Digitized Fossil Brains
Harry Jerison <http:\\ hjerison. Bol. Ucla.edu>
1. Paleoneurology: Fossil Brains
2. Digitization: laser and other scanning
3. Measurement: surface area, volume
4. Encephalization and Neocorticalization
5. Language (speculation)
This is a fossil “brain,” an endocranial cast
from a small late Eocene artiodactyl about
37 million years old. Its “brain” volume is
about 10 ml; the animal was about the size
of a living cat. You recognize olfactory
bulbs, forebrain, and hindbrain on this rock.
Laser scans of the endocast and
measurements on this Bathgenys reevesi.
(Green marked“neocortex” by laser software)
Artist’s drawing of the oreodont Merycoidodon
culbertsoni, a relative, presumably a
descendant, of a few million years later.
Imagine it cat-size.
Paleoneurology (Fossil “Brains”)
The slides so far cover the elements of paleoneurology
applied to a particular species, Bathygenys reevesi.
Endocasts are good enough pictures of brains that
one can treat them as the brains of particular
animals. They are fossil finds, either from skulls or
from natural endocasts like the one I showed you.
Digitization converts the physical endocast into a
computer image that can be measured. The most
important measurement on these fossils is of
surface area. Before showing you why this is
especially important I must show you some human
endocasts, about which you must think a bit harder.
Two human endocasts (laser scan, left, and mri
scan, right). What’s wrong? Poor image of the
Sylvian fissure -- no good data for guesses on
language. Next: the living human brain.
Left and right human hemispheres (Wisconsin
Brain Collection 69-314). Note left Sylvian is a
bit longer than right. Heschl’s gyrus (“language
area”) is buried within posterior left fissure.
3-D Digitizing Laser Scanning
You saw a digitized scan before
3-D Digitizing Laser Scanning
“Rhinal fissure” is outlined (blue) on scan.
A second example compares the endocast of a
late Eocene prosimian Adapis parisiensis with the
brain of a living galago Galago senegalensis.
These were both about 10 ml in volume. I
marked “neocortex” in green on the fossil. You
can see a bit of rhinal fissure on galago.
Laser scanning: The scanner is adjacent to my
computer monitor, which shows the image of the
Adapis endocast on the screen -- “neocortex” is
marked green. The picture at the right shows the
specimen on the scanner platform. (The disarray of
skulls and endocasts is normal in my workplace.)
The “Cyberware”scanner platform with the endocast of
Adapis parisiensis in place. Images in the scanner to
the left of the platform are of laser-optics prisms.
My scan of the Eocene prosimian Adapis
parisiensis is FMNH 59259. Galago is from
Wisconsin (61-686 ), an invaluable resource for
comparisons of living brains.
Measurement : Cortical
surface area and
This graph is from my James
Arthur Lecture on “Brain Size
and the Evolution of Mind”
(American Museum of
Natural History, 1991). A few
of its 50 species are named to
indicate their great variety.
The correlation of 1.00 to two
significant figures indicates
the strong relationship: brain
size is an almost perfect
between-species estimator of
cortical surface area. Surface
area is an excellent estimator
(between-species) of the total
number of neurons in a brain,
hence of processing capacity.
Neocortex and Mind
Neuroscientists have generally assumed that mind
is a consequence of cortical activity, in particular
of neocortex. To determine more about the
evolution of neocortex the first issue was to
estimate neocortical size in endocasts of fossil
mammals. This is surprisingly easy, because the
rhinal fissure is a visible boundary line in most
endocasts. The uniformitarian hypothesis
assumes it is true for fossil as well as living brains.
In living mammals neocortex is cortex dorsal to the
rhinal fissure as is probably true in fossils.
Rhinal fissure as ventral boundary to
neocortx: Armadillo brain and cross-section
(Wisc 60-465) showing border of neocortex.
Rhinal fissure is partly
hidden in most primate
endocasts. The slide has
human and Adapis at the
left, which you have seen
before, and an endocast and
brain of the mandrill, Papio
sphinx. The adapid and
monkey are shown
ventrolaterally as well as
dorsally, but the mandrill
brain from the Wisconsin
collection is shown only
laterally. My green marking
of left neocortex is manual
and only approximate. Both
the human and mandrill
neocortex are about 80 % of
the brain’s surface area. Human endocast 1370 ml;
The Eocene adapid’s is mandrill 132 ml; Eocene adapid, 8.3 ml.
about 65 %.
Encephalization • Encephalization is measured by
and brain size relative to its expected
Neocorticalization size. In this example, an early
A, rendered endocast. Eocene equoid,“Hyracotherium.”
B, tesselations (voxels)
for the scan. C,
Accurate scale model
that had been scanned.
Endocast volume is 24.1
ml. Scaling up the model
leads to a body size
estimate of 10.7 kg.
Digitization provides the
best guess on body size
Encephalization in Amniotes
(Next slide places living and fossil horses in the
mammal convex polygon.)
Encephalization of “Hyracotherium” 1
(points are living and fossil equids)
“Hyracotherium” is point 1 at
left. Dashed line is average
living mammal allometry: Y =.05
X.74. Solid line: Y = 0.12 X2/3.
(T is a condylarth endocast
misidentified as an equoid.)
Digitization is not a major
contribution for encephalization,
which can be estimated from
routine brain and body weights
determined as in the past.
Digitization is the only way to
on the irregular surface of a
brain or endocast. It is the
ratio of surface area of
neocortex to total surface
area of the brain or
To return to an early slide in which the measurements were
shown, the idea is to measure total surface area and then the total
neocortical area. For technical reasons I eliminate olfactory bulb
area from total endocast area. The neocorticalization ratio is 9.4/
31.5 = 30%. This is done for all of my 150 or so specimens. A
preliminary graph shows the evidence clearly on the next slide.
Preliminary Result Mammals
There are three Eocene
prosimians and two Plio-
Pleistocene australopithecines in
this sample. Living primates
included two humans and a
mangabey at the same
The main result is that taxa
differed. Primates have always
been more neocorticalized than
other mammals, marsupials
always less. Over the 60 million
year span there was an average
increase of neocorticalization at
about 5% per 10 million years.
There may be an error in the ordinate in the previous
(preliminary) slide. Here is a later graph with more species.
DISCUSSION: Why did neocorticalization increase (even though it's only
5% per 10 million years). There was evidently some selective advantage
to more neocortex. How does language fit in? In addition to its role in
communication, language is a uniquely human cognitive trait. Years ago
Joe Bogen, who did the cutting that showed the left brain and right
brain as different, told me that as a neurosurgeon he was especially
careful about cutting in ventral temporal lobe. It was that region rather
than Heschl's gyrus or Wernicke's area that worried him the most.
Removing it had terrible effects on language. Our view of the precise
localization of language is part of our general disposition to find
localized centers. Language certainly has these, but like other cognitive
controls it is dispersed through much of the neocortex. Enlarged
neocortex occurs to the same proportional change in primates. It's only
our very large brain, the absolute size of our proportion, that is so
remarkable. I would not look for specific areas in the fossil record of the
hominine brain for a localized language area. The increase in the
hominine neocortex really explains the changing cognitive capacities.
I have too many colleagues to thank for
access to their collections. The Field
Museum of Natural History in Chicago
and the late Len Radinsky, who left his
collection of endocasts to the Museum,
deserve special thanks. The Cyberware
Corporation in California provided the
apparatus that supports my scanning.