 |
 Previous Article | Next Article 
The Journal of Neuroscience, February 15, 1998, 18(4):1491-1504
Synchronizing Retinal Activity in Both Eyes Disrupts Binocular
Map Development in the Optic Tectum
Stephen G.
Brickley1,
Elizabeth A.
Dawes1,
Michael J.
Keating1, and
Simon
Grant1, 2
1 Division of Neurophysiology, National Institute for
Medical Research, London NW7 1AA, United Kingdom, and
2 Department of Sensorimotor Systems, Division of
Neuroscience, Imperial College School of Medicine, London W6 8RF,
United Kingdom
 |
ABSTRACT |
Spatiotemporal correlations in the pattern of spontaneous and
evoked retinal ganglion cell (RGC) activity are believed to influence
the topographic organization of connections throughout the developing
visual system. We have tested this hypothesis by examining the effects
of interfering with these potential activity cues during development on
the functional organization of binocular maps in the
Xenopus frog optic tectum. Paired recordings combined with cross-correlation analyses demonstrated that exposing normal frogs
to a continuous 1 Hz of stroboscopic illumination synchronized the
firing of all three classes of RGC projecting to the tectum and induced
similar patterns of temporally correlated activity across both lobes of
the nucleus. Embryonic and eye-rotated larval animals were reared until
early adulthood under equivalent stroboscopic conditions. The maps
formed by each RGC class in the contralateral tectum showed normal
topography and stratification after strobe rearing, but with
consistently enlarged multiunit receptive fields. Maps of the
ipsilateral eye, formed by crossed isthmotectal axons, showed
significant disorder and misalignment with direct visual input from the
retina, and in the eye-rotated animals complete compensatory
reorientation of these maps usually induced by this procedure failed to
occur. These findings suggest that refinement of retinal arbors in the
tectum and the ability of crossed isthmotectal arbors to establish
binocular convergence with these retinal afferents are disrupted when
they all fire together. Our data thus provide direct experimental
evidence that spatiotemporal activity patterns within and between the
two eyes regulate the precision of their developing connections.
Key words:
retinal ganglion cell; nucleus isthmi; visual topography; stroboscopic illumination; correlated activity; synaptic plasticity; Xenopus laevis
| |
INTRODUCTION |
Activity contributes importantly to
the organization of synaptic connectivity at many levels of the
developing visual system. For example, topographic projections from the
retina and between different visual centers undergo activity-dependent
refinement during development, in which connections most appropriate
for visual information processing become stabilized and misplaced inputs are withdrawn (for review, see Udin and Fawcett, 1988 ; Constantine-Paton et al., 1990; Rauschecker, 1991; Goodman and Shatz,
1993). It is generally proposed that synchronous activity between
convergent inputs reinforces developing connections, whereas asynchrony
leads to their removal. Two distinct sources of afferent activity, with
topographical origins needed to mediate these processes, now have been
identified; correlated spontaneous firing of neighboring retinal
ganglion cells (RGCs) may drive structural refinements of primary
retinal maps, and correlated visually evoked activity in the two eyes
appears to be necessary for organizing binocularly convergent
projections.
Although these proposals enjoy considerable theoretical support, the
experimental evidence remains indirect. Previous work on the formation
of topographic projections from the contralateral and ipsilateral eyes
onto the frog midbrain tectum typifies this problem. Activity clearly
influences the organization of retinal and crossed isthmotectal
connections underlying these projections (see Fig.
1), because the refinement and proper
alignment of the two maps are disrupted when afferent activity or
synaptic activation of tectal NMDA receptors is reduced (Cline and
Constantine-Paton, 1989, 1990; Grant and Keating, 1989b). These
procedures also block more extreme forms of connectional plasticity in
the two pathways under conditions in which between-eye activity
correlations are assumed to be altered significantly, such as when two
retinal projections form eye-specific stripes in one tectum (Reh and
Constantine-Paton, 1985; Cline et al., 1987), and the ipsilateral map
changes orientation after early eye rotation in Xenopus
(Keating and Feldman, 1975; Scherer and Udin, 1989; Grant and Keating,
1992). However, the main treatment effects in these studies were on the
overall level or balance of neural activity, rather than on its
spatiotemporal pattern or degree of synchrony, so the connectional
disruptions cannot be attributed specifically to interference with
correlated activity-based mechanisms.
View larger version (21K):
[in this window]
[in a new window]
|
Figure 1.
Schematic illustration of the topographic
projections from the two eyes onto the frog tectum and predicted
effects of stroboscopic stimulation on spatiotemporal firing patterns.
A, In normal frogs the retina forms a point-to-point map
across the opposite tectum, with the axons of neighboring RGCs
terminating on adjoining tectal cells. Corresponding tectal positions
are linked by an "intertectal" relay, involving uncrossed
tecto-isthmic and crossed isthmotectal axons, which supply each lobe
with a convergent point-to-point map of the ipsilateral eye.
Consequently, a localized stimulus in the binocular field coactivates
convergent retinal and crossed isthmotectal inputs. B,
Stroboscopic illumination should synchronize activity in both eyes,
subverting spontaneous and evoked activity correlations that may be
involved in refining retinal and crossed isthmotectal connections. Note
that, for clarity, the intertectal relay associated with the left
nucleus isthmi has been omitted, as has the postoptic commissure in
which crossed isthmotectal axons actually decussate.
|
|
To test this possibility more directly, we generated abnormal
spatiotemporal patterns of afferent activity in the developing Xenopus tectum by stimulating animals under constant
stroboscopic light. Previous studies have reported connectional
disorders after similar stimulation of developing (Berman and Payne,
1985; Grigonis and Murphy, 1991; Schmidt and Buzzard, 1993) and
regenerating (Schmidt and Eisele, 1985; Cook, 1987) visual pathways,
but the studies provided no evidence to support the assumption that
activity was correlated in the species used. We demonstrate that during stroboscopic illumination RGC axons fire in synchronized bursts, which
correlate spontaneous and evoked activity right across the tectum,
rather than just locally as under more natural stimulus conditions.
Electrophysiological mapping assays demonstrated that the precision and
plasticity of retinotectal and crossed isthmotectal connections were
disrupted significantly in normal and early eye-rotated frogs reared
under our experimental conditions. The disruptions, although generally
less severe, are similar in nature to those observed when activity in
these developing pathways is reduced.
| |
MATERIALS AND METHODS |
Rearing and recording procedures. Xenopus
embryos were obtained after injection of human chorionic gonadotrophin
into the dorsal lymph sac of adult breeding pairs. Embryos derived from each mating were divided at hatching (stage 35/36; Nieuwkoop and Faber,
1967) into two groups: one group was reared normally under natural
diurnal lighting conditions, and the other was raised in an environment
consisting of continuous stroboscopic illumination (strobe-reared).
Postembryonic animals of each group were maintained initially in
opaque, open-top tanks containing oxygenated Stearns solution and fed
with strained baby food. On reaching larval stages 56-58, individual
animals were transferred into separate black perspex containers fitted
with transparent perforated lids so that incident light could enter the
box only from above. They were fed tubifex worms twice weekly, when
their water was replaced, and remained on this regime until late
juvenile-young adult life, at 1-11/2 years postmetamorphosis
(PM).
Strobe-reared animals were placed inside a light-tight cabinet with
featureless walls housing the xenon strobe unit suspended from its
roof. Strobe flashes were maintained at a constant frequency of 1 Hz
and 10 µsec duration (at
1/2Vmax) throughout the rearing period. The intensity of illumination varied from ~0.5
kW/m2 at the edge of the cabinet to ~10
kW/m2 directly below the unit ("average"
daylight ranges from ~1-3 kW/m2). For this
reason, animals were switched at feeding times to different positions
in the cabinet. To minimize further the availability of visual contrast
information during this regime, we conducted animal maintenance away
from the cabinet under conditions of dim red illumination (compare
Keating et al., 1986).
Uniocular eye rotations of 90-180° were performed on normally reared
stages 56-58 larval animals, anesthetized by immersion in MS222
(tricaine-methanesulphate; 1:1500 in aqueous solution), as previously
described (Grant and Keating, 1992). Once the operated eye was
repositioned successfully, the animals were allowed to recover fully
and then either were returned to their original environment or were
strobe-reared, as above, until the recording experiment.
Electrophysiological recordings, undertaken to assess both the acute
and long-term effects of exposure to stroboscopic illumination, followed several standardized procedures. Young adult frogs were anesthetized by injection of MS222 (2-5 mg, i.p). The optic tectum was
exposed by a partial craniotomy and photographed at 50× magnification onto Polaroid film. Animals were placed at the center of an Aimark projection perimeter (Bausch and Lomb, London, UK), with the optic axis
of one eye centered on the origin of the perimeter arc. Their bodies
were covered with tissue paper that was saturated with oxygenated MS222
solution to ensure that adequate anesthesia was maintained throughout
the recording session. Responses to various visual stimuli were
recorded with tungsten-in-glass microelectrodes of tip diameter 2-5
µm and impedance 0.5-2 M inserted into the superficial tectal
neuropil, guided by reference to a grid drawn on the photographic
enlargement of its surface. The responses were amplified conventionally
and monitored on an oscilloscope and through an audio amplifier. Window
discriminators were used to time the arrival of action potentials at
recording sites and to eliminate background noise levels, as
required.
Effects of stroboscopic illumination on temporal firing patterns
in the tectum. These recording techniques selectively sample from
the active zones of axonal arbors: that is, RGC arbors from the
contralateral eye (George and Marks, 1974) and crossed isthmotectal arbors via the ipsilateral eye (Udin and Keating, 1981). Consequently, we were able to evaluate the responses of normal RGCs to stroboscopic illumination by recording from their arbors in the optic tectum. The
Xenopus tectum receives projections from three classes of RGC, from which the axons terminate in separate strata of the superficial layers (9-8) and from which the responses to conventional visual stimuli differ markedly (Maturana et al., 1960; Chung et al.,
1975; Keating et al., 1986). Class I or "sustained" units are
located just below the pial surface and fire in prolonged bursts when a
small (~5°) visual stimulus is moved into their excitatory
receptive field (RF) and held stationary there, but they are relatively
insensitive to light onset or offset. Class III or "event" units
occupy the next level. They respond optimally to the movement of small
(5-10°) visual stimuli within their RF and also give strong
transient "ON" and "OFF" responses to changes in luminance.
Class IV or "dimmer" units are found deepest down and are
responsive to large (~20°) moving stimuli, but their cardinal response involves sustained bursts of activity to stepwise decrements in RF illumination.
A requirement for initiating the strobe-rearing experiments was that
each of these RGC classes should respond consistently to the chosen
temporal frequency of stroboscopic illumination and that it should
alter the spontaneous discharges of each. In preliminary experiments
(data not shown) we monitored the single-unit responses of each RGC
class to brief (5-10 min) presentations of a range of strobe
frequencies (between 0.1 and 10 Hz). These recordings indicated that
frequencies of  0.5 and of  2.5 Hz were unsuitable for rearing
purposes; all unit classes responded to low-frequency strobe flashes,
but the interstimulus intervals were too long to prevent the return of
spontaneous activity; at the higher frequencies, only event units were
capable of consistent responses. Another requirement was that the
strobe-induced RGC firing should activate postsynaptic tectal cells.
Because we are unable to record from these with our techniques, this
was assessed indirectly by examining the selective outflow of their
activity via the tecto-isthmic and crossed isthmotectal pathways (Fig. 1A). In these experiments frogs were enucleated
monocularly and multiunit recordings made from just below the pial
surface in the tectum ipsilateral to the remaining eye, where the
arbors of fine-caliber crossed isthmotectal axons terminate (Udin,
1989). Ipsilateral units are known to habituate during repetitive
stimulation in anesthetized frogs, but successive responses to
low-frequency (1 Hz) strobe flashes were obtained, indicating that
these arbors were firing because of activation of cells in the opposite
tectum.
Because these observations suggested that a 1 Hz frequency might confer
optimal conditions for strobe rearing, a more detailed series of
control experiments was undertaken to determine whether this strobe
frequency could synchronize and correlate RGC firing during much longer
exposures. In these experiments two microelectrodes were used for
simultaneous paired recordings in the same or opposite tectal lobes,
with their depths and positions carefully adjusted to record multiunit
activity from the same class of RGC having response fields that were
either partially overlapping or completely nonoverlapping in visual
space. Then the strobe unit was moved into position ~33 cm from the
animal, either within the region of RF overlap or at a strategic
distance between the two nonoverlapping fields that optimized their
firing to test flashes. Spontaneous discharges (in the dark) were
monitored at the two recordings sites for 5-15 min before the strobe
was activated, and responses to constant stroboscopic illumination (10 µsec duration at 1 Hz) were recorded for a further period of 2-6 hr.
Toward the end of the session the response or responses of one or both
multiunits to a 5, 10, or 20° black disk introduced into their RF or
RFs also were recorded to evaluate any modulatory effects of visual contrasts on the strobe-induced firing pattern.
Signals from the two recording sites were processed through separate
channels and logged for the entire course of the experiment with a CED
1401 (Cambridge Electronic Design, Cambridge, UK) laboratory interface
on which on-line and subsequent analyses could be performed via an IBM
computer. Strobe-induced changes in the temporal pattern of RGC firing
were examined by depicting the data from each channel as raster plots
showing the time of occurrence of individual spikes in relation to the
strobe flash for every 1 sec epoch of the experiment. Changes in firing
rates were calculated from the number of spikes per second after each
strobe flash. Cross-correlation (CC) analyses were performed on the
simultaneously recorded responses. For this purpose one channel was
treated as a "reference," and for each spike in this channel the
times at which any spikes occurred in the other channel during the
interval 0 ± 500 msec were computed, after dividing this channel
into 2 msec bins. The time-averaged CCs determined in this way from
selected periods in each recording session were plotted as histograms
showing the temporal phase and relative strengths of the correlations
that were present.
To provide a quantitative indication of the degree of coincident firing
occurring under the different conditions of these recordings, which
subsumed differences in firing rates associated with them, we divided
the average number of events occurring in the CC histograms at
time = 0 ± 10 msec by the average number of noncoincident
events (i.e., in the remaining bins at t = ± 10-500
msec) over the same sampling period. This measure was termed the
"coincidence index" (CI10);
increasing values >1.0 on this index indicate increasing degrees of
in-phase, synchronous activity at the two recording sites (while
disregarding any anti-phase or asynchronous relationships). As an
indication of any broader temporal correlations that were present, a
"correlation index" (CI50) also was
calculated from the data by similarly comparing the events occurring at
t = 0 ± 50 msec with those in the remaining bins.
Our reason for selecting these particular intervals is that 10 and 50 msec correspond, respectively, to the average initial delay and
subsequent overlap in the activity evoked in the normal Xenopus tectum by ON/OFF stimuli at corresponding positions
in the contralateral and ipsilateral eyes (Scherer and Udin, 1991). Either or both intervals thus may have physiological relevance for
stabilizing the developing visual connections under investigation, as
well as falling within known integration times for inducing long-term
potentiation of synaptic strength after stimulating paired inputs
(i.e., associative LTP) in the adult hippocampus (Levy and Steward,
1983; Gustafsson and Wigstrom, 1986). In these analyses we made no
attempt to separate direct stimulus-induced correlations in firing from
those arising via "effective connectivity" (Aertson et al., 1989)
between the paired recording sites. The reason for this is that
effective connectivity attributable to common input (e.g., onto
neighboring class-specific RGCs) or to direct (e.g., intertectal)
connections is more likely to exist at corresponding than at
noncorresponding sites, and so the analysis should properly reflect
this difference by including all possible sources of coincident or
correlated activity. In fact, the completed analyses revealed no
meaningful differences between the two indices used, so only the data
relating to CI10 are shown.
Effects of stroboscopic illumination on binocular map
development. Visual inputs to the tectum were mapped at multiple
recording sites in normal and strobe-reared frogs by using conventional visual stimuli presented against the uniform background of the perimeter arc and RFs plotted onto polar coordinate charts for future
analyses. Initial efforts focused on the direct retinotectal projection
from the contralateral eye. To assess its topographic order, we made
stepwise electrode penetrations into the tectum, separated by 100-200
µm across its surface dimensions and by 25-50 µm in depth while
plotting minimum response fields and classifying the sequence of units
that were encountered. In some of these experiments and in later ones
dedicated to the purpose, the RF sizes of single- and multiunits
recorded at different depths were evaluated quantitatively, using
computer-controlled methods of visual stimulation and data capture
(Keating et al., 1986). These measures can provide an index of the
detailed precision of the retinotectal map; enlarged multiunit RFs
(MURFs), in the absence of changes in single-unit RF size, indicate
that the RGCs projecting to a given tectal site arise from a wider
territory than normal and/or that their arbors are more diffuse. To
minimize other variables that can affect these measures, we used the
same set of low-impedance electrodes for all of the recordings in both
groups of frog, and window discrimination of the amplified signals was
always set at 100% above the noise level. RFs were plotted on a large
television screen, positioned 38 cm from the stimulated eye, and
covered an effective stimulus area of 64.5° in the nasotemporal axis
and 34.5° in the superoinferior axis of its visual field. The visual stimulus (a 6° black square) was moved either horizontally or vertically in 1.5° steps across the screen at a constant speed of
35°/sec. Spikes occurring during each traverse were timed and stored
on computer. The direction of stimulus traverse was interleaved randomly between runs, with a least three sets of traverses completed for each possible direction. For each MURF plotted, electrode depth was
set to ensure that only one RGC class was being sampled and that the
response contained at least three separate units, as revealed by
differences in their spike height and width. To analyze the data, we
produced two-dimensional matrix representations of the screen, in which
the effective stimulus area was divided into 1.5° × 1.5° pixels
and the number of spikes per pixel was presented as an element of the
matrix. The RF size was measured by pooling matrices: the horizontal
diameter by combining matrices from the two vertical directions of
stimulus movement, the vertical diameter from its two horizontal
directions of movement, and its area from the outline of the combined
matrices. To avoid inclusion of elements arising from occasional
spontaneous or artifactual bursts of spikes, we considered only pixels
containing 10% of the maximum number of spikes per pixel to be
genuine components of the RF.
In other animals of both groups, including those with early eye
rotations, the visual projections from both eyes to one tectum were
mapped to examine the topographic order of the crossed isthmotectal map
from the ipsilateral eye and to assess its spatial alignment with the
direct retinal map. In these experiments we plotted minimum response
fields in the contralateral and ipsilateral eyes independently on the
polar coordinate system while alternately covering the nontested eye
with an opaque shield. The center of each RF was defined as the
geometric center of the response field, a position that typically
approximated the point of maximal activation. Spatial alignment between
the two maps was quantified by calculating the disparity (in degrees)
between the binocular RF centers at each tectal site receiving input
from the two eyes. Disparity was expressed both by its absolute value
and by its horizontal (nasotemporal) component. Data derived from the
strobe-reared frogs were compared with the controls from the present
investigation and also with those previously obtained from age-matched
dark-reared frogs in which these features were found to be disrupted
systematically (Grant and Keating, 1989b). These and other statistical
comparisons were conducted with the software package SIGMA STAT.
| |
RESULTS |
The results are presented in two sections. First, we provide
evidence that exposure to 1 Hz of stroboscopic illumination interferes with the temporal pattern of RGC firing, in support of the rationale for the rearing experiments (Fig. 1B). Second, we
examine the consequences of development under these conditions for the
organization and plasticity of binocular maps in the optic tectum.
Control experiments: correlated firing in the tectum
Simultaneous recordings were undertaken in 12 normal frogs to
monitor the effects of prolonged 1 Hz of stroboscopic illumination on
RGC firing patterns. The recordings were obtained from pairs of
sustained (n = 4), event (n = 6), or
dimmer (n = 2) RGC classes, comprising six pairs from
corresponding sites with spatially overlapping RFs and six from
noncorresponding ones with disparate, completely nonoverlapping RFs.
These and other details are summarized in Table
1.
To establish a baseline against which any strobe-induced effects could
be compared, we conducted CC analyses of visually evoked and
spontaneous activity for each RF pair before activating the strobe.
Figure 2A shows typical
responses obtained from corresponding recording sites when a visual
stimulus was introduced into the region of RF overlap. The CC
histograms derived from such data always revealed a broad (100-200
msec) peak of correlated activity centered at t = 0, with a CI10 between 2.1 and 3.5, indicating synchronous firing at the two tectal sites. This pattern of correlated activity could be replicated at noncorresponding tectal sites by
perfect in-phase stimulation of the two disparate RFs, but with more
natural out-of-phase stimulation their activity was anticorrelated with
a similarly broad, but asymmetric, peak in the CC histogram and
sometimes with a depression at t = 0, indicative of
noncoincident firing (CI10 0.8). By contrast,
CCs derived from visual stimulation of only one RF of the
nonoverlapping pair were always flat (Fig. 2B),
indicating that correlations between evoked activity at stimulated
tectal sites and spontaneous firing at other sites were random
(CI10 = 1 ± 0.1).
View larger version (23K):
[in this window]
[in a new window]
|
Figure 2.
Correlations in firing at corresponding and
noncorresponding tectal sites evoked by localized binocular visual
stimulation. A, The top and bottom
traces show responses recorded simultaneously from paired sites
in opposite tectal lobes receiving corresponding retinal input during
stimulation of their overlapping RFs (with a 10° black disk moved
against the background of the illuminated perimeter arc). The
top bars indicate stimulation periods and refer to both
traces. The middle panel inset shows the time-averaged CC histogram computed from these recording data: the
abscissa (in seconds) is divided into 2 msec bins, and
the ordinate indicates the number of events per bin. A
broad peak of correlated activity is centered at t = 0 (±100 msec): CI10 = 2.2. B, Simultaneous recording from sites in opposite tectal
lobes receiving noncorresponding retinal input during visual
stimulation of one RF (top trace) of the nonoverlapping
pair. The CC histogram (middle panel inset) derived from
this recording indicates that the correlation between the evoked and
spontaneous activity is random: CI10 = 1.0. Other conventions are as in A. Calibration refers to all
traces.
|
|
Spontaneous activity generally was maintained at low levels regardless
of RGC class, both in the light (Fig. 2) and in the dark (Fig.
3), with just occasional bursts of spikes
detected. Nonetheless, the spontaneous discharges occurring at
neighboring locations in the same tectum consistently exhibited a
degree of synchrony (CI10 > 1.0; Table 1)
presumably arising via effective connectivity, whereas random activity
correlations were present among most of the other paired sites (compare
Fig. 2B).
View larger version (67K):
[in this window]
[in a new window]
|
Figure 3.
Entrained patterns of RGC firing induced by
exposure to a continuous 1 Hz of stroboscopic illumination in
(A) sustained, (B) event,
and (C) dimmer unit classes. In these raster
plots each dot represents an action potential recorded
during the first 3 hr of an uninterrupted recording session in three
different experiments (A, B, and C; one of the unit pairs str-03,
str-02, and str-07). The x-axes represent the time (in
seconds) for this part of each recording session, and the
y-axes correspond to the 1 sec interval between strobe
flashes (occurring at the abscissa in each plot). For
the period indicated by the short open bar at the
top of each panel, the frog was maintained in darkness
while spontaneous discharges were monitored; the long black
bar indicates the period of exposure to stroboscopic
illumination. Within a few minutes of strobe onset, an entrained
triphasic response pattern emerges and is maintained throughout the
recording period. Calibration (bottom left) refers to
all panels.
|
|
Stroboscopic illumination entrains the firing of all three
RGC classes
All 24 sustained, event, or dimmer unit types that were examined
acquired and maintained similar patterns of entrained firing during
prolonged exposure to 1 Hz of stroboscopic illumination. The raster
plots in Figure 3 illustrate the time course of this entrainment effect
for representatives of each RGC class. Initial responses to the strobe
usually were vigorous but variable in rate (Table 1) and temporal
pattern. With continued exposure the firing rate became more stable,
and a distinctly triphasic response pattern emerged, comprising short,
middle, and long latency bursts of periodic activity with silent or
near-silent periods in between, each element of which was reproduced
after successive strobe flashes. The triphasic latency profile was
remarkably similar across RGC classes (see also Fig. 5). The early and
middle components of the response were most prominent at 100-150 msec
and at 300-500 msec poststimulus, respectively, and the late component
began at ~700 msec, usually to be curtailed only by the next strobe flash in the sequence. Moreover, the entrained firing patterns remained
essentially invariant and without obvious contamination by spontaneous
discharges until the recording was interrupted. These results thus
imply that stroboscopic illumination causes RGC axons in all regions
and depths of the Xenopus tectum to fire in synchrony.
The response of normal RGCs to 1 Hz of stroboscopic illumination has
been examined previously during several minutes of continuous recording
from the goldfish tectum and consisted of just a single short latency
burst, ~50-200 msec in duration, followed by quiescence (Schmidt and
Eisele, 1985). The more intermittent bursting activity in
Xenopus almost certainly results from differences in
intraretinal processing between frogs and fish, a difference that
begins in the photoreceptor layer, which is known to be rod- and
cone-dominated, respectively, in these species. Indeed, intermittent
periodic bursts of activity, lasting for as long as 1 sec poststimulus, also appear to characterize the firing patterns of frog RGCs responding to the onset and offset of bright light flashes (25-1000 msec duration), which are confined to their excitatory RF [Rana,
Stiles et al. (1985); Xenopus, S. Brickley and S. Grant,
unpublished data]. The main difference between these responses and
those elicited by (more transient and diffuse) strobe flashes is that
the earliest component (up to 250 msec poststimulus) is more periodic
and begins at a shorter (50-75 msec) latency. The strobe-entrained
firing pattern of Xenopus RGCs thus appears to be a variant
on their normal temporal response to abrupt changes in
illumination.
Stroboscopic illumination induces temporal correlations in firing
at all tectal locations
Representative data showing the effects of stroboscopic
entrainment on correlations in firing at corresponding and at
noncorresponding tectal locations are presented in Figure
4. CC histograms computed from all of the
paired recording sites in positions of tectal correspondence (Fig.
4A), whether in the same or opposite sides of the
brain, exhibited a sharp (20-50 msec) peak of coincident activity
centered at t = 0 as soon as the strobe flashes began within their region of RF overlap (Table 1). With more prolonged exposure and consequent entrainment of firing at the two recording sites, this correlated firing usually improved but was accompanied by
the emergence of secondary peaks of anticorrelated activity (at > ± 50 msec), the number and distribution of which depended on subtle
phase differences in the periodic activity of the two strobe-entrained
responses. At paired noncorresponding tectal locations, correlations in
firing were closer to random at strobe onset, but a similar pattern of
correlated and anticorrelated firing emerged during strobe entrainment
(Fig. 4B). Most strikingly, a sharp peak of
coincident activity, at t = 0 ± 10 msec, appeared in all of the CC histograms computed from these paired recordings once
entrainment was well established (Table 1). These results indicate,
therefore, that stroboscopic synchronization of RGC firing induces
similar temporal patterns of correlated and anticorrelated activity at
all tectal locations.
View larger version (22K):
[in this window]
[in a new window]
|
Figure 4.
Stroboscopic entrainment leads to similar temporal
correlations in firing at (A) corresponding and
(B) noncorresponding tectal sites. Each histogram
shows time-averaged CCs computed from 60 sec episodes of simultaneous
recording from the onset of stroboscopic illumination (at time 0, as
indicated, right) and at subsequent 15 min intervals
during the entrainment of the responses. In A, a sharp
peak of coincident activity was present throughout the recording, which
increased slightly during strobe entrainment (unit pair, str-05); in
B, activity was correlated randomly at strobe onset, but
a peak of coincident activity emerged as a result of strobe entrainment
(unit pair, str-11). Calibration conventions, as in Figure 2
insets, refer to each histogram in the landscaped presentations.
|
|
Effects of visual contrasts on the rate and temporal pattern of
stroboscopically entrained firing
Although we took steps to limit visual contrast information in the
strobe-rearing environment, some potential sources of this (e.g., the
air holes in the animal container lids and food supplied) could not be
excluded. To evaluate their possible effects on strobe-induced firing
patterns, we introduced visual stimuli (black disks) into the RF or RFs
of each unit pair once stable strobe entrainment and activity
correlations had become established. Figure
5, A and C, shows
segments of the raster plots for unit pairs of each RF condition for
the period immediately before, during, and after visual stimulation of
both (overlapping) or one of their (nonoverlapping) RFs. The visual
contrast significantly increased the rate of firing of the stimulated
units (by 50-100%; Table 1), most prominently in the interval between
200 and 700 msec after each strobe flash. As indicated in the raster
plots of Figure 5 and confirmed by analyses of cumulative poststimulus
time histograms compiled from these recording periods (data not shown),
this contrast-evoked activity was superimposed on the middle component
of the strobe-entrained firing and on the silent periods on either side
of it, with little effect on the early or late components of the
entrained response.
View larger version (45K):
[in this window]
[in a new window]
|
Figure 5.
Effects of localized binocular visual stimulation
on stroboscopic entrainment and correlated firing at (A,
B) corresponding and (C, D) noncorresponding
tectal sites. The raster plots in each top panel
(A, C) show 10 min segments (open bar
above) of strobe-entrained responses recorded toward the end of
a 4 hr exposure to a continuous 1 Hz of stroboscopic illumination.
During the middle of the period shown (solid black bar
above), a visual stimulus (10° diameter black disk) was moved
against the background of the perimeter arc through the region of RF
overlap (in A; str-10) or through one RF of the
nonoverlapping pair (left, in C; str-11) while the strobe continued to flash. Additional activity evoked by this
contrast stimulation is evident in the raster plots and is associated
with an increase in the firing rate of the stimulated units (see Table
1). B and D show residual time-averaged
CC histograms, produced by subtracting the cross-correlations in firing
for the 60 sec periods immediately before and during contrast
stimulation in each unit pair. These indicate that correlations in the
contrast-evoked activity are random; however, note the central
depression at t = 0 ± 10 msec in
B, reflecting the general reduction in levels of
coincident activity resulting from visual stimulation in the strobe
(see Table 1). Other conventions are as in Figure 2
insets.
|
|
This increased firing would be expected to impact on the pattern of
activity correlations, because the added activity presents more
opportunities for temporal coincidences and noncoincidences to be
established. However, CC histograms computed for the period of visual
contrast stimulation, although differing markedly from those obtained
for comparable stimulation under normal viewing conditions (see Fig.
2A), showed remarkably similar patterns of correlated
and anti-correlated activity to those derived from the strobe-entrained
periods immediately before and afterward. Indeed, the main difference
appeared to be in the height of the peaks, all of which were elevated
in the CCs computed from the responses during contrast stimulation,
suggesting that absolute levels of both coincident and
noncoincident activity were increased. This was confirmed by
subtracting one CC from another (e.g., during contrast stimulation
minus before). As typified in Figure 5, B and D,
the residual histograms were flat for both overlapping and
nonoverlapping RF conditions, indicating that contrast-evoked correlations in firing were essentially random (compare Fig.
2B) during stroboscopic entrainment.
To determine the effects of contrast-evoked firing on the
relative levels of correlated and anti-correlated activity,
which might provide a basis for distinguishing corresponding from
noncorresponding tectal locations, we calculated coincidence indices
for the period of stimulation and compared them with those for
preceding periods of stable strobe entrainment. As shown in Table 1,
these comparisons revealed that during visual stimulation the
CI10 was generally reduced, and by a
similar degree, at both corresponding and noncorresponding sites. These
findings strongly suggest that the additional firing induced by visual
contrast produces an equivalent net loss in the probability of
coincident activity at all tectal locations. The explanation for this
rests with the temporal profile of the contrast-evoked RGC firing,
which occurred sporadically over a long (500 msec) period after each
strobe flash, rather than in synchrony with elements of the entrained
response.
Effects of strobe rearing on binocular map development
and plasticity
Organizational features of the maps formed by the retinotectal and
crossed isthmotectal (intertectal) projections were examined in 41 normal and 48 strobe-reared frogs.
The topography of the retinotectal map is normal in
strobe-reared frogs but may lack detailed precision
The topographic order of the retinotectal projection in
animals strobe-reared from embryonic stage 35/36 (see Figs.
6, 8) was indistinguishable from normal;
stepwise microelectrode penetrations along the rostrocaudal or
mediolateral tectal axes yielded similar nasotemporal or superoinferior
progressions in RF position through the contralateral eye in both
groups of frog. These features of normality after strobe rearing were
shared by all three classes of RGC projecting to the tectum, as was
their relative depth of termination within the superficial neuropil. In
89 penetrations made into the tectum of these animals
(n = 14) in which two or three classifiable unit types
were encountered, a complete sustained-event-dimmer sequence was
obtained from superficial to deep in 31, and combinations of two unit
classes in the same appropriate sequence were obtained in the
remainder: that is, a sustained unit first and an event (n = 38) or a dimmer unit deeper down
(n = 5) or just an event-dimmer (n = 15) combination. Similar proportions of classifiable unit types and
sequences were recorded in 65 penetrations in control (n = 10) animals ( 2 = 8;
df = 7; p > 0.3). In addition, as indicated in
Figure 6, RFs of the different units encountered in any given
penetration were, in large part, overlapping in space, especially in
recordings from central tectum in which such penetrations should be
near perpendicular to its surface. The scatter in the RF centers of different unit classes mapped sequentially at such central tectal sites
was also comparable in strobe-reared (mean = 6.0 ± 3.0°, SD; n = 66) and normally reared frogs (mean = 5.5 ± 2.7°, SD; n = 57; Student's t
test, p > 0.4). These results indicate that all three
RGC classes in strobe-reared Xenopus are able to deploy their arbors with appropriate topographic order in the opposite tectum,
both across its surface and in depth.
View larger version (23K):
[in this window]
[in a new window]
|
Figure 6.
The retinotectal projection in a strobe-reared
frog. The bottom diagram shows the outline of the right
tectal lobe traced from a photograph of its dorsal surface, with the
positions of microelectrode penetrations shown by
numbers in the grid array. R, Rostral
(medial is to the right). The large
circle derives from the polar coordinate chart representing the
hemispheric mapping perimeter, which closely approximates the visual
field of the animal's left eye. The optic axis of this eye was
centered on the origin of the perimeter. N, T, S, I,
Nasal, temporal, superior, and inferior aspects of its visual field. In
the chart representation, minimum response fields of single or
multiunit event-type RGC projections recorded at each numbered tectal
position are shown in outline and are marked by the
corresponding number situated at or close to the RF
center. At each tectal position sustained- and/or dimmer-type RGC
projections with overlapping response fields also were mapped more
superficially or at deeper locations, respectively, in the same
penetration; s and d indicate the RF
centers of these respective unit types. Note that the apparent
enlargement of event-class response fields from center to periphery in
the chart arises from the polar transformation of visual coordinates
that were used and not from a genuine increase in RF size with
eccentricity.
|
|
In these recordings the response fields of the multiunits that were
plotted were typically 20-40° in diameter (Fig. 6), irrespective of
rearing conditions. However, to obtain a more accurate index of this
aspect of the mapping fidelity, RF sizes in the retinotectal projection
were measured by quantitative methods. Single-unit (class III) RF
diameters were identical in the two groups of frog, averaging 19° in
both nasotemporal and superoinferior dimensions (n = 10/group). MURF sizes, measured for each of the three RGC classes, are
presented in Table 2. Illustrative
three-dimensional matrix plots of the data that were obtained are shown
in Figure 7. The average MURF diameters
and their areas increased significantly with depth in both groups of
frog (i.e., sustained < event < dimmer; Student's
t tests, p < 0.05), but all equivalent
dimensions were larger in the strobe-reared animals (n = 10) than in the controls (n = 12). The increases were
statistically significant for sustained multiunits and for some of the
event MURF dimensions.
View larger version (51K):
[in this window]
[in a new window]
|
Figure 7.
Three-dimensional representations of event-type
MURFs recorded in (A, C) normal and
(B, D) strobe-reared frogs. The
x- and y-axes correspond to 64.5° of
the nasotemporal and 34.5° of the superoinferior visual field,
equivalent to the area covered by the television monitor on which the
MURFs were plotted. The z-axis represents the cumulative
number of spikes recorded in each 1.5 × 1.5° pixel on the screen
from three traverses in each orthogonal direction from which the MURF
areas were measured. Asterisks (C, D) denote double activation peaks.
|
|
Probing these differences further, it was evident from the matrix
plots that the MURF profiles fell into two categories: those possessing
a single, central peak of maximal activation (Fig. 7A,B) and
others possessing two or more regions of peak activity (Fig.
7C,D). This latter category of multiunit response could imply that the RGCs contributing to it had terminal arbors of normal
size but cell bodies that were spatially separated on the retina,
rather than immediate neighbors, and might be an expected trait of
strobe-induced disorder in the retinotectal map (Cook and Rankin, 1986;
Cook, 1987). Such responses were, however, recorded with similar
frequency in the normal (10 of 61; 16%) and strobe-reared animals (14 of 61; 23%). We believe, therefore, that the consistent trend toward
MURF enlargement across RGC classes after strobe rearing more likely
reflects a general increase in retinal arbor size, so that our sampling
was obtained from a more diffuse set of active terminal branches. This
interpretation is consistent with recent findings that retinal MURFs
and arbor sizes are enlarged in the tectum of goldfish reared under a
mixed strobe/dark cycle (Schmidt and Buzzard, 1993).
The topographic order and alignment of the crossed isthmotectal map
is disrupted in strobe-reared frogs
In 18 strobe-reared frogs the binocular inputs to one or both
tectal lobes were mapped in sufficient detail to evaluate the order and
alignment of the crossed isthmotectal projection. Two classes of result
were obtained. In 10 animals the ipsilateral map or maps appeared
completely normal, displaying good topographic order and spatial
alignment with the map from the contralateral eye on the same tectal
lobe. In the other frogs, an example of which is shown in Figure
8, the ipsilateral map or maps displayed abnormal features reminiscent of dark-reared Xenopus
(Keating and Kennard, 1987; Grant and Keating, 1989b). Their overall
topography showed some order, with sequences of electrode penetrations
across the rostrocaudal or mediolateral axes of the tectum usually
revealing appropriate temporonasal or superoinferior shifts in RF
position. However, there were also signs of misalignment with the
contralateral map, which were manifest at some recordings sites (e.g.,
4, 7, and 27 in Fig. 8) by abnormally large spatial disparities between the RFs in the two eyes and manifest at others on the margins of the
map (e.g., 1-3 and 31-33) by systematic inward shifts in ipsilateral
relative to contralateral RF positions, giving the maps a compressed
appearance. At some of these recording sites (33 of 207; 16%),
ipsilateral multiunit response fields were almost double the normal
size, ranging from 60 to 90° in diameter.
View larger version (20K):
[in this window]
[in a new window]
|
Figure 8.
Binocular maps on the tectum of a strobe-reared
frog. The dorsal outline of the optic tectum shows the recording
positions (numbers and intermediate dots)
on the right tectum in this mapping experiment. The corresponding
numbers and intermediate dots in the chart representations indicate the
minimum response field centers mapped through the tested eye.
Filled symbols linked by solid lines
represent RF centers recorded via the direct retinotectal projection
from the contralateral eye (filled arrow);
open symbols linked by discontinuous
lines represent RF centers recorded via the crossed
isthmotectal projection from the ipsilateral eye (open arrow). All other conventions, including left eye centering,
are as in Figure 6. The retinotectal map is organized normally, but the
ipsilateral map shows topographic disorder and distortion and lacks
proper alignment with the contralateral map at most recording
positions.
|
|
Quantitative analyses of the strobe-reared ipsilateral maps were
undertaken to establish more firmly the nature and severity of the
abnormalities present and to better assess their similarity with those
known to exist in Xenopus that were dark-reared to the same
age. For these purposes, data from the strobe-reared frogs with
apparently "unaffected" and "abnormal" maps were treated separately. The spatial alignment between the ipsilateral and contralateral maps, expressed by their mean absolute and nasotemporal binocular RF disparities, is given in Table
3. These latter disparities were grouped
by mediolateral row of tectal recording sites and the mean values
computed for each such row along the rostrocaudal axis. On our
conventions, indications that the ipsilateral map is systematically
compressed are revealed by large positive nasotemporal disparities
rostrally and by large negative disparities caudally. There were no
significant disparity differences between the control and the
unaffected strobe-reared animals: absolute disparities averaged
~10°, and nasotemporal disparities were within ± 5° of zero
for all rows, as previously reported in normal Xenopus at 1 year PM (Grant and Keating, 1989a). In the abnormal strobe-reared frogs, absolute disparities were significantly larger than normal, as
were the positive and negative nasotemporal disparities in the most
rostral and caudal tectal rows (outcomes that also were obtained after
pooling these data with those from the "unaffected" animals).
However, all of these disparity values were significantly smaller than
in dark-reared frogs. Taken together, these results show that strobe
rearing can lead to disorders in the crossed isthmotectal map, which
are qualitatively similar to, but less severe than, those resulting
from total visual deprivation.
Only limited intertectal plasticity occurs after larval eye
rotation in strobe-reared frogs
The much larger scale reorienting of this map that usually follows
early eye rotation also was affected by strobe rearing. In a control
group of eye-rotated frogs reared under natural lighting conditions
(n = 4), the binocular inputs mapped at one or both tectal lobes were found to be aligned spatially, despite the eye rotation present (which ranged from 60-180°). Absolute binocular RF
disparities in these frogs (mean = 8.7 ± 4.8°, SD;
n = 128) were no different from those in normal
age-matched animals (Table 3), indicating that the reoriented map
formed by crossed isthmotectal arbors in eye-rotated Xenopus
allowed postoperative vision can be as accurate as a normally ordered
one.
The results obtained from the strobe-reared frogs (n = 8) differed markedly from those of the controls; the ipsilateral maps showed both disorder and compression and no or very little plasticity. In five frogs with eye rotations of between 120 and 180°, the binocular inputs to one or both tecta were mapped at 89 recording sites
and found to be completely misaligned by an amount equivalent to the
degree and direction of the eye rotation present. An example is shown
in Figure 9. This result implies that the
crossed isthmotectal arbors in these frogs, although disordered, had
maintained their normal orientation and is identical to the effects of
dark rearing (Keating and Feldman, 1975) or of blocking tectal NMDA
receptors (Scherer and Udin, 1989) after larval eye rotation in
Xenopus.
View larger version (22K):
[in this window]
[in a new window]
|
Figure 9.
Binocular maps on the tectum of an eye-rotated,
strobe-reared frog showing no evidence of intertectal plasticity. In
this animal the left eye was rotated counterclockwise by 180° at
larval stage 56. All conventions are as in Figure 8, except that the contralateral and ipsilateral maps on both tectal lobes are shown with
the animal positioned so that its rotated eye is centered on the origin
of the coordinate system. Both visual projections recorded through this
operated eye are rotated 180° counterclockwise, and both visual
projections through the unoperated right eye are in a normal
orientation. As a result, the binocular maps on each tectum are out of
alignment by the degree and direction of the eye rotation present
(e.g., compare RF positions 1-3 in the two eyes). The
ipsilateral maps also show signs of disorder and compression.
|
|
The other three frogs, which had smaller eye rotations ranging
from 60 to 120°, showed some evidence of intertectal plasticity. As
illustrated in Figure 10, the binocular
maps in these animals contained a mixture of predominantly misaligned
with some aligned RFs in the two eyes. In most regions of the tectum
the ipsilateral maps were disordered and out of register with the input
from the contralateral eye by the degree of rotation present (mean
disparity = 89.8 ± 36.8°, SD; n = 83), but
in regions representing the visual field around the axis of the
rotation, binocular RFs showed evidence of spatial correspondence. The
mean disparities of these matching RFs (15.2 ± 7.2°, SD;
n = 25) were significantly larger than in normal
animals and in the control eye-rotated frogs (Student's t
tests; p < 0.0001) and no different from those in the
"affected" (unrotated) strobe-reared group (Table 3). Nonetheless,
such partial map alignment indicates that a cohort of crossed
isthmotectal arbors had reorganized in these animals, almost
compensating for the rotated visual input. A similar partial
intertectal plasticity is known to occur during the initial process of
map reorientation in Xenopus, but it has never been seen in
eye-rotated animals reared in total darkness (Keating and Feldman,
1975; Grant and Keating, 1992; Keating and Grant, 1992).
View larger version (28K):
[in this window]
[in a new window]
|
Figure 10.
Binocular maps on the tectum of an
eye-rotated, strobe-reared frog showing partial intertectal plasticity.
All conventions are as in Figures 8 and 9. In this animal the left eye
was rotated at larval stage 56. For the most part, the binocular maps
on each tectum are out of alignment by the degree and direction of the eye rotation present (equaling 120° counterclockwise) at the time of
recording. However, a component of each of the two ipsilateral maps
arising from the central portion of the chart representation of the
visual field (outlined) and recorded from a restricted portion of each tectal lobe (outlined) is rotated by
120° counterclockwise in approximate spatial alignment with the
corresponding RFs recorded through the contralateral eye (also
outlined) at the same tectal positions. This implies
that some arbors in the crossed isthmotectal projections of this animal
had reorganized to compensate for the eye rotation. Note that at all
five recording sites in the left tectum at which rotated RF positions
were plotted through the ipsilateral eye (e.g., 57"), a second RF was
also present in a normally oriented position (e.g., 57') within the
more disordered and compressed part of its map. Such dual fields are
common during the early postmetamorphic phase of eye rotation-induced
map plasticity in Xenopus and indicate the coexistence
of altered and normal arbors at the same tectal site (Grant and
Keating, 1992).
|
|
| |
DISCUSSION |
The results show that (1) stroboscopic illumination induces high
rates of entrained firing in Xenopus RGCs, masking any local correlations in their spontaneous discharges and synchronizing activity
in the two eyes, and (2) in frogs reared under these conditions the
precision, alignment, and plasticity of visual maps in the tectum are
disrupted. We attribute these disruptions to strobe-induced levels of
correlated afferent activation sufficient to obscure temporal
differences between the pattern of activity at topographically
appropriate and inappropriate synapses.
Strobe rearing was initiated mainly at embryonic hatching, when
all retinal neuron classes are present in Xenopus, and just before the normal onset of photoreceptor function at stage 39/40 (Witkovsky et al., 1976). During the next 10-12 d (up to midlarval stage 49/50), RGC responses to visual stimulation, including transient light flashes, usually become increasingly reliable, coinciding with
the emergence of distinct plexiform layers (Chung et al., 1975) and
synaptic inputs from bipolar and amacrine cells onto discrete
morphological classes of RGC resembling those of the adult retina
(Fisher, 1976; Sakaguchi et al., 1984). This early period is also a key
stage in the formation of the Xenopus retinotectal map.
Previous studies have shown that retinal axons first form terminal
arborizations in the tectum at stage 39 and that they undergo
substantial refinements, especially along the rostrocaudal axis, until
the appearance of well ordered topography at stage 49/50 (Gaze et al.,
1974; Holt and Harris, 1983; O'Rourke and Fraser, 1986; Fujisawa,
1987). The arbor refinements have been visualized in situ
and shown to be influenced by synaptic activity at NMDA receptors
(O'Rourke et al., 1994). Whether the retinal origin of this activity
is generated spontaneously, as occurs in other vertebrate classes while
translaminar retinal inputs are being assembled (Maffei and
Galli-Resta, 1990; Meister et al., 1991; Wong et al., 1993; Sernagor
and Grzywacz, 1996), or is light-evoked remains unknown. Regardless, we
doubt that strobe entrainment of RGC activity would have begun until
these midlarval stages, when the retinal circuitry needed to drive
their responses to light becomes fully established and the early map
refinements are completed.
Major adjustments in retinotectal connectivity would be required during
subsequent development, however, when activity is synchronous under
strobe conditions. New RGCs are added constantly to the retinal
periphery throughout later larval and juvenile postmetamorphic life in
Xenopus, unmatched by the pattern of tectal cell addition
(Gaze et al., 1979; Grant and Keating, 1986b). To preserve the
topography of the map during this period, resident retinal arbors are
forced to shift in an orderly manner across the tectum, changing
postsynaptic partners as they go. It is also in this period that arbors
belonging to the different RGC classes normally become sorted into
distinct strata of the superficial tectum (Chung et al., 1975). The
lack of effect of correlating activity on either the topography or the
stratification of RGC class-specific projections strongly suggests that
the underlying shifts in arbor position are achieved in
Xenopus by activity-independent mechanisms, such as graded
molecular affinities and selective adhesion in the target (see Sanes,
1993; Yamagata et al., 1995). This conclusion accords with reports that
correct retinotectal arbor position, including lamination by RGC type,
can be established independently of patterned activity in other
amphibia (Harris, 1980) and in teleost fish (Stuermer et al., 1990;
Schmidt and Buzzard, 1993); it also is in accord with the specific
finding of Fraser et al. (1984) that brief application of neural cell adhesion molecule antibodies to early juvenile Xenopus
tectum distorts the existing retinal map and almost doubles its MURF sizes.
A small, but consistent, increase in MURF sizes was also present after
synchronizing RGC activity under strobe conditions. As in other frogs
and in goldfish, Xenopus RGC arbors continue remodeling as
they shift across the tectum, retracting older branches and synaptic
contacts while extending new ones that are necessary to preserve their
detailed point-to-point maps (Constantine-Paton et al., 1983; Easter
and Stuermer, 1984; Fujisawa, 1987). The implication that these arbors
are enlarged after strobe rearing suggests that the proper elimination
of older branches and synapses may be repressed when their activity is
highly correlated with that of more topographically appropriate arbors
converging on the same tectal neurons. Such a temporary stabilization
of coactive, although inappropriate, synapses supports the hypothesis
that associative reinforcements of coactive retinal afferents normally contribute to maintaining their mapping precision (Cline and
Constantine-Paton, 1989, 1990; Schmidt and Buzzard, 1993).
We can be confident that strobe-induced activity correlations
were present throughout the development of the ipsilateral map. Crossed
isthmotectal axons normally invade the Xenopus tectum after
stage 50 and only form terminal arborizations at later metamorphic stages (Udin, 1989), coinciding with the appearance of visually driven,
topographically organized ipsilateral units (Grant and Keating, 1989a).
During subsequent postmetamorphic development in both normal and
eye-rotated frogs, crossed isthmotectal arbors shift their position
under the influence of visually evoked activity (Grant and Keating,
1986a, 1989a,b, 1992). These shifts are protracted, serving to maintain
the topography of the map and match its RFs to corresponding retinal
inputs in normal animals in which eye positions are changing
continuously or to reacquire such spatial correspondence after eye
rotation. In normal animals this process probably is accomplished by
systematic remodeling of existing arbor structures, whereas eye
rotation demands radical alterations in their trajectory, apparently
involving a random search strategy (Udin, 1983). Both processes,
however, are thought to be mediated by common mechanisms involving
associative reinforcements or the elimination of crossed isthmotectal
connections on the basis of their temporal activity correlations with
convergent retinal afferents that activate postsynaptic NMDA receptors
(Scherer and Udin, 1989).
Strobe rearing led to defective ipsilateral maps, and it blocked their
complete reorganization in eye-rotated animals. We conclude that both
systematic and radical shifts in crossed isthmotectal arbor position
were disrupted because they had difficulty in distinguishing between
retinal coactivity at inappropriate versus appropriate sites for
binocular convergence. Importantly, these defects have been shown to
resemble those that follow visual deprivation during the same
developmental period, although they are generally less severe and less
consistent. There are two likely explanations for this. First, because
ipsilateral units are capable of responding to 1 Hz strobe flashes,
they probably are activated with a degree of synchrony in the rearing
environment, and it is known that correlated visual activity between
neighboring crossed isthmotectal arbors can preserve their mapping
order even when mismatched with rotated retinal input (Grant and
Keating, 1992; Keating and Grant, 1992). Second, visual contrasts in
the strobe environment may have provided additional spatiotemporal
information for some relatively minor or gradual reorganizations to
occur, such as remodeling of the normal map, but not for the larger
scale plasticity required to compensate fully for an eye rotation. The
nature of the partial plasticity observed in some of the eye-rotated
frogs supports this possibility. As we have noted before (Grant and
Keating, 1992), because crossed isthmotectal arbor dimensions average
~200 µm (Udin, 1989), ipsilateral map reorganization in central
tectum representing the axis of the rotated eye and in the opposite
rostral tectum (Fig. 10) could be achieved by remodeling the existing
arbor framework without the need for changes in trajectory.
Although we have no explanation for why this information apparently was
used more by some normal and eye-rotated animals than by others, a more
interesting question concerns its content. Analyses of contrast-evoked
responses during established strobe entrainment showed that they
reduced the relative level of synchronous activity at all tectal sites
while increasing noncoincident activity over intervals of ± 500 msec, some of which presumably were translated into asynchronous
activation of crossed isthmotectal synapses. By analogy with the limits
of temporal coincidence (up to 50 msec) between the activity of paired
inputs capable of inducing associative LTP in the adult hippocampus
(Levy and Steward, 1983; Gustafsson and Wigstrom, 1986), these
conditions would seem to be inadequate for reinforcing appropriate
crossed isthmotectal synapses. Indeed, given the persistent associative
depression of weaker inputs reported in these and other hippocampal
studies after asynchronous (by 100 msec) afferent activations
(Stanton and Sejnowski, 1989; Otani and Connor, 1995), they might favor
their elimination.
The temporal constraints within which developmental processes of
"coincidence detection" actually operate, however, are unknown, and
so the possibility remains that afferent activity correlations over
hundreds, rather than tens, of milliseconds are of considerable significance. In fact, this possibility is consistent with evidence that NMDA receptor-mediated synaptic currents are this prolonged during
periods of structural synaptic plasticity in the developing visual
system (Carmignoto and Vicini, 1992; Hestrin, 1992) and that binocular
visual stimulation needs to be asynchronous by at least 500 msec to
induce ocular dominance shifts in kitten cortex (Blasdel and Pettigrew,
1979; Altmann et al., 1987). The present demonstration that the
detailed temporal patterning of neural activity within and between the
two eyes influences their synaptic connectivity adds urgency to future
characterization of the coincidence sensitivity of developing visual
neurons.
| |
FOOTNOTES |
Received Sept. 16, 1997; revised Nov. 21, 1997; accepted Dec. 2, 1997.
This research was supported by the Medical Research Council.
Correspondence should be addressed to Dr. Simon Grant, Department of
Sensorimotor Systems, Division of Neuroscience, Imperial College School
of Medicine, Fulham Palace Road, London W6 8RF, UK.
Dr. Brickley's present address: Department of Pharmacology, University
College London, Gower Street, London WC1E 6BT, UK.
| |
REFERENCES |
-
Aertson AMHJ,
Gerstein GL,
Habib MK,
Palm G
(1989)
Dynamics of neuronal firing correlation: modulation of "effective connectivity."
J Neurophysiol
61:900-917[Abstract/Free Full Text].
-
Altmann L,
Luhmann HJ,
Greuel JM,
Singer W
(1987)
Functional and neuronal binocularity in kittens raised with rapidly alternating monocular occlusion.
J Neurophysiol
58:965-980[Abstract/Free Full Text].
-
Berman NEJ,
Payne BR
(1985)
An exuberant retinocollicular pathway in Siamese kittens: effects of competition and abnormal activity on its maturation.
Dev Brain Res
22:197-209.
-
Blasdel GG,
Pettigrew JD
(1979)
Degree of interocular synchrony required for maintenance of binocularity in kitten's visual cortex.
J Neurophysiol
42:1692-1710[Abstract/Free Full Text].
-
Carmignoto G,
Vicini S
(1992)
Activity-dependent decrease in NMDA receptor responses during development of the visual cortex.
Science
258:1007-1011[Abstract/Free Full Text].
-
Chung S-H,
Stirling V,
Gaze RM
(1975)
The structural and functional development of the retina in larval Xenopus.
J Embryol Exp Morphol
33:915-940[Web of Science][Medline].
-
Cline HT,
Constantine-Paton M
(1989)
NMDA receptor antagonists disrupt the retinotectal topographic map.
Neuron
3:413-426[Web of Science][Medline].
-
Cline HT,
Constantine-Paton M
(1990)
NMDA receptor agonist and antagonists alter retinal ganglion cell arbor structure in the developing frog retinotectal projection.
J Neurosci
10:1197-1216[Abstract].
-
Cline HT,
Debski E,
Constantine-Paton M
(1987)
NMDA receptor antagonist desegregates eye-specific stripes.
Proc Natl Acad Sci USA
84:4342-4345[Abstract/Free Full Text].
-
Constantine-Paton M,
Pitts EC,
Reh TA
(1983)
The relationship between retinal axon ingrowth, terminal morphology, and terminal patterning in the optic tectum of the frog.
J Comp Neurol
218:297-313[Web of Science][Medline].
-
Constantine-Paton M,
Cline HT,
Debski E
(1990)
Patterned activity, synaptic convergence, and the NMDA receptor in developing visual pathways.
Annu Rev Neurosci
13:129-154[Web of Science][Medline].
-
Cook JE
(1987)
A sharp retinal image increases the topographic precision of the goldfish retinotectal projection during optic nerve regeneration in stroboscopic light.
Exp Brain Res
68:319-328[Web of Science][Medline].
-
Cook JE,
Rankin ECC
(1986)
Impaired refinement of the regenerated retinotectal projection of the goldfish in stroboscopic light: a quantitative WGA-HRP study.
Exp Brain Res
63:421-430[Web of Science][Medline].
-
Easter SS,
Stuermer CAO
(1984)
An evaluation of the hypothesis of shifting terminals in goldfish optic tectum.
J Neurosci
4:1052-1063[Abstract].
-
Fisher LJ
(1976)
Synaptic arrays of the inner plexiform layer in the developing retina of Xenopus.
Dev Biol
50:402-412[Web of Science][Medline].
-
Fraser SE,
Murray BA,
Chuong C-M,
Edelman GM
(1984)
Alteration of the retinotectal map in Xenopus by antibodies to neural cell adhesion molecules.
Proc Natl Acad Sci USA
81:4222-4226[Abstract/Free Full Text].
-
Fujisawa H
(1987)
Mode of growth of retinal axons within the tectum of Xenopus tadpoles, and implications for the ordered neuronal connections between the retina and tectum.
J Comp Neurol
260:127-139[Web of Science][Medline].
-
Gaze RM,
Keating MJ,
Chung S-H
(1974)
The evolution of the retinotectal map during development in Xenopus.
Proc R Soc Lond [Biol]
185:301-330[Medline].
-
Gaze RM,
Keating MJ,
Ostberg A,
Chung S-H
(1979)
The relationship between retinal and tectal growth in larval Xenopus: implications for the development of the retinotectal projection.
J Embryol Exp Morphol
53:103-143[Web of Science][Medline].
-
George SA,
Marks W
(1974)
Optic nerve terminal arborizations in the frog: shape and orientation inferred from electrophysiological measurements.
Exp Neurol
42:467-482[Web of Science][Medline].
-
Goodman CS,
Shatz CJ
(1993)
Developmental mechanisms that generate precise patterns of neuronal connectivity.
Cell
72:77-98.
-
Grant S,
Keating MJ
(1986a)
Normal maturation involves systematic changes in binocular visual connections in Xenopus laevis.
Nature
322:258-261.
-
Grant S,
Keating MJ
(1986b)
Ocular migration and the metamorphic and postmetamorphic maturation of the retinotectal system in Xenopus laevis: an autoradiographic and morphometric study.
J Embryol Exp Morphol
92:43-69[Web of Science][Medline].
-
Grant S,
Keating MJ
(1989a)
Changing patterns of binocular visual connections in the intertectal system during development of the frog, Xenopus laevis. I. Normal maturational changes in response to changing binocular geometry.
Exp Brain Res
75:99-116[Web of Science][Medline].
-
Grant S,
Keating MJ
(1989b)
Changing patterns of binocular visual connections in the intertectal system during development of the frog, Xenopus laevis. II. Abnormalities following early visual deprivation.
Exp Brain Res
75:117-132[Web of Science][Medline].
-
Grant S,
Keating MJ
(1992)
Changing patterns of binocular visual connections in the intertectal system during development of the frog, Xenopus laevis. III. Modifications following early eye rotation.
Exp Brain Res
89:383-396[Web of Science][Medline].
-
Grigonis AM,
Murphy EH
(1991)
Organization of callosal connections in the visual cortex of the rabbit following neonatal enucleation, dark rearing, and strobe rearing.
J Comp Neurol
312:561-572[Web of Science][Medline].
-
Gustafsson B,
Wigstrom H
(1986)
Hippocampal long-lasting potentiation produced by pairing single volleys and brief conditioning tetani evoked in separate afferents.
J Neurosci
6:1575-1582[Abstract].
-
Harris WA
(1980)
The effects of eliminating impulse activity on the development of the retinotectal projection in salamanders.
J Comp Neurol
194:303-317[Web of Science][Medline].
-
Hestrin S
(1992)
Developmental regulation of NMDA receptor-mediated synaptic currents at a central synapse.
Nature
357:686-689[Medline].
-
Holt CE,
Harris WA
(1983)
Order in the initial retinotectal map in Xenopus: a new technique for labeling growing nerve fibres.
Nature
301:150-152[Medline].
-
Keating MJ,
Feldman JD
(1975)
Visual deprivation and intertectal neuronal connections in Xenopus laevis.
Proc R Soc Lond [Biol]
191:467-474[Medline].
-
Keating MJ,
Grant S
(1992)
The critical period for experience-dependent plasticity in a system of binocular visual connections in Xenopus laevis: its temporal profile and relation to normal developmental requirements.
Eur J Neurosci
4:27-36[Web of Science][Medline].
-
Keating MJ,
Kennard C
(1987)
Visual experience and the maturation of the ipsilateral visuotectal projection in Xenopus laevis.
Neuroscience
21:519-527[Web of Science][Medline].
-
Keating MJ,
Grant S,
Dawes EA,
Nanchahal K
(1986)
Visual deprivation and the maturation of the retinotectal projection in Xenopus laevis.
J Embryol Exp Morphol
91:101-115[Web of Science][Medline].
-
Levy WB,
Steward O
(1983)
Temporal contiguity requirements for long-term associative potentiation/depression in the hippocampus.
Neuroscience
8:791-797[Web of Science][Medline].
-
Maffei L,
Galli-Resta L
(1990)
Correlation in the discharges of neighboring rat retinal ganglion cells during prenatal life.
Proc Natl Acad Sci USA
87:2861-2864[Abstract/Free Full Text].
-
Maturana HR,
Lettvin JY,
McCulloch WS,
Pitts WH
(1960)
Anatomy and physiology of vision in the frog (Rana pipiens).
J Gen Physiol
43:129-171[Free Full Text].
-
Meister M,
Wong ROL,
Baylor DA,
Shatz CJ
(1991)
Synchronous bursts of action potentials in ganglion cells of the developing mammalian retina.
Science
252:939-943[Abstract/Free Full Text].
-
Nieuwkoop PD,
Faber J
(1967)
In: A normal table of Xenopus laevis (Daudin), 2nd Ed . Amsterdam: North-Holland.
-
O'Rourke NA,
Fraser SE
(1986)
Dynamic aspects of retinotectal map formation revealed by a vital dye fiber-tracing technique.
Dev Biol
114:265-276[Web of Science][Medline].
-
O'Rourke NA,
Cline HT,
Fraser SE
(1994)
Rapid remodeling of retinal arbors in the tectum with and without blockade of synaptic transmission.
Neuron
12:921-934[Web of Science][Medline].
-
Otani S,
Connor JA
(1995)
Long-term depression of naive synapses in adult hippocampus induced by asynchronous synaptic activity.
J Neurophysiol
73:2596-2601[Abstract/Free Full Text].
-
Rauschecker JP
(1991)
Mechanisms of visual plasticity: Hebb synapses, NMDA receptors, and beyond.
Physiol Rev
71:587-615[Free Full Text].
-
Reh TA,
Constantine-Paton M
(1985)
Eye-specific segregation requires neural activity in three-eyed Rana pipiens.
J Neurosci
5:1132-1143[Abstract].
-
Sakaguchi DS,
Murphey RK,
Hunt RK,
Tompkins R
(1984)
The development of retinal ganglion cells in a tetraploid strain of Xenopus laevis: a morphological study utilizing intracellular dye injection.
J Comp Neurol
224:231-251.
-
Sanes JR
(1993)
Topographic maps and molecular gradients.
Curr Opin Neurobiol
3:67-74[Medline].
-
Scherer WJ,
Udin SB
(1989)
N-methyl-D-aspartate antagonists prevent interaction of binocular maps in Xenopus tectum.
J Neurosci
9:3837-3843[Abstract].
-
Scherer WJ,
Udin SB
(1991)
Latency and temporal overlap of visually elicited contralateral and ipsilateral firing in Xenopus tectum during and after the critical period.
Dev Brain Res
58:129-132[Medline].
-
Schmidt JT,
Buzzard M
(1993)
Activity-driven sharpening of the retinotectal projection in goldfish: development under stroboscopic illumination prevents sharpening.
J Neurobiol
24:384-399[Web of Science][Medline].
-
Schmidt JT,
Eisele LE
(1985)
Stroboscopic illumination and dark-rearing block the sharpening of the regenerated retinotectal map in goldfish.
Neuroscience
14:535-546[Web of Science][Medline].
-
Sernagor E,
Grzywacz NM
(1996)
Influence of spontaneous activity and visual experience on developing retinal receptive fields.
Curr Biol
6:1503-1508[Web of Science][Medline].
-
Stanton PK,
Sejnowski TJ
(1989)
Associative long-term depression in the hippocampus induced by Hebbian covariance.
Nature
339:215-218[Medline].
-
Stiles M,
Tzanakou E,
Michalak R,
Unnikrishnan KP,
Goyal P,
Harth E
(1985)
Periodic and nonperiodic burst responses of frog (Rana pipiens) retinal ganglion cells.
Exp Neurol
88:176-197[Web of Science][Medline].
-
Stuermer CAO,
Rohrer B,
Munz H
(1990)
Development of the retinotectal projection in zebrafish embryos under TTX-induced neural impulse blockade.
J Neurosci
10:3615-3626[Abstract].
-
Udin SB
(1983)
Abnormal visual input leads to development of abnormal axon trajectories in frogs.
Nature
301:336-338[Medline].
-
Udin SB
(1989)
Development of the nucleus isthmi in Xenopus. II. Branching patterns of contralaterally projecting isthmotectal axons during maturation of binocular maps.
Vis Neurosci
2:153-163[Web of Science][Medline].
-
Udin SB,
Fawcett JW
(1988)
Formation of topographic maps.
Annu Rev Neurosci
11:289-327[Web of Science][Medline].
-
Udin SB,
Keating MJ
(1981)
Plasticity in a central nervous pathway in Xenopus: anatomical changes in the isthmotectal projection after larval eye rotation.
J Comp Neurol
203:575-594[Web of Science][Medline].
-
Witkovsky P,
Gallin E,
Hollyfield JG,
Ripps H,
Bridges CDB
(1976)
Photoreceptor thresholds and visual pigment levels in normal and vitamin A-deprived Xenopus tadpoles.
J Neurophysiol
39:1272-1287[Abstract/Free Full Text].
-
Wong ROL,
Meister M,
Shatz CJ
(1993)
Transient period of correlated bursting activity during development of the mammalian retina.
Neuron
11:923-938[Web of Science][Medline].
-
Yamagata M,
Herman J-P,
Sanes JR
(1995)
Lamina-specific expression of adhesion molecules in developing chick optic tectum.
J Neurosci
15:4556-4571[Abstract].
Copyright © 1998 Society for Neuroscience 0270-6474/98/1841491-14$05.00/0
This article has been cited by other articles:
|
|
|
|
 
P. J. Drew and D. E. Feldman
Intrinsic Signal Imaging of Deprivation-Induced Contraction of Whisker Representations in Rat Somatosensory Cortex
Cereb Cortex,
February 1, 2009;
19(2):
331 - 348.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
L. I. Zhang, S. Bao, and M. M. Merzenich
Disruption of primary auditory cortex by synchronous auditory inputs during a critical period
PNAS,
February 12, 2002;
(2002)
261707398.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
Y. Guo and S. B. Udin
The Development of Abnormal Axon Trajectories after Rotation of One Eye in Xenopus
J. Neurosci.,
June 1, 2000;
20(11):
4189 - 4197.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
L. I. Zhang, S. Bao, and M. M. Merzenich
Disruption of primary auditory cortex by synchronous auditory inputs during a critical period
PNAS,
February 19, 2002;
99(4):
2309 - 2314.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
Top
Abstract
Introduction
