NewYork.

Larsson, H.C.E., P.C. Sereno and J.A. Wilson (2000) Forebrain enlargement among nonavian theropoddinosaurs. J. Vert. Paleontology, 20: 615-618.

Lefebvre, L., P. Whittle, E. Lascaris and A. Finkelstein (1997) Feeding innovations and forebrain size in birds. Animal Behav., 53: 549-560.

Lefebvre, L., A. Gaxiola, S. Dawson, S. Timmermans, L. Rosza and P. Kabai (1998) Feeding innovationsand forebrain size in Australasian birds. Behaviour, 135: 1077-1097.

Lowry, P. (1991) The blue-headed Pionus parrot. American Cage-Bird Magazine, 63: 5-10, 13-14.

Martin, R.D. and P.H. Harvey (1985) Brain size allometry: ontogeny and phylogeny. In Size and scalingin primate biology(ed. by W.L. Jungers), Plenum Press, New York, pp. 147-173.

Molnar, R.E. (1985) Alternatives to Archaeopteryx: a survey of proposed early or ancestral birds. In The beginning of birds(ed. by M.K. Heckht, J.H. Ostrom, G. Viohl and P. Wellnhofer), Freunde des Jura-Museums Eichstätt, Willibaldsburg, Eichstätt, pp. 209-217.

Nealen, P.M. and R.E. Ricklefs (2001) Early diversification of the avian brain:body relationship. J. Zool.(Lond.), 253: 391-404.

Nicolakakis, N. and L. Lefebvre (2000) Forebrain size and innovation rate in European birds: feeding,nesting and confounding variables. Behaviour, 137: 1415-1429.

Niewenhuys, R., H. ten Donkelaar and C. Nicholson (1998) The central nervous system of vertebrates,Springer, Berlin.

Nottebohm, F. (1981) A brain for all seasons: cyclical anatomical changes in song control nuclei of thecanary brain. Science, 214: 1368-1370.

Pepperberg, I.M. and M.V. Brezinsky (1991) Acquisition of a relative class concept by an African gray parrot ( Psittacus erithacus): discriminations based on relative size. J. Comp. Psych., 105: 286-294.

Pepperberg, I.M. (2002) In search of King Solomon's Ring: cognitive and communicative studies of grey parrots ( Psittacus erithacus). Brain Behav. Evol., 59: 54-67.

Platel, R. (1976) Analyse volumétrique comparée des principales subdivisions encéphaliques chez lesreptiles sauriens. J. Hirnforsch., 17: 513-537.

Portmann, A. (1946) Études sur la cérébralisation chez les oiseaux: I. Alauda, 14: 2-20.Portmann, A. (1947a) Études sur la cérébralisation chez les oiseaux: II. Les indices intra-cérébraux.Alauda, 15: 1-15.

Portmann, A. (1947b) Études sur la cérébralisation chez les oiseaux: III. Cérébralisation et modeontogénétique. Alauda, 15: 161-171.

Reader, S.M. and K.N. Laland (2002) Social intelligence, innovation, and enhanced brain size in primates. Proc. Natl. Acad. Sci. USA, 99: 4436-4441.




25

Rehkämper, G., H.D. Frahm and K. Zilles (1991) Quantitative development of brain structures in birds(Galliformes and Passeriformes) compared to that in mammals (Insectivores and Primates). Brain Behav.Evol., 37: 125-143.

Ridet, J.M. and R. Bauchot (1990) [Quantitative analysis of the teleost brain: evolutionary and adaptive

features of encephalization. II. Primary brain subdivisions]. J. Hirnforsch., 31: 433-458.Rogers, S.W. (1998) Exploring dinosaur neuropaleobiology: viewpoint computed tomography scanningand analysis of an Allosaurus fragilis endocast. Neuron, 21: 673-679.

Sereno, P.C. (1999) The evolution of dinosaurs. Science, 284: 2137-2147.

Sibley, C.G. and J.E. Ahlquist (1991) Phylogeny and classification of birds: a study in molecular evolution, Yale University Press, New Haven.

Stephan, H., H. Frahm and G. Baron (1981) New and revised data on volumes of brain structures ininsectivores and primates. Folia Primatologica, 35: 1-29.

Stevens, C.F. (2001) An evolutionary scaling law for the primate visual system and its basis in cortical

function. Nature, 411: 193-195.Thompson, D.W. (1942) On growth and form, The University Press, Cambridge.

Venables, W. and B. Ripley (1994) Modern applied statistics with S-PLUS, Springer-Verlag, New York.

Wang, S.S.-H., P.P. Mitra and D.A. Clark (2002) How did brains evolve? Nature, 415: 135.

Weir, A.A., J. Chappell and A. Kacelnik (2002) Shaping of hooks in New Caledonian crows. Science,297: 981.

West, G.B., J.H. Brown and B.J. Enquist (1999) The fourth dimension of life: fractal geometry andallometric scaling of organisms. Science, 284: 1677-1679.

Whiten, A. and R.W. Byrne (1988) The Machiavellian intellect hypotheses. In Machiavellian

Intelligence(ed. by R.W. Byrne and A. Whiten), Oxford University Press, Oxford, pp. 1-9.Wittmer, L.M. (1990) The craniofacial air sac system of Mesozoic birds (Aves). Zool. J. Linnean Soc.,100: 327-378.

Wyles, J.S., J.G. Kunkel and A.C. Wilson (1983) Birds, behavior, and anatomical evolution. Proc. Natl.Acad. Sci. USA, 80: 4394-4397.

Zann, R.A. (1996) The zebra finch: a synthesis of field and laboratory studies, Oxford University Press,Oxford.

Zhang, K. and T.J. Sejnowski (2000) A universal scaling law between gray matter and white matter of cerebral cortex. Proc. Natl. Acad. Sci. USA, 97: 5621-5626.



0.8

0.6

0.4

0.2

0

F r a c t i o n

o f t o t a l b r a i n

S o r i c o m o r p h a

H o m i n

o i d e a

C e r c o p i t h e c

o i d e a

S

i r e n i a

P l a t y r r h i n i

C e

t a c e a

T a r s i u s

A r t i o d a c t y l a

S t r e p s i r h i n i

M e g a c h i r o

p t e r a

S c a n d e n t i a

M i c r o c h i r o p

t e r a

M a c r o s c e l i d i d a e

C h r y s o c h l o

r i d a e

E r i n a c e o m o r p h a

Neocortex

Cerebellum

* **

R o d e n t i a

M a r s u

p i a l i a

L a g o m o r p h a

C a r n i v o r a

21 117 232 471522523 414122 1 8 15 12 6

E d e n t a t a

Burish, Kueh and Wang

Figure 1



Insectivores

PrimatesTree shrews

Lemurs/ lorises/ tarsier

Monkeys/ apes

10% Euclideandistance

Tarsier

Great apes

Old World monkeysNew World monkeys

Lemurs/lorises

Pan

Homo

Gorilla Pongo

Tenrecs

Tree shrewsSolenodon

Golden moles

Hedgehogs

Elephant shrewsShrews

Moles

Potamogale velox Limnogale mergulus

Micropotamogale lamottei

a

b

c

1 0 % E

u c l i d e a n d i s t a n c e

At.Lag.

Ceb.

Sai.


Figure 2



4

3

2

1 2 3 4

Telencephalon

Non-telencephalic

components

Log (body mass) (g)10

L o g

( b r a i n v o l u m e ) ( m m

)

1 0

3


Figure 3



|

Aerial 48.6% (4)

Terrestrial/

Flightless

No mimicry 58.0% (7)

Learned

Vocalizations

Transactional 71.6% (6)

Social structure

0.577

Social structure

0.135

Vocalizations

0.032

Flight

0.044

Solitary/Covey/Colonial

Col.

Sol./

Cov.

54.2% (13)

63.9% (

a b

0.45

0.50

0.55

0.60

0.65

0.70

0.75

Telencephalic

fraction of total brain

Social

Structure

Eating

HabitsMigration Flight Mating

TypeSong

T r a n s a c t i o n a l

S o l i t a r y

C o v e y

C o l o n i a l

H e r b i v o r o u s

O m n i v o r o u s

C a r n i v o r o u s

S e d e n t a r y

M i g r a t o r y

N o m a d i c

T e r r e s t r i a l / F l i g h t l e s s

A e r i a l

M o n o g a m o u s

P o l y g y n o u s

N o m i m i c r y

L e a r n e d v o c a l i z a t i o n s

613104 21120 9249195 1914 825


Figure 4



10%

bTransactional

Colonial

Covey

Solitary

10%

a

Ga.

Ps.

Ap.

Ch.

Co.

c

TransactionalColonial

Covey

Solitary

10%


Figure 5



0.5

0.6

0.7

T e l e n c e p h a l i c f r a c t i o n o f t o t a l

b r a i n

Birds

T r a n s a c t i o n a l

S o l i t a r y

C o v e y

C o l o n i a l

613104

0.4

0.3

0.2

Archaeopteryx

Tyrannosaurus

35

0.8

R e p t i l e s

0.1

0

Cold-blooded

vertebrates

Carcharodont.

Allosaurus

72

F i s h e s

Burish, Kueh and WangFigure 6

Burish Wang03 BBE Preprint

Documents

Transcript of Burish Wang03 BBE Preprint