Professional Documents
Culture Documents
Introduction
Edmund T. Rolls
DOI:10.1093/acprof:oso/9780199232703.003.0001
Page 1 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 2 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 3 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 4 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 5 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 6 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
(p.4)
1.2
Neurons in
the brain,
and their
Page 7 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
1.3 A
formalism
for
(p.5)
Page 8 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
where j indicates that the sum is over the C input axons (or
connections) indexed by j to each neuron. The multiplicative form
here indicates that activation should be produced by an axon only if
it is firing, and depending on the strength of the synapse wij from
input axon j onto the dendrite of the receiving neuron i. Equation
1.1 indicates that the strength of the activation reflects how fast
the axon j is firing (that is xj), and how strong the synapse wij is. The
sum of all such activations expresses the idea that summation (of
synaptic currents in real neurons) occurs along the length of the
dendrite, to produce activation at the cell body, where the
activation hi is converted into firing yi. This conversion can be
expressed as
(1.2)
(p.6)
Page 9 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
which
indicates that
the firing rate
is a function (f)
of the
postsynaptic
activation. The
function is
called the
activation
function in this
case. The
function at its
simplest could
be linear, so
that the firing
rate would be
proportional to
the activation
(see Fig. 1.3).
Real neurons
have
thresholds,
with firing
occurring only
if the
activation is
above the
threshold. A
threshold
linear
activation
function is
shown in Fig.
is small, the output goes smoothly and
1.3. This has
slowly from 0 to 1 as hi goes from to
been useful in
+. If is large, the curve is very steep,
formal analysis
and approximates a binary threshold
of the
activation function. (d) Binary threshold.
properties of
neural
networks.
Neurons also
have firing rates that become saturated at a maximum rate, and we
could express this as the sigmoid activation function shown in Fig.
1.3c. Another simple activation function, used in some models of
neural networks, is the binary threshold function (Fig. 1.3d), which
Page 10 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 11 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
A property
implied by
equation 1.1
is that the
postsynaptic
membrane is
electrically
short, and so
summates its
inputs
irrespective of
where on the
dendrite the
input is
received. In
real neurons,
the
transduction
of current
into firing
frequency
(the analogue
of the transfer
function of
equation 1.2)
is generally
studied not
with synaptic
inputs but by
applying a
steady
Page 12 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
(p.8)
Page 13 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 14 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
(p.9)
Page 15 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
could be weak
or strong) to
the cells. The
associativity
arises because
it is only when
sufficient
activation of
the
postsynaptic
neuron to
exceed the
threshold of
NMDA receptors (see below) is produced that any learning can
occur. The two weak inputs summate to produce sufficient
depolarization to exceed the threshold. This associative property is
shown very clearly in experiments in which LTP of an input to a
single cell only occurs if the cell membrane is depolarized by
passing current through it at the same time as the input arrives at
the cell. The depolarization alone or the input alone is not sufficient
to produce the LTP, and the LTP is thus associative. Moreover, in
that the presynaptic input and the postsynaptic depolarization must
occur at about the same time (within approximately 500 ms), the
LTP requires temporal contiguity. LTP is also synapse-specific, in
that, for example, an inactive input to a cell does not show LTP
even if the cell is strongly activated by other inputs (Fig. 1.5b, input
B).
(p.10)
Page 16 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 17 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
acid), activated
by the
glutamate
released by the
presynaptic
terminals. The
NMDA
receptor
channels are
normally
blocked by
Mg2+, but
when the cell
is strongly
depolarized by
strong tetanic
stimulation of
the type
necessary to
induce LTP,
the Mg2+ block
is removed, and Ca2+ entering via the NMDA receptor channels
triggers events that lead to the potentiated synaptic transmission
(see Fig. 1.6). Part of the evidence for this is that NMDA
antagonists such as AP5 (D-2-amino-5-phosphonopentanoate) block
LTP. Further, if the postsynaptic membrane is voltage clamped to
prevent depolarization by a strong input, then LTP does not occur.
The voltage-dependence of the NMDA receptor channels
introduces a threshold and thus a non-linearity that contributes to a
number of the phenomena of some types
(p.11)
of LTP, such as
Page 18 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
There are a
number of
possibilities
about what
change is
triggered by
the entry of
Ca2+ to the
postsynaptic
cell to
mediate LTP.
One
possibility is
that somehow
a messenger
reaches the
presynaptic
terminals
from the
postsynaptic
membrane
and, if the
terminals are
active, causes
them to
Bliss 1987.)
release more
transmitter in
future
whenever
they are activated by an action potential. Consistent with this
possibility is the observation that, after LTP has been induced,
more transmitter appears to be released from the presynaptic
endings. Another possibility is that the postsynaptic membrane
changes just where Ca2+ has entered, so that AMPA receptors
become more responsive to glutamate released in future.
Consistent with this possibility is the observation that after
LTP, the postsynaptic cell may respond more to locally applied
glutamate (using a microiontophoretic technique).
The rule that underlies associative LTP is thus that synapses
connecting two neurons become stronger if there is
Page 19 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 20 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
(p.12)
neuron is strongly
Page 21 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
(p.13)
Page 22 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 23 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
where yi is the firing rate of the ith neuron in the set of N neurons
(Treves and Rolls 1991). This is referred to as the population
sparseness, and measures of sparseness are considered in detail in
Section C.3.1. A low value of the sparseness a p indicates that few
neurons are firing for any one stimulus.
(p.14)
Page 24 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 25 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
(p.15)
Appendix C). A similar result has been found for the encoding
of position in space by the primate hippocampus (Rolls,
Treves, Robertson, Georges-Franois and Panzeri 1998).
It is particularly important that the information can be read
from the ensemble of neurons using a simple measure of the
similarity of vectors, the correlation (or dot product, see
Appendix A) between two vectors. The importance of this is
that it is essentially vector similarity operations that
characterize the operation of many neuronal networks (see
Appendix B). The neurophysiological results show that both
the ability to reflect similarity by vector correlation, and the
utilization of exponential coding capacity, are properties of
real neuronal networks found in the brain.
To emphasize one of the points being made here, although the
binary encoding used in the 8-bit vector described above has
optimal capacity for binary encoding, it is not optimal for
vector similarity operations. For example, the two very similar
numbers 127 and 128 are represented by 01111111 and
10000000 with binary encoding, yet the correlation or bit
overlap of these vectors is 0. The brain, in contrast, uses a
code that has the attractive property of exponentially
increasing capacity with the number of neurons in the
representation, though it is different from the simple binary
Page 26 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 27 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
(p.16)
outputs, which
are often
crucial to the
way in which
the problem is
solved. In
addition, they
may use
learning
algorithms that
are really too
powerful for
the brain to
perform, and
Fig. 1.7 Three network architectures that
therefore they
use local learning rules: (a) pattern
can be taken
association introduced with a single
only as a guide
output neuron; (b) pattern association
to how
network; (c) autoassociation network; (d)
cognitive
functions
competitive network.
might be
implemented
by neuronal
networks in the brain. In this book, we focus on more biologically
plausible neuronal networks.
Page 28 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 29 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 30 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 31 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Information
processing
along this
stream is
primarily
unimodal, as
shown by the
fact that
inputs from
other
modalities
(such as taste
or smell) do
not
anatomically
have
significant
inputs to
these regions, and by the fact that neurons in these areas
respond primarily to visual stimuli, and not to taste or
olfactory stimuli, etc. (Rolls 2000a, Baylis, Rolls and
(p.19)
Page 32 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 33 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Leonard 1987,
Ungerleider
1995, Rolls and
Deco 2002).
The
representation
built along this
pathway is
mainly about
what object is
being viewed,
independently
of exactly
where it is on the retina, of its size, and even of the angle with
which it is viewed (see Chapter 4 and Rolls and Deco (2002)), and
for this reason it is frequently referred to as the what visual
pathway. The representation is also independent of whether the
object is associated with reward or punishment, that is the
representation is about objects per se (Rolls, Judge and Sanghera
1977). The computation that must be performed along this stream
is thus primarily to build a representation of objects that shows
invariance. After this processing, the visual representation is
interfaced to other sensory systems in areas in which simple
associations must be learned between stimuli in different
modalities (see Chapters 2 and 3). The representation must thus be
in a form in which the simple generalization properties of
associative networks can be useful. Given that the association is
about what object is present (and not where it is on the retina), the
representation computed in sensory systems must be in a form that
allows the simple correlations computed by associative networks to
reflect
(p.20)
Page 34 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
similarities
between
objects, and
not between
their positions
on the retina.
The way in
which such
invariant
sensory
Page 35 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
representations
could be built
in the brain is
the subject of
Chapter 4.
Page 36 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
The ventral
visual stream
converges
with other
mainly
unimodal
information
processing
streams for
taste,
olfaction,
touch, and
hearing in a
number of
areas,
particularly
the amygdala
(p.21)
Page 37 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
and
orbitofrontal
cortex (see
Figs. 1.8, 1.9,
and 1.10).
These areas
appear to be
necessary for
learning to
associate
sensory stimuli
with other
Fig. 1.11 Lateral view of the macaque
reinforcing
brain showing the connections in the
(rewarding or
dorsal or where visual pathway from V1
punishing)
to V2, MST, LIP, VIP, and parietal cortex
stimuli. For
area 7a, with some connections then
example, the
reaching the dorsolateral prefrontal
amygdala is
cortex. Abbreviations as in Fig. 1.9. FEF involved in
learning
frontal eye field.
associations
between the
sight of food
and its taste. (The taste is a primary or innate reinforcer.) The
orbitofrontal cortex is especially involved in rapidly relearning
these associations, when environmental contingencies change (see
Rolls (2005) and Rolls (2000e)). They thus are brain regions in
which the computation at least includes simple pattern association
(e.g. between the sight of an object and its taste). In the
orbitofrontal cortex, this association learning is also used to
produce a representation of flavour, in that neurons are found in
the orbitofrontal cortex that are activated by both olfactory and
taste stimuli (Rolls and Baylis 1994), and in that the neuronal
responses in this region reflect in some cases olfactory to taste
association learning (Rolls, Critchley, Mason and Wakeman 1996b,
Critchley and Rolls 1996b). In these regions too, the representation
is concerned not only with what sensory stimulus is present, but for
some neurons, with its hedonic or reward-related properties, which
are often computed by association with stimuli in other modalities.
For example, many of the visual neurons in the orbitofrontal cortex
respond to the sight of food only when hunger is present. This
probably occurs because the visual inputs here have been
associated with a taste input, which itself in this region only occurs
to a food if hunger is present, that is when the taste is rewarding
(see Chapter 3, Rolls (2005), and Rolls (2000e)). The outputs from
these associative memory systems, the amygdala and orbitofrontal
Page 38 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
(p.22)
and Haxby (1994); and Rolls and Deco (2002)). This where
pathway for primate vision is involved in representing where
stimuli are relative to the animal (i.e. in egocentric space), and
the motion of these stimuli. Neurons here respond, for
example, to stimuli in visual space around the animal,
including the distance from the observer, and also respond to
optic flow or to moving stimuli. Outputs of this system control
eye movements to visual stimuli (both slow pursuit and
saccadic eye movements). These outputs proceed partly via
the frontal eye fields, which then project to the striatum, and
then via the substantia nigra reach the superior colliculus
(Goldberg 2000). Other outputs of these regions are to the
dorsolateral prefrontal cortex, area 46, which is important as a
short-term memory for where fixation should occur next, as
shown by the effects of lesions to the prefrontal cortex on
saccades to remembered targets, and by neuronal activity in
this region (Goldman-Rakic 1996). The dorsolateral prefrontal
cortex short-term memory systems in area 46 with spatial
information received from the parietal cortex play an
important role in attention, by holding on-line the target being
attended to, as described in Chapters 6, 8 and 9.
The hippocampus receives inputs from both the what and the
where visual systems (Chapter 2). By rapidly learning
associations between conjunctive inputs in these systems, it is
able to form memories of particular events occurring in
particular places at particular times. To do this, it needs to
store whatever is being represented in each of many cortical
areas at a given time, and later to recall the whole memory
from a part of it. The types of network it contains that are
involved in this simple memory function are described in
Chapter 2.
Page 39 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
(p.23)
Page 40 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 41 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
cells are typically small pyramidal cells (Lund 1984). (It may
be noted here that spiny stellate cells and small pyramidal
cells are similar in many ways, with a few main differences
including the absence of a major apical dendrite in a spiny
stellate which accounts for its non-pyramidal, star-shaped,
appearance; and for many spiny stellate cells, the absence of
an axon that projects outside its cortical area.) The terminals
presumably make synapses with these small pyramidal cells,
and also presumably with the dendrites of cells from other
layers, including the basal dendrites of deep layer 3 pyramidal
cells (see Fig. 1.13).
The axons of the superficial (layer 2 and 3) pyramidal cells
have collaterals and terminals in layer 5 (see Fig. 1.13), and
synapses are made with the dendrites of the layer 5 pyramidal
cells (Martin 1984). The axons also typically project out of that
cortical area, and on to the next cortical area in sequence,
where they terminate in layer 4, forming the forward corticocortical projection. It is also from these pyramidal cells that
projections to the amygdala arise in some sensory areas that
are high in the hierarchy (Amaral, Price, Pitkanen and
Carmichael 1992).
The axons of the layer 5 pyramidal cells have many collaterals
in layer 6 (see Fig. 1.1), where synapses could be made with
the layer 6 pyramidal cells (based on indirect evidence, see
Fig. 13 of Martin (1984)), and axons of these cells typically
leave the cortex to project to subcortical sites (such as the
striatum), or back to the preceding cortical area to terminate
in layer 1. It is remarkable that there are as many of these
backprojections as there are forward
(p.24)
connections
Page 42 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 43 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
(p.25)
Page 44 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Andersen,
Dingledine,
Gjerstad,
Langmoen and
Laursen (1980)
for
hippocampal
pyramidal
cells), so that
the effect of
shunting is to
produce
division (i.e. a
multiplicative
reduction) of
the excitatory
inputs received
by the cell, and
not just to act
by subtraction
(see further
Bloomfield
(1974), Martin
(1984),
Douglas and
Martin (1990)).
Thus, when
modelling the
normalization
of the activity
of cortical
pyramidal
cells, it is
common to
include
Page 45 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
(p.26)
Page 46 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 47 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
(p.27)
20,00040,000/
mm3
Neuronal composition:
Pyramidal
75%
Spiny stellate
10%
15%
Synaptic density
8 108/mm3
9,000
2,000
10 mm
18,000
2,000
300
400 m/mm3
3,200 m/mm3
Page 48 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
2 mm
Cortical area
human (assuming 3 mm for cortical
thickness)
300,000 mm2
30,000 mm2
thickness)
rat (assuming 2 mm for cortical
thickness)
300 mm2
for example Rolls and Tovee (1995b), Rolls, Treves, Tovee and
Panzeri (1997d), and Franco, Rolls, Aggelopoulos and Jerez (2007)),
which means that even any small additional input may produce
some spikes sooner than would otherwise have occurred, because
some of the neurons may be very close to a threshold for firing. It
might also show some of the autoassociative retrieval of
information typical of autoassociation networks, if the synapses
between the nearby pyramidal cells have the appropriate (Hebbian)
modifiability. In this context, the value of 0.1 for the probability of a
connection between nearby neocortical pyramidal cells is of
interest, for the connection probability between hippocampal CA3
pyramidal is approximately 0.020.04, and this is thought to be
sufficient to sustain associative retrieval (see Appendix B and Rolls
and Treves (1998)).
Page 49 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
(p.28)
Page 50 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
(p.29)
Page 51 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
were di-(or
poly-)
synaptically
activated by
stimulation of
the afferents
from the
thalamus, but
also had S-type
receptive
fields.
Inputs could
reach the
layer 5
pyramidal
cells from the
pyramidal
cells in layers
2 and 3, the
which the
layer 5
pyramidal
cells have
axons of
which ramify
extensively in
layer 5, in
widespread
basal
dendrites (see
Fig. 1.1), and also perhaps from thalamic afferents. Many
layer 5 pyramidal cells are di-or trisynaptically activated by
stimulation of the thalamic afferents, consistent with them
receiving inputs from monosynaptically activated deep layer 3
pyramidal cells, or from disynaptically activated pyramidal
cells in layer 2 and upper layer 3 (Martin 1984). Interestingly,
many of the layer 5 pyramidal cells had C-type receptive
fields, that is they did not have distinct on and off regions, but
did respond with orientation tuning to elongated visual stimuli
(Martin 1984) (see Rolls and Deco (2002)).
Page 52 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 53 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 54 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
(p.31)
Page 55 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
(p.32)
Page 56 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
in the
hierarchy. The
reasons for the
asymmetry
(including the
need for
Page 57 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
One point made in Figs. 1.8, 2.1 and 3.3 is that the amygdala
and hippocampus are stages of information processing at
which the different sensory modalities (such as vision,
hearing, touch, taste, and smell for the amygdala) are brought
together, so that correlations between inputs in different
modalities can be detected in these regions, but not at prior
cortical processing stages in each modality, as these cortical
processing stages are mainly unimodal. As a result of bringing
together any two sensory modalities, significant
correspondences between the two modalities can be detected.
One example might be that a particular visual stimulus is
associated with the taste of food. Another example might be
that another visual stimulus is associated with painful touch.
Thus at these limbic (and orbitofrontal cortex, see Chapter 3)
stages of processing, but not before, the significance of, for
example, visual and auditory stimuli can be detected and
signalled. Sending this information back to the neocortex
could thus provide a signal that indicates to the cortex that
information should be stored. Even more than this, the
backprojection pathways could provide patterns of firing that
could help the neocortex to store the information efficiently,
one of the possible functions of backprojections within and to
the neocortex considered next.
1.11.2 Learning
(p.33)
Page 58 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 59 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
(p.34)
would be
Page 60 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 61 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
(p.35)
Page 62 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
Page 63 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016
Introduction
backward connected cortical areas could also be used in shortterm memory functions (see Section 5.1), to implement the
types of short-term memory effect described in Appendix B
and Chapter 5. Such connections could also be used to
implement a trace learning rule as described in Chapter 4.
With this overview, it is now time to consider memory, object
recognition, attention, and decision-making systems in the
brain.
Page 64 of 64
PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2015. All
Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a
monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber:
Chinese University of Hong Kong; date: 10 April 2016