Neurobiology 303 -- Chapter 23 Outline
Developmental Plasticity
development of the vertebrate brain is intrinsic in that it is directed by the
genes, which code for the differentiation of neuroblasts and for the
expression of markers and trophic factors that guide axon growth and
establish synaptic connections
this aspect of development is termed "activity independent"
however, brain development can be influenced by neural activity, either
generated spontaneously or driven by sensory input from the
environment, which can plastically modify developing connections
this aspect of development is termed "activity dependent"
development of the visual system involves both activity independent and
activity dependent processes
studies of development in the visual system have focused on the
formation of topographic projections between regions
topographic projections in the visual system include the following:
retina to optic tectum (or superior colliculus in mammals)
retina to lateral geniculate nucleus (LGN)
LGN to primary visual cortex
primary to higher-order parts of visual cortex
these topographic projections are initially established through activity
independent mechanisms (e.g. neurotrophic factors released by
target structures)
but the initial projections are course, because the axons branch
extensively in the target structure and the large amount of
overlap makes the topographic map very imprecise
after birth, stimulation of the eye by light causes activity in the visual
system, and this causes the projecting axon branches to narrow,
leading to finer projections that confer greater visual acuity
development of the fine-grained map is activity dependent --
elimination of activity in the visual system, by crushing and
poisoning the optic nerve, prevents formation of the fine-
grained map but preserves the rough projections
other activity dependent developmental features include:
segregation of projections into eye specific layers in the LGN
formation of ocular dominance columns in the visual cortex
the mammalian LGN is a layered structure in which adjacent layers receive
projections from different eyes
the projection of retinal ganglion cells to LGN is activity independent,
but the formation of eye specific layers is activity dependent
experiments involve poisoning the retina with TTX, which eliminates
all action potential generation in the optic nerve
treatment with TTX does not affect the projection from retina to LGN,
but it completely abolishes the formation of eye specific layers
the formation of eye specific layers requires a specific pattern of
activity from the retina -- just random activity is not enough
calcium imaging experiments reveal that coherent waves of activity
regularly sweep over the retina -- this provides a temporally
structured retinal input to the LGN that is necessary for the
formation of eye-specific layers
the waves are generated independently at different times and at
different places in the two eyes
this causes the activity in nearby cells in the same retina to be
temporally correlated, but the activity in retinal cells from
opposite eyes will be uncorrelated
activity dependent connection formation usually involves coincident
activity in the pre- and post-synaptic neurons (retinal ganglion
cells and LGN cells in this case)
correlated activity in retinal ganglion cells from one eye activates a
subset of the LGN cells -- those on which they converge
the synapses between those retinal ganglion cells and the activated
LGN cells becomes strengthened
correlated activity in retinal ganglion cells from the other eye activates
a separate subset of LGN cells, and the synapses between those
other retinal ganglion and LGN cells becomes strengthened
eye-specific layers in the LGN appear to form in this way from the
correlated activity of retinal ganglion cells from either eye
autoradiography -- a histological technique based on radioactivity
ordinary molecules, like amino acids or hormones, can be made
radioactive by substituting some of their hydrogen atoms for
tritium, which decays by emitting a proton and a beta particle
the beta particle (high energy electron) can expose photographic film
after treating the brain with radioactive molecules, their location is
determined by placing sections of the brain against unexposed
film -- grains in the film will be exposed only over the molecule
binocular vision
binocular vision (a.k.a. stereopsis) is the ability to see in depth
it arises form the overlap of the visual fields of the two eyes and is
most well developed in frontally eyed animals
stereopsis is the result of visual processing in which the images from
the two eyes are compared -- differences between the eyes in
the locations of features in the image are used by the brain to
compute the distances of the features from the eyes
ocular dominance
in order to produce stereopsis the brain must have images that are
derived primarily or exclusively from one eye or the other
the visual cortex in mammals is organized into ocular dominance
columns, which are columns of neurons that respond primarily
to input from one eye or the other
ocular dominance columns were demonstrated autoradiographically
radioactive proline injected into one eye is taken up by retinal
ganglion cells and transported back to the LGN and from there
to the visual cortex -- only cells receiving input from that eye
are labeled with the radioactive proline
autoradiographic processing then shows ocular dominance columns
in normal mammals, the ocular dominance columns for one eye are
about the same as for the other eye in terms of width, length,
and the total area they take up in the cortex
most neurons in the ocular dominance columns response primarily to
input from the corresponding eye, but they can also respond to
a lesser degree to input form the other eye, and some cells even
receive input that is about equal from both eyes
thus, cortical cells in normal mammals vary in the degree to which
their responses are dominated by input from one eye
ocular dominance also depends upon the species of mammal studied:
cats -- visual cortical cells tend to be binocular
monkeys -- visual cortical cells tend to be monocular
development of ocular dominance columns is activity dependent
this was shown in experiments in which one eye in a kitten is sutured
shut at birth -- the procedure is called "monocular depravation"
when the eyelid is opened after 3 months, the kitten is blind in that
eye even though the activity of retinal ganglion cells from that
eye, and the LGN cells to which they project, are normal
the monocularly deprived kittens are blind in the previously shut eye
because their visual cortex is not normal
after monocular deprivation, almost all of the cells in the visual cortex
of the kitten respond to input only from the intact eye -- this is
especially striking because visual cortex cells tend to be
binocular in cats
after monocular deprivation, the ocular dominance columns become
asymmetrical -- nearly all of the space in cortex is devoted to
input from the intact eye
thus, even though the retina and LGN can respond to input from the
deprived eye, the cortex receives almost zero input from it
since the visual cortex is needed for most forms of visual processing
in the brain, the monocularly deprived kittens are blind in the
previously shut eye
very similar results are obtained in monkeys
development of ocular dominance columns occurs during a critical period
in kittens, the critical period for development of ocular dominance
columns is from 23 to 40 days (about week 3 to week 6)
suturing shut the eye before or after the period has no effect on the
development of ocular dominance columns in cats
monkeys also have a critical period, but its duration is a few months
rather than a few weeks
ocular dominance columns are formed through competitive interactions
among neurons in visual cortex
a projection from LGN to cortex must be active to compete
successfully for synaptic sites in cortex
since the spontaneous activity waves in retina no longer occur after
birth, this activity must come from the external environment
suturing both eyes shut during the critical period blinds the animal in
both eyes, even though the retinas and LGNs are both normal
now the neurons projecting from the LGN to the cortex are both
equally inactive, so neither has a competitive advantage
ocular dominance columns appear normal, indicating that the
competitive advantage of the open eye caused the asymmetry
in ocular dominance following monocular deprivation
however the neurons have abnormal responses -- they respond almost
exclusively to input from one eye or the other -- there are
almost zero binocular neurons
the abnormality in cortical neurons responses accounts for the
blindness of binocularly deprived animals
activity dependent mechanisms of visual system development in cortex
require formed features as input, not just diffuse light
placing a frosted contact lens over the eye, which transmits only
diffuse light and removes all visual features, has the same
adverse effects on ocular dominance development as eyelid
suturing during the critical period
thus, the final development of the visual system requires the same sort
of visual input that it will need to process in its mature state
nerve growth factor (NGF) affects ocular dominance column formation
depriving one eye of normal visual input causes growth and branching
of the axons that project to cortex from LGN cells serving the
intact eye -- axonal projections from the deprived eye wither
the withering of the projections from the deprived eye can be reversed
by infusion of nerve growth factor (NGF)
infusion of antibodies to NGF prevent its normal trophic effects -- this
causes a prolongation of the critical period
it appears that NGF works against the competitive effects by
stabilizing all connections regardless of activity, thereby
shortening the period during which competitive interactions can
take place
other neurotrophic factors can also influence ocular dominance column
formation
these other neurotrophic factors include:
brain derived neurotrophic factor (BDNF)
neurotrophin-3 (NT-3)
neurotrophin-4/5 (NT-4/5)
these and other neurotrophic factors interact together and with NGF
during the development of ocular dominance columns
their complex interactions are incompletely understood
activity dependent mechanisms of ocular dominance column formation
involves activation of NMDA receptors
NMDA (N-methyl-D-aspartate) receptors are ionotropic glutamate
receptors that are well known for their involvement in long-
term potentiation (LTP) in the hippocampus
glutamate is the transmitter at many synapses in the LGN and in the
visual cortex
NMDA receptors are blocked by amino phosphonovaleric acid (APV)
infusion of APV prevents the competitive shift toward monocularity
in monocularly deprived animals
interestingly, the occurrence of functional NMDA receptors falls off
sharply in visual cortex after the critical period
it appears that activity-dependent synapse formation, and activity-
dependent competition, may require active NMDA receptors
however, these results must be interpreted with a grain of salt, because
NMDA is found widely throughout the brain and so may not be
involved in development specifically
development is also plastic in the somatosensory systems
in the somatosensory system of the rat, the representation for the
forelimb is adjacent to that for the hindlimb
if the forelimb is amputated during development, the region of
somatosensory cortex that would have represented it is instead
innervated by input from the hindlimb
again it appears that inputs to cortex compete for synaptic sites
in the developing rat somatosensory cortex, when a region became
vacant it was taken over by projections form an adjoining
region with still active inputs
even in invertebrates, where development is deterministic, it can still be
affected by experience
for example, if the circus is amputated from an immature cricket, then
the sensory neurons that would have conveyed information
from it to their giant neuron fail to develop
if the circus is removed early enough in the period between molts it
will regenerate, but the corresponding giant neuron never
develops a normal, fully-sized dendritic tree
external input from the environment can also affect connections in the brains
of adult animals
experiments have shown that the somatosensory cortex is capable of
substantial reorganization even in full adulthood
for example, in monkeys, the median nerve innervates the palm side
of the hand, and carries input from there into the CNS and to
its representation in the somatosensory cortex
several months after the median nerve is cut, the part of the
somatosensory cortex that had received input from the palm
side of the hand now receives input from the back of the hand
however, this sort of reorganization in somatosensory cortex is limited
for example, if an entire limb is amputated in an adult, those areas of
somatosensory cortex that represented it will not receive new
inputs but will remain largely silent
somatosensory cortex can also become reorganized by changing the
correlation of its inputs
for example, the representations for two fingers can merge if the
fingers are tied together so that they are always stimulated
together in a correlated way
conversely, the representation for one finger can split if adjacent
regions of that finger are stimulated separately in an
uncorrelated way
adult animals exposed to complex environments have more synaptic
connections than those kept in deprived environments
William Greenough has shown that this increased synaptology
improves learning and memory ability in rats kept in "complex"
as opposed to "simple" environments
effects are observed even in adults after exposure to different
environments for periods as short as a week, showing that there
is no critical period for this sort of new synapse growth