Full text of DOI:10.1017/S0952523803203011

Visual Neuroscience (2003), 20, 221–230. Printed in the USA.
Copyright © 2003 Cambridge University Press 0952-5238003 $16.00
DOI: 10.10170S0952523803203011
Gain control from beyond the classical receptive field
in primate primary visual cortex
BEN S. WEBB, CHRIS J. TINSLEY, NICK E. BARRACLOUGH, AMANDA PARKER,
and ANDREW M. DERRINGTON
School of Psychology, University of Nottingham, University Park, Nottingham, NG7 2RD, UK
(Received September 6, 2002; Accepted April 8, 2003)
Abstract
Gain control is a salient feature of information processing throughout the visual system. Heeger (1991, 1992)
described a mechanism that could underpin gain control in primary visual cortex (V1). According to this model, a
neuron’s response is normalized by dividing its output by the sum of a population of neurons, which are selective
for orientations covering a broad range. Gain control in this scheme is manifested as a change in the semisaturation
constant (contrast gain) of a V1 neuron. Here we examine how flanking and annular gratings of the same or
orthogonal orientation to that preferred by a neuron presented beyond the receptive field modulate gain in V1
neurons in anesthetized marmosets (Callithrix jacchus). To characterize how gain was modulated by surround
stimuli, the Michaelis–Menten equation was fitted to response versus contrast functions obtained under each
stimulus condition. The modulation of gain by surround stimuli was modelled best as a divisive reduction in
response gain. Response gain varied with the orientation of surround stimuli, but was reduced most when the
orientation of a large annular grating beyond the classical receptive field matched the preferred orientation of
neurons. The strength of surround suppression did not vary significantly with retinal eccentricity or laminar
distribution. In the marmoset, as in macaques (Angelucci et al., 2002a,b), gain control over the sort of distances
reported here (up to 10 deg) may be mediated by feedback from extrastriate areas.
Keywords: Marmoset, Primary visual cortex, Receptive field, Gain, Context
Introduction
neurons with large receptive fields—may be carried out as early as
V1. Several studies with higher order stimuli have shown that
interactions between stimuli inside and outside the receptive field
of V1 neurons may form the substrate for contour integration
The classical receptive field has long been regarded as the funda-
mental functional unit in primary visual cortex (V1) (Hubel &
Wiesel, 1962, 1968). Numerous physiological studies have shown,
however, that descriptions of the classical receptive field may also
need to take account of input from beyond the classical receptive
field (for reviews see, Allman et al., 1985; Fitzpatrick, 2000;
Lamme & Spekreijse, 2000; Wörgötter & Eysel, 2000; Albright &
Stoner, 2002). A stimulus beyond the classical receptive field is, by
definition, not able to drive a V1 neuron, but typically it can
modulate its response (Blakemore & Tobin, 1972; Maffei & Fioren-
tini, 1976; Nelson & Frost, 1978; Gilbert & Wiesel, 1990; DeAn-
gelis et al., 1994; Li & Li, 1994; Sillito et al., 1995; Levitt & Lund,
(
Kapadia et al., 1995), perceptual grouping (Mizobe et al., 2001),
perceptual “pop-out” (Kastner et al., 1997; Nothdurft et al., 1999),
and figure-ground segregation (Lamme, 1995; Zipser et al., 1996,
but see Rossi et al., 2001). These perceptual phenomena tend to be
represented in V1 as a relative increase in neural activity.
Research on the properties of surround interactions has shown
that the response of a V1 neuron can be increased or decreased by
changing the relative contrast and0or orientation (Toth et al., 1996;
Levitt & Lund, 1997; Polat et al., 1998; Chen et al., 2001; Mizobe
et al., 2001) or direction of motion (Palmer & Nafziger, 2002)
inside and outside the receptive field. However, there is skepticism
whether increases in the activity of V1 neurons due to the presence
of surround stimuli genuinely originate from beyond the classical
receptive field (Walker et al., 2000). Because of the dynamic
nature of the receptive field (Kapadia et al., 1999; Sceniak et al.,
1997; Polat et al., 1998; Sugita, 1999; Walker et al., 1999; Palmer
&
Nafziger, 2002).
V1 neurons are thus able to integrate information over large
areas, and may be able to perform much more complex operations
than previously thought. Some higher order visual processing—
traditionally postulated as an emergent property of extrastriate
1999) and underestimation of its size in some studies (Walker
et al., 2000), such increases in activity may be explained by
surround stimuli encroaching onto and activating the receptive
field (Walker et al., 2000; Cavanaugh et al., 2002a,b). Much more
robust is the finding that surround interactions in V1 are predom-
Address correspondence and reprint requests to: Ben Webb, School of
Psychology, University of Nottingham, University Park, Nottingham, NG7
2
RD, UK. E-mail: bsw@psychology.nottingham.ac.uk
221

222
B.S. Webb et al.
inantly suppressive, and maximal when surround stimuli match
stimuli on the receptive field (DeAngelis et al., 1994; Walker et al.,
solution, at a rate of 1.5 ml0kg0h. Skeletal muscles were paralyzed
Ϫ1 Ϫ1
with vecuronium bromide (0.1 mg{kg
h
). End-expired CO
2
2
000; Cavanaugh et al., 2002b).
Little attention has been given to the function(s) served by
was maintained at between 4 and 5% in the expired air by
adjusting the stroke volume of the animal. Body temperature was
maintained close to 37.58C by an electric blanket controlled by a
rectal thermistor. An appropriate depth of anesthesia was main-
tained by continuous monitoring of electrocardiogram and electro-
encephalogram activity. Supplementary anesthesia (fluothane) was
given if necessary.
input from beyond the classical receptive field. Gain control,
which is fundamental to information processing in V1, is a strong
candidate (Albrecht & Hamilton, 1982; Ohzawa et al., 1985; Sclar
et al., 1989; Heeger, 1991, 1992; Geisler & Albrecht, 1992;
Carandini et al., 1997; Truchard et al., 2000). Heeger (1991,1992)
proposed a mechanism for contrast gain control in V1. The model
posits that the range of contrasts over which a V1 neuron responds
is reset by dividing its output by the total output of a nearby
ensemble of neurons. Contrast gain control scales the semisatura-
tion constant by a factor k; put another way, effective stimulus
contrast is changed by 10k. Moreover, according to this model,
gain controls signals are pooled over a broad range of orientations
and spatial frequencies (Heeger, 1991, 1992).
Sengpiel et al. (1998) tested whether a range of inhibitory
phenomena could be explained in terms of a single divisive
gain-control mechanism. Contrary to the predictions of Heeger’s
model, they found that surround suppression, cross-orientation
suppression, and interocular suppression, in cat V1, were best
explained by different inhibitory mechanisms. A response gain
model provided the best description of surround suppression, a
contrast gain model described cross-orientation suppression best,
and a combined response gain and contrast gain model provided
the best description of interocular suppression (Sengpiel et al.,
Physiological recording
A craniotomy was made over V1 and a dural flap was reflected to
expose the surface of the cortex. Extracellular responses of single
neurons were recorded with either glass-insulated tungsten elec-
trodes (Merrill & Ainsworth, 1972) or epoxy-coated tungsten
electrodes (FHC Inc., Maine). The signal from the electrode was
amplified, band-pass filtered, sampled, and time-stamped with a
resolution of 100 ms. Single units were isolated by matching
incoming spikes to a template constructed from the shape of action
potential traces. To prevent respiratory movement of the brain
during recording sessions, the craniotomy was filled with agar (3%
in saline) and sealed with candle wax.
Optics
The pupils were dilated with atropine sulphate and the eyes were
protected by gas-permeable contact lenses of zero-added power.
The refractive error of each eye was corrected with miniature
spectacle lenses that optimized the response of an isolated neuron
to a high spatial frequency sine-wave grating. The optic disk and
fovea of each eye were plotted on a tangent screen 57 cm in front
of the animal, with a reversing ophthalmoscope.
1
998). However, surround modulation of gain was measured only
with annular gratings of the same orientation as that preferred by
neurons. It is not clear from this study whether gain-control signals
from beyond the classical receptive field are pooled over a range
of orientations or activated only by a restricted range of orientations.
In the current experiment, we examine how flanking and
annular gratings of the same and orthogonal orientation to that
preferred by neurons modulate the gain of V1 neurons in anesthe-
tized marmosets. In the lateral geniculate nucleus (LGN) of the
marmoset, we found that large surround stimuli of the same or
orthogonal orientation to that on the receptive field reduced re-
sponse gain in a divisive fashion (Webb et al., 2002). In the LGN,
gain control is probably mediated by the feedback projection from
V1 (Przybyszewski et al., 2000; Webb et al., 2002). Here we find
that gain-control signals from beyond the classical receptive field
of V1 neurons are also modelled best as a reduction in response
gain. We also found that response gain was not the same for all
orientations: gain was reduced most when the orientation of sur-
round stimuli matched the preferred orientation of the neuron.
Visual stimuli
Stimuli were generated by a Macintosh computer using a Radius
10-bit graphics card and presented initially on a tangent projection
screen, subtending approximately 87 deg ϫ 67 deg at a viewing
distance of 57 cm. On the tangent projection screen, we searched
for cells with a large patch of drifting sinusoidal grating that varied
its orientation. Receptive fields were mapped with a drifting
grating at the preferred orientation, and then positioned on the
center of a CRT display monitor with a front-surfaced mirror. The
CRT display (Sony Model No. GDM 200PST) subtended 15.5
deg ϫ 11.5 deg at a viewing distance of 114 cm, had a mean
2
luminance of approximately 50 cd0m , and a frame rate of 120 Hz.
The display nonlinearity was corrected using a lookup table.
Contrast of visual stimuli was specified by Michelson contrast
Materials and methods
~
Lmax Ϫ Lmin)0~Lmax ϩ Lmin).
On the CRT display, we used drifting sinusoidal gratings to
Animal preparation
Extracellular recordings were made in four adult marmosets (Cal-
lithrix jacchus) that weighed between 362–414 g. One of the
animals was male and the others were female. All surgical and
preparatory procedures were in accordance with the guidelines of
the UK Animals (Scientific Procedures) Act of 1986.
map the receptive field and obtain tuning curves for each neuron.
We mapped receptive fields by measuring the response of the
neuron as the length and width and position of a patch of grating
were independently changed. The position of a 0.5–2 deg patch of
grating that evoked the maximum obtainable response was desig-
nated as the center of the receptive field. All stimuli used here were
centered on this location. We regularly checked the centering of the
receptive field. If it shifted from its original position, it was
recentered using the same method.
Animals were initially anesthetized with Saffan (Alphadalone0
Alphaxalone acetate; 1.5 ml0kg i.m.). The lateral tail veins and the
trachea were cannulated. In the male the urethra was catheterized.
Animals were artificially respired with a N O (70%), O (30%)
2
2
mixture. Surgical anesthesia was maintained with a venous infu-
Spatial-frequency, temporal-frequency, orientation, and size tun-
ing curves were obtained in separate tests by varying a circular
Ϫ1 Ϫ1
sion of fentanyl citrate (20 mg{kg
h
) in a saline-glucose

Surround modulation of gain in V1
223
(2)
patch of drifting grating along each of these dimensions. The peak
values for each of these measures were used in subsequent tests.
We then compared the response of the neuron to an annular patch
of drifting grating of different inner diameters and a blank screen
of the same mean luminance. The inner diameter of the annular
grating that evoked no response from the neuron was designated as
the diameter of the summation field.
In the experiment reported here a drifting sinusoidal grating, at
the preferred orientation, spatial frequency, and temporal fre-
quency of the neuron, was presented within the classical receptive
field for 1000 ms. The grating could have several different con-
trasts, covering the range 0–1, and was presented alone or sur-
rounded by a drifting grating of the same or orthogonal orientation
contained within either (1) a larger annular field that was 10 deg ϫ
~RBD Ϫ ROD!0~RBD ϩ ROD!,
where RBD is the response to the preferred direction of motion
minus the spontaneous activity, and ROD is the response to the
opposite direction of motion minus the spontaneous activity. Cells
were designated as direction selective if the index of direction
selectivity was Ͼ0. 3^ .
Results
The data described here were obtained from 34 simple and 33
complex cells in primary visual cortex (V1) of the marmoset.
Neurons were recorded at eccentricities between 1.02 and 9.9 deg.
Fig. 1 shows typical responses of a simple cell (left column) and
complex cell (right column). The top row of Fig. 1 illustrates that
a grating drifting across the receptive field at the preferred orien-
tation, spatial frequency, and temporal frequency evoked a re-
sponse that peaks at 180 impulses0s from the simple cell (Fig. 1A
left) and one that peaks at 120 impulses0s from the complex cell
10 deg in size, (2) flanks at the preferred orientation that were the
length of the receptive field and 10 deg wide, or (3) flanks at the
orthogonal orientation that were the width of the receptive field
and 10 deg in length. The spatial phase of surround stimuli was the
same as the stimulus on the classical receptive field.
The different conditions were presented 60 times (10 repeti-
tions of 6 contrast levels of the grating presented to the classical
receptive field), in an interleaved fashion with interstimulus inter-
vals of 500 ms.
(
Fig. 1A right). The presence of an annular grating beyond the
classical receptive field of the same orientation as that preferred by
the neuron reduced the peak response of the simple cell by
approximately 130 impulses0s (Fig. 1B left), and of the complex
cell by about 60 impulse0s. When an annular grating was pre-
sented alone, it evoked no response from the simple cell (compare
Fig. 1C left with Fig. 1D left) and suppressed the spontaneous
activity of the complex cell (compare Fig. 1C right with Fig. 1D
right).
Histology and track tracing
Microlesions (5–10 mA for 5–10 s, electrode negative) were made
at different depths on each penetration. These were used to recon-
struct each electrode track and assign cells to layers. At the end of
each experiment an overdose of pentobarbitone (Sagatal; 60 mg0
kg) was given. Once electrocardiogram, electroencephalogram,
Surround modulation of gain
and CO traces were flat, animals were perfused through the left
2
The primary aim of the study was to examine how flanking and
annular stimuli of the same and orthogonal orientation to the
receptive field presented beyond the receptive field modulated the
gain of responses to an optimal stimulus on the receptive field. To
characterize how different surround stimuli modulated gain, the
Michaelis–Menten equation was fitted to response versus contrast
functions obtained under seven stimulus conditions (Fig. 2A). We
ventricle, initially with phosphate buffer, and then with 4% form-
aldahyde. The brain was removed and stored in a 30% sucrose
solution until it sank. Sagittal sections were taken every 60 mm
with a freezing microtome; sections were mounted and stained
with cresyl violet.
Data analysis
estimated values of each of the parameters Rmax, c , and n that
50
were required to produce curves that fit the neuron’s response
versus contrast function under the seven stimulus conditions. For
each neuron, the seven curves were fit simultaneously and two
parameters were constrained to have the same value for all seven
curves, while the third parameter and M were allowed to vary from
curve to curve. We compared the fits produced by varying each of
the three curve parameters, Rmax, c , and n. Allowing c to vary
To quantify the differential modulation of the response versus
contrast function by stimuli beyond the classical receptive field,
the response of a neuron to each stimulus configuration was
compared, by fitting the Michaelis–Menten equation to each cell’s
response versus contrast function by minimizing the squared error.
The equation is
50
50
explained more variance than allowing n to vary, but allowing
Rmax to vary explained the most variance (see Table 1). The mean
fraction of variance explained by allowing Rmax to vary was 91%.
More than 90% of the variance was explained in 30 (88%) simple
cells and 27 (82%) complex cells. The following analysis is
confined to Rmax values obtained from these neurons.
Figs. 2B and 2C are examples of constrained fits to data
obtained under seven stimulus conditions from a simple and a
complex cell, respectively. In both examples, surround stimuli
reduced Rmax to the grating on the classical receptive field by
different amounts. Rmax was reduced most when the surround
stimulus was a large annular grating of the same orientation as the
grating on the classical receptive field. The response was reduced
by up to 75% of the Rmax in the simple cell, and by up to 60% in
the complex cell. There is no evidence that surround stimulation
n
n
n
R ϭ Rmax c 0~c ϩ c ! ϩ M.
(1)
50
R
max is the maximum attainable response, c50 is the contrast at
which the response reached half its maximum value, n indicates
the steepness of the curve, and M is the spontaneous firing rate. We
used the Nelder–Mead simplex search method to generate the
curve fits. The fraction of variance explained by the curve fits was
used as a goodness-of-fit criterion.
Cells were designated as simple or complex by calculating the
ratio of the first harmonic ~F ) and mean of the response ~F ) to an
1
0
optimal drifting grating (Skottun et al., 1991). F
1
and F
0
amplitude
were used as the measure of a simple and complex cell’s response,
respectively. We calculated an index of direction selectivity with
the following equation:

2
24
B.S. Webb et al.
(
Figs. 3A, 3B, 3D, & 3E), the Rmax of some neurons was usually
reduced but rarely increased. When the orientation of an annular
grating was orthogonal to that of the grating on the classical
receptive field, the Rmax was reduced in a small number of cells
(
Fig. 3F). When the orientation of an annular grating matched that
of the grating on the classical receptive field, the Rmax of most
neurons was reduced.
Selectivity and spatial distribution of surround
interactions
To examine whether particular stimuli at different locations be-
yond the classical receptive field were more effective than others
in modulating the gain of V1 neurons, we calculated a Rmax ratio:
@~RmaxEC Ϫ RmaxC !0RmaxC #.
(3)
RmaxEC is the maximum firing rate obtained during a surround
condition and RmaxC is the maximum firing rate obtained when
only the classical receptive field was stimulated. A ratio of 0
indicates that the Rmax was unaffected by the presence of a
surround stimulus, a positive ratio indicates that the Rmax was
increased by the presence of a surround stimulus, and a negative
ratio indicates that it was reduced by the presence of a surround
stimulus. Histograms of the Rmax ratio for each surround stimulus
condition are plotted in Fig. 4. The main effect of surround
stimulation was to reduce the Rmax ratio. The proportion of neu-
rons in which the reduction in the Rmax ratio was not significant
(i.e. the reduction was less than 10%) in each stimulus condition
depicted in Fig. 4 was as follows: panel A, 30057, panel B, 33057,
panel C, 20057, panel D, 40057, panel E, 39057, panel F, 33057. It
was increased for a minority of cells, but this increment does not
appear related to any particular stimulus configuration. One-
sample t-tests revealed that the mean Rmax ratio was significantly
less than zero in all stimulus conditions except that depicted in
panel E of Fig. 4. Surround gratings with the same orientation as
the grating on the classical receptive field reduced the mean Rmax
ratio more (panel A, P Ͻ 0.001, mean Ϫ0.1, SEM 6 0.03; panel
B, P Ͻ 0.001, mean Ϫ0.1, SEM 6 0.03; panel C, P Ͻ 0.000001,
mean Ϫ0.26, SEM 6 0.04) than surround gratings with a different
orientation (panel D, P Ͻ 0.01, mean Ϫ0.08, SEM 6 0.02; panel
E, P ϭ 0.125, mean Ϫ0.03, SEM 6 0.02 ; panel F, P Ͻ 0.01, mean
Ϫ0.08, SEM 6 0.03). But, the mean Rmax ratio was reduced most
by large surround stimuli of the same orientation as that preferred
by neurons (panel C, P Ͻ 0.000001, mean Ϫ0.26, SEM 6 0.04).
The mean Rmax ratio and standard error of the mean in each
surround stimulus condition for simple and complex cells are
plotted in Fig. 5A. The Rmax ratio for both simple and complex
cells was reduced most by annular gratings of the same orientation
as the grating on the receptive field. Both flanking and annular
gratings of the same orientation as that on the receptive field
seemed to be more effective in reducing the Rmax ratio of complex
cells than simple cells. A 2 (cell type: simple vs. complex) ϫ 2
(grating orientation: horizontal vs. vertical) ANOVA revealed there
was a main effect of orientation @F~1,169! ϭ 25.53, P Ͻ 0.000001]
and an interaction between cell type and grating orientation
@F~1,169! ϭ13.04, P Ͻ 0.001]. This analysis shows that surround
gratings of the same orientation as the grating on the receptive
field were more effective than surround gratings of the orthogonal
orientation at reducing the Rmax ratio of V1 neurons. The inter-
action indicates that the Rmax ratio in complex cells was reduced
Fig. 1. The response of a simple cell (left column) and complex cell (right
column) is reduced by the presence of an annular grating beyond the
classical receptive field. The top row shows the response of a simple cell
(
Fig. 1A left) and complex cell (Fig. 1A right) to a grating drifting across
the classical receptive field at the optimal orientation, spatial frequency,
and temporal frequency. The presence of a large annular grating at the same
orientation reduced the peak response of the simple cell by approximately
1
30 impulses0s (Fig. 1B left) and of the complex cell by approximately 60
impulses0s (Fig. 1B right). The presence of the annular grating alone
evoked no response from the simple cell (compare Fig. 1C left with Fig. 1D
left), and suppressed the spontaneous activity of the complex cell (compare
Fig. 1C right with Fig. 1D right).
increases the response to the grating on the classical receptive
field.
The Rmax values obtained in each stimulus condition for each
neuron are plotted in Figs. 3A–3F. In each graph, Rmax values for
simple and complex cells obtained in the grating-alone condition
are plotted against the Rmax values obtained in each of the surround
stimulus conditions. Data points above the diagonal indicate that
the Rmax was reduced by the presence of a surround stimulus. Data
points below the diagonal indicate that the Rmax was increased by
the presence of a surround stimulus. In the flanking conditions

Surround modulation of gain in V1
225
Fig. 2. Effects of each surround stimulus (A) on response gain
of a simple cell (B) and complex cell (C). In both cells, the
seven curves are best-fitting results of eq. (1), with the
parameters fit simultaneously with the following constraints.
Parameters c50 and n were constrained to have the same value
for all seven curves while Rmax and M were allowed to vary
from curve to curve. The presence of a large annular grating
of the same orientation as the grating on the classical recep-
tive field was most effective in reducing the response gain of
these neurons.
more than the Rmax ratio in simple cells by surround gratings of the
same orientation as the grating on the receptive field.
receptive field. To test whether there was a locus of surround
suppression, we examined whether the strength of suppression
originating from the ends of the receptive field and the sides of the
receptive field were correlated. We consider only cells for which
the Rmax ratio was clearly reduced (i.e. by more than 10%) by the
presence of a large annular grating of the same orientation as that
preferred by the neuron. There was no evidence of a statistically
significant correlation between the reductions in the Rmax ratio
caused by stimulating the ends and sides of the receptive field with
gratings of the same orientation as that on the receptive field
[Pearson’s r~38! ϭ 0.08, P ϭ 0.65]. This suggests that suppression
may originate from localized regions beyond the classical recep-
tive field.
An alternative measure of gain to changes in Rmax is the slope
of the response versus contrast function, expressed as the incre-
ment in firing rate for each 1% increase in contrast. For compar-
ison, we conducted the same analysis as above on the average
slope of the response versus contrast function measured over the
range of contrasts from 0% to 25%, for each surround stimulus
condition. To show how this measure of gain was modulated by
surround stimulation, these values were normalized, by subtracting
the slope measured in the absence of a surround grating. The mean
reduction in gain and standard error of the mean are plotted in
Fig. 5B. An ANOVA showed that there was a main effect of grating
orientation @F~1,169! ϭ 22.79, P Ͻ 0.00001] and an interaction
between cell type and grating orientation @F~1,169! ϭ 12.36, P Ͻ
Direction-selective and bi-directional neurons
0.002]. As was found with the Rmax ratio, surround gratings with
the same orientation as the grating on the receptive field were more
effective than orthogonal gratings at reducing the gain of V1
neurons. The gain of complex cells was also reduced more than
simple cells by gratings of the same orientation as the classical
receptive field.
Our sample consisted of 38 direction-selective neurons, 15 bi-
directional neurons, and four neurons that were not selective for
orientation or direction of motion. To ensure that the surround
effects reported here are not due to this sampling bias, we con-
ducted the same analyses as above on direction-selective and
bi-directional cells. The mean Rmax ratio and standard error of the
mean and mean reduction in gain (as above, measured at 25%
contrast) and standard error of the mean, for each surround stim-
ulus condition for direction and bi-directional cells, are plotted in
Figs. 6A and 6B, respectively. Separate ANOVAs revealed that
Walker et al. (1999) found that suppression frequently origi-
nated from a single asymmetric location beyond the classical
there were only main effects of grating orientation on the Rmax
Table 1. The fraction of variance explained (FOVE) by fitting
curves to response versus contrast functions obtained
in seven stimulus conditionsa
ratio @F~1,157! ϭ 10.17, P Ͻ 0.01] and on gain measured at 25%
contrast @F~1,157! ϭ 15.78, P Ͻ 0.001] . The Rmax ratio and gain
of both direction-selective and bi-directional neurons was reduced
most by surround gratings with the same orientation as the grating
on the classical receptive field. The important point here, however,
is that there were no differences between direction-selective and
bi-directional neurons in the extent of modulation from beyond the
classical receptive field.
Best-fitting parameters
Rmax
c50
n
Simple cells (FOVE Ͼ 90%)
Complex cells (FOVE Ͼ 90%)
All cells (Mean FOVE)
88%
82%
91%
73%
52%
83%
47%
39%
69%
aThe seven curves were fit simultaneously, while each of the three param-
eters, Rmax, c50, and n [see eq. (1)], were varied and two of them were
constrained to have the same value for all seven curves. Varying Rmax while
c50 and n were held constant explained the largest fraction of variance in
both simple cells and complex cells.
Laminar distribution and retinal eccentricity
We were interested to know if there was any laminar organization
to neurons that are susceptible to modulation from beyond the

2
26
B.S. Webb et al.
Fig. 3. Effects of each surround stimulus on the Rmax
of each neuron. The best-fitting Rmax values for the
grating-alone condition are plotted against the best-
fitting Rmax values for each surround stimulus, for
each neuron. Rmax is reduced to the greatest degree in
most neurons by a large annular grating of the same
orientation as the grating on the classical receptive
field. Other surround stimuli reduced Rmax in some
neurons, but to a lesser degree.
classical receptive field. Ten neurons could not be assigned to
layers, so are not considered further here. Fig. 7 shows the Rmax
ratio for neurons recorded when an annular grating was at the same
orientation as the grating on the classical receptive field, plotted
against normalized percent depth from the cortical surface. There
does not appear to be any laminar organization to neurons that are
susceptible to reductions in the Rmax ratio. Unfortunately, there
were not sufficient neurons in each layer to perform any statistics.
We recorded data from neurons at a range of retinal eccentric-
ities (1.02–9.9 deg). To test whether there were any difference in
the strength of surround suppression at different eccentricities, we
looked for differences between foveal (1–5 deg) and peripheral
marmoset. Surround stimulation reduced the response gain of V1
neurons in a divisive fashion. Response gain varied with the
orientation of the surround stimulus and was reduced most by large
surround stimuli of the same orientation as that preferred by
neurons. The strength of surround suppression did not vary with
laminar distribution or retinal eccentricity.
Below, we compare our results with previous studies on sur-
round interactions and the predictions derived from a model of
gain control. We also discuss the anatomical substrates of gain
control from beyond the classical receptive field.
Surround interactions in V1
(
5–9.9 deg) neurons on the Rmax ratio. Even though there was a
trend towards stronger surround suppression in peripheral neurons
Ϫ0.34, 6 0.07) than in foveal neurons (Ϫ0.21, 6 0.06), a t-test
Research on surround interactions in V1 has produced contradic-
tory results. The definition of the classical receptive field, or center
field, adopted by different research groups and the different meth-
ods used to map it, are probably at the heart of many discrepancies
in the literature (Fitzpatrick, 2000; Walker et al., 2000). The
minimum response field method (Barlow et al., 1967) tends to
underestimate the size of receptive fields (Bishop & Henry, 1972;
Schiller et al., 1976). It provides a smaller estimate of receptive-
(
revealed no statistical difference.
Discussion
We measured how stimulation beyond the classical receptive field
modulated the gain of neurons in primary visual cortex (V1) of the

Surround modulation of gain in V1
227
Fig. 4. Population histograms of the Rmax ratio [eq. (3)] for each surround
stimulus condition. A large annular grating with the same orientation as the
grating on the classical receptive field was most effective in reducing the
Rmax ratio. (The distribution is markedly skewed to the left in panel C.)
Vertical dashed line represents the mean Rmax ratio.
field size than summation field measurements do with spatially
extensive stimuli, such as sinusoidal gratings (Walker et al., 2000;
Cavanaugh et al., 2002a). A stimulus that is beyond the minimum
response field and evokes no excitatory response is not necessarily
beyond the summation field of a neuron. Therefore, some reports
that stimuli outside the receptive field increase responses evoked
from inside the receptive field (e.g. Gilbert & Wiesel, 1990;
Kapadia et al., 1995) may simply be caused by surround stimuli
encroaching onto and activating relatively insensitive regions of
the summation field. When stimulated alone these regions may
evoke only a subthreshold response, but when stimulated in con-
junction with the minimum response field, they may evoke a
suprathreshold response that is greater than stimulating the mini-
mum response field alone. Here we attempted to overcome many
of these problems by mapping the summation field of V1 neurons
with spatially extensive grating stimuli. We gave particular atten-
tion to the centering of stimuli on the classical receptive field and
ensuring that surround stimuli did not encroach onto the summa-
tion field of V1 neurons (see Materials and methods).
Fig. 5. Effects of each surround stimulus on the gain of simple and
complex cells. A: Mean Rmax ratio [eq. (3)] for each surround stimulus
condition. B: Average slope of the response versus contrast function
measured over the range of contrasts from 0% to 25% (normalized by
subtracting the slope measured in the absence of a surrounding grating) for
each surround stimulus condition. Both measures of gain were reduced
most by surround gratings of the same orientation as the grating on the
classical receptive field.
evoked by an optimal stimulus within the summation field of a V1
neuron. An increase in response seems to be mediated by the
relative configuration and sign of stimuli within and beyond the
summation field. The response of a V1 neuron can be increased
when the luminance contrast (Toth et al., 1996; Levitt & Lund,
1997; Polat et al., 1998; Chen et al., 2001; Mizobe et al., 2001),
orientation (Sillito et al., 1995; Jones et al., 2002), or direction of
motion (Jones et al., 2001; Palmer & Nafziger, 2002) of stimuli
within and beyond the summation field are different. Changing the
relative luminance contrast and0or orientation of stimuli inside and
outside the receptive field can change the sign of the response
(Sillito et al., 1995; Toth et al., 1996; Levitt & Lund, 1997; Polat
et al., 1998; Chen et al., 2001; Mizobe et al., 2001; Jones et al.,
2002). At all luminance contrasts within the summation field, we
found that the activity of most V1 neurons was only ever reduced
by the presence of a surround stimulus. When stimuli inside and
outside the summation field had orthogonal orientations—a con-
figuration sometimes found to increase the response of V1 neurons
(Sillito et al., 1995; Cavanaugh et al., 2002b; Jones et al., 2002)—we
found that they were simply less effective at reducing the response
of V1 neurons than when they were at the same orientation. The
activity of a minority of neurons in our sample was increased by
The extraclassical surround in V1 is broadly tuned and less
selective than the classical receptive field (DeAngelis et al., 1994;
Li & Li, 1994; Cavanaugh et al., 2002b). Suppression of activity
in V1 neurons by stimuli in the surround is independent of spatial
phase, and is strongest when the parameters of stimuli outside the
classical receptive field match the preferences of the neuron
(
Blakemore & Tobin, 1972; DeAngelis et al., 1994; Li & Li, 1994;
Cavanaugh et al., 2002b). Consistent with this work, we found
here that responses of marmoset V1 neurons were reduced most
when the orientation of surround stimuli matched what neurons
preferred on their receptive field. Cavanaugh et al. (2002b) have
recently extended this finding and shown that surround suppres-
sion is strongest when the parameters of stimuli inside and outside
the receptive field match each other, regardless of whether they
match the preferences of the neuron.
Stimuli outside the excitatory summation field can, under cer-
tain conditions, increase the activity of V1 neurons. Here we use
the terms “increase” and “reduce” to mean relative to the response

2
28
B.S. Webb et al.
Fig. 7. Laminar distribution of surround suppression. The Rmax ratio,
for neurons recorded when an annular grating was at the same orientation
as the grating on the classical receptive field, is plotted against normalized
percent depth from the cortical surface. There appears to be no laminar
organization to neurons whose responses were reduced by surround
stimulation.
in V1 (Albrecht & Hamilton, 1982). An increase in the semisatu-
ration constant shifts the contrast–response curve rightwards, re-
ducing contrast sensitivity by a constant proportion (contrast gain
control). A reduction in the maximum obtainable response com-
presses the contrast–response curve downwards, scaling the re-
sponse at all contrasts (response gain control). An increase in the
steepness of the contrast–response curve limits the range of con-
trasts to which a neuron can increase its response.
Fig. 6. Effects of each surround stimulus on the gain of direction-selective
see eq. (2)] and bi-directional neurons. A: Mean Rmax ratio [eq. (3)] for
each surround stimulus condition. B: Average slope of the response versus
[
contrast function measured over the range of contrasts from 0% to 25%
(
normalized by subtracting the slope measured in the absence of a sur-
Heeger’s contrast normalization model (Heeger, 1991, 1992)
predicts that inhibition in V1 can be explained by a single contrast
gain-control mechanism. We (results reported here), and others
rounding grating) for each surround stimulus condition. Surround stimuli
had equivalent effects on the gain of direction-selective and bi-directional
neurons.
(Sengpiel et al., 1998; Cavanaugh et al., 2002a), have found that
a response gain model provides the best description of the effects
of surround stimulation on gain (but see, DeAngelis et al., 1994).
Other cortical phenomena, such as cross-orientation suppression
stimulation beyond the classical receptive field. But these small
response increments were rare and were not produced systemati-
cally by any particular stimulus configuration.
(
Bonds, 1989; Sengpiel et al., 1998) and contrast adaptation (Ohzawa
et al., 1985; Sclar et al., 1989), are better described by a contrast
gain model. A model that includes both response gain and contrast
gain mechanisms provides a better description of interocular sup-
pression (Sengpiel et al., 1998). Taken together, these results
suggest that surround suppression has a neural substrate that is
distinct from other cortical inhibitory phenomena. Heeger’s model
also predicts that gain control signals are pooled over a range of
orientations. Contrary to these predictions, we found that surround
suppression is described best by a response gain mechanism that is
selective for orientation.
Although we presented stimuli in apertures covering different
regions of the visual field—one that covered the summation field
and one that surrounded it—spatial summation experiments sug-
gest that center and surround mechanisms actually overlap. The
center and surround fields of V1 neurons are modelled as over-
lapping excitatory and inhibitory Gaussian sensitivity functions.
Typically, the surround mechanism is more spatially extensive and
has a subtractive influence on the center mechanism (DeAngelis
et al., 1994; Sceniak et al., 1999; Walker et al., 2000; Sceniak et al.,
2001; Levitt & Lund, 2002). One implementation of this model
revealed that the summation field of a V1 neuron is dynamic: it
Walker et al. (1999) found that surround suppression frequently
originated from a single asymmetric region beyond the classical
receptive field. The stimuli we used here were not appropriate to
test for such effects. But the lack of a correlation between the
strength of suppression originating at the ends and sides of the
receptive field that we found here suggests that surround effects in
marmoset V1 may originate from localized regions beyond the
receptive field. Indeed, we suspect that, as Walker and his col-
leagues found in cats, surround effects in marmoset V1 originate
from asymmetric, symmetric, and broadly distributed locations
beyond the classical receptive field.
Gain control
There are several different ways in which surround interactions
might modulate the contrast–response relationship of V1 neurons.
These can be represented by changes in the parameters of the
Michaelis–Menten equation, which is known to provide a good
description of the contrast–response relationship of most neurons

Surround modulation of gain in V1
229
samples more of the visual field at low contrast than at high
contrast (Sceniak et al., 1999). Our data, however, are clearly more
consistent with a model in which the surround has a divisive
influence on the center mechanism (Cavanaugh et al., 2002a;
Sceniak et al., 2001). Cavanaugh et al. (2002a) have devised a
model that implements a general form of divisive suppression.
Center and surround mechanisms are modelled as overlapping
Gaussians, but each has its own gain control. Independent gain-
control mechanisms in the center and surround enable the divisive
influence of the surround to scale responses up and down, depend-
ing on the relative strength of each mechanism. The apparent
increase in size with contrast occurs because of this scaling of
sensitivity (Cavanaugh et al., 2002a).
Conclusion
Input from beyond the classical receptive field reduced response
gain in a divisive fashion. Reductions in response gain were most
prominent when the orientation of large stimuli beyond the clas-
sical receptive field matched that preferred by neurons. Response
gain control in the early visual system may be mediated by the
interaction between cortical feedback and inhibitory influences
from beyond the classical receptive field.
Acknowledgments
This work was supported by grants from the Wellcome Trust, the BBSRC,
and a Wellcome Prize studentship to Ben S. Webb. We thank Carl Espin for
technical support, Peter Lennie for allowing us to use his software, and the
two anonymous reviewers for their comments.
Anatomical substrates for surround interactions
Surround interactions were not localized to any particular cortical
layer(s). In each layer from which we obtained data, the strength
of signals from the surround field of neurons ranged from negli-
gible to weak to very strong. As others (Sceniak et al., 2001;
Cavanaugh et al., 2002a; Levitt & Lund, 2002; Yao & Li, 2002)
have found, it seems that surround interactions are a general
property of V1, with different neurons being affected to different
extents.
What is the circuitry that mediates surround interactions in V1?
Surround interactions may be mediated by intrinsic lateral and0or
feedback connections in V1 (Rockland & Lund, 1982, 1983;
Livingstone & Hubel, 1984; Kennedy & Bullier, 1985; Gilbert &
Wiesel, 1989; Malach et al., 1993; Budd, 1998; Lamme & Roelf-
sema, 2000; Bullier, 2001). Combined anatomical and physiolog-
ical studies conducted by Angelucci and her colleagues (Angelucci
et al., 2002a,b) suggest that feedback connections from extrastri-
ate cortex are the most likely substrate. They compared the spatial
extent of feedforward input from the lateral geniculate nucleus
References
Albrecht, D.G. & Hamilton, D.B. (1982). Striate cortex of monkey and
cat: Contrast response function. Journal of Neurophysiology 48, 217–237.
Albright, T.D. & Stoner, G.R. (2002). Contextual influences on visual
processing. Annual Review of Neuroscience 25, 339–379.
Allman, J., Miezin, F. & McGuinness, E. (1985). Stimulus specific
responses from beyond the classical receptive field: Neurophysiologi-
cal mechanisms for local-global comparisons in visual neurons. Annual
Review of Neuroscience 8, 407–430.
Angelucci, A., Levitt, J.B. & Lund, J.S. (2002a). Anatomical origins of
the classical receptive field and modulatory surround field of single
neurons in macaque visual cortical area V1. Progress in Brain Research
136, 373–388.
Angelucci, A., Levitt, J.B., Walton, E.J.S., Hupé, J.M., Bullier, J. &
Lund, J.S. (2002b). Circuits for local and global signal integration in
primary visual cortex. Journal of Neuroscience 22, 8633–8646.
Barlow, H.B., Blakemore, C. & Pettigrew, J.D. (1967). The neural
mechanism of binocular depth discrimination. Journal of Physiology
193, 327–342.
Bishop, P.O. & Henry, G.H. (1972). Striate neurons: Receptive field
concepts. Investigative Ophthalmology 11, 346–354.
Blakemore, C. & Tobin, E.A. (1972). Lateral inhibition between orien-
tation detectors in the cat’s visual cortex. Experimental Brain Research
(
LGN), intrinsic lateral connections, and feedback connections
from extrastriate cortex with that of the minimum response field,
the summation field, and extraclassical surround field of V1 neu-
rons. They concluded that in V1 LGN input mediates activity in
summation fields measured with high contrast stimuli, horizontal
connections mediate activity in summation fields measured with
low contrast stimuli, and feedback connections mediate inter-
actions between stimuli within and beyond summation fields (An-
gelucci et al., 2002a,b).
Gain control appears to be a fundamental property of early
visual processing. We have found that, in both the LGN (Webb
et al., 2002) and V1 of marmosets (results reported here), response
gain is reduced in a divisive fashion by input from beyond the
classical receptive field. In the LGN, gain control may be mediated
by the feedback projection from V1 (Przybyszewski et al., 2000;
Webb et al., 2002). In V1, modulation of gain-control signals from
beyond the summation field are probably mediated by feedback
from extrastriate cortex (Angelucci et al., 2002a,b).
Cortical feedback projections may control information flowing
through the visual hierarchy by modulating the gain of neurons at
the preceding stage. In the LGN, this mechanism is mediated by an
excitatory feedback projection from V1 that influences responses
evoked from inside the classical receptive field (Przybyszewski
et al., 2000). Inhibition from outside the classical receptive field
modulates the excitatory influence of the cortex in the LGN (Webb
et al., 2002). In V1, feedback from extrastriate cortex may have a
more direct effect on inhibitory influences from outside the sum-
mation field, possibly via inhibitory interneurons.
15, 439–440.
Bonds, A.B. (1989). Role of inhibition in the specification of orientation
selectivity of cells in the cat striate cortex. Visual Neuroscience 2,
41–55.
Budd, J.M.L. (1998). Extrastriate feedback to primary visual cortex in
primates: A quantitative analysis of connectivity. Proceedings of the
Royal Society (B) (London) 265, 1037–1044.
Bullier, J. (2001). Integrated model of visual processing. Brain Research
Reviews 36, 96–107.
Carandini, M., Heeger, D.J. & Movshon, J.A. (1997). Linearity and
normalization in simple cells of the macaque primary visual cortex.
Journal of Neuroscience 17, 8621–8644.
Cavanaugh, J.R., Bair, W. & Movshon, J.A. (2002a). Nature and
interaction of signals from the receptive field center and surround in
macaque V1 neurons. Journal of Neurophysiology 88, 2530–2546.
Cavanaugh, J.R., Bair, W. & Movshon, J.A. (2002b). Selectivity and
spatial distribution of signals from the receptive field surround in
macaque V1 neurons. Journal of Neurophysiology 88, 2547–2556.
Chen, C.C., Kasamatsu, T., Polat, U. & Norcia, A.M. (2001). Contrast
response characteristics of long-range lateral interactions in cat striate
cortex. Neuroreport 12, 655–661.
Deangelis, G.C., Freeman, R.D. & Ohzawa, I. (1994). Length and
width tuning of neurons in the cat’s primary visual cortex. Journal of
Neurophysiology 71, 347–374.
Fitzpatrick, D. (2000). Seeing beyond the receptive field in primary
visual cortex. Current Opinion in Neurobiology 10, 438–443.
Geisler, W.S. & Albrecht, D.G. (1992). Cortical neurons: Isolation of
contrast gain control. Vision Research 32, 1409–1410.
Gilbert, C.D. & Wiesel, T.N. (1989). Columnar specificity of instrinsic
horizontal and corticocortical connections in cat visual cortex. Journal
of Neuroscience 9, 2432–2442.

2
30
B.S. Webb et al.
Gilbert, C.D. & Wiesel, T.N. (1990). The influence of contextual stimuli
on the orientation selectivity of cells in primary visual cortex of the cat.
Vision Research 30, 1689–1701.
Heeger, D.J. (1991). Nonlinear model of neural responses in cat visual
cortex. In Computational Models of Visual Processing, ed. Landy, M.
Nothdurft, H.C., Gallant, J.L. & Van Essen, D.C. (1999). Response
modulation by texture surround in primate area V1: Correlates of
“popout” under anesthesia. Visual Neuroscience 16, 15–34.
Ohzawa, I., Sclar, G. & Freeman, R.D. (1985). Contrast gain control in
the cat’s visual system. Journal of Neurophysiology 3, 651–667.
Palmer, L.A. & Nafziger, J.S. (2002). Effects of surround motion on
receptive-field gain and structure in area 17 of the cat. Visual Neuro-
science 19, 335–353.
Polat, U., Mizobe, K., Pettet, M.W., Kasamatsu, T. & Norcia, A.M.
(1998). Collinear stimuli regulate visual responses depending on cell’s
contrast threshold. Nature 391, 580–584.
Przybyszewski, A.W., Gaska, J.P., Foote, W. & Pollen, D.A. (2000).
Striate cortex increases contrast gain of macaque LGN neurons. Visual
Neuroscience 17, 485–494.
&
Movshon, J.A., pp. 119–133. Cambridge, Massachusetts: MIT
Press.
Heeger, D.J. (1992). Normalization of cell responses in cat striate cortex.
Visual Neuroscience 9, 181–197.
Hubel, D.H. & Wiesel, T.N. (1962). Receptive fields, binocular inter-
action and functional architecture in cat’s visual cortex. Journal of
Physiology (London) 160, 106–154.
Hubel, D.H. & Wiesel, T.N. (1968). Receptive fields and functional
architecture of monkey striate cortex. Journal of Physiology (London)
1
95, 215–243.
Rockland, K.S. & Lund, J. (1982). Widespread periodic intrinsic con-
nections in the tree shew visual cortex. Science 215, 1532–1534.
Rockland, K.S. & Lund, J. (1983). Intrinsic laminar lattice connections
in primate visual cortex. Journal of Comparative Neurology 216,
303–318.
Rossi, A.F., Desimone, R. & Ungerleider, L.G. (2001). Contextual
modulation in primary visual cortex of macaques. Journal of Neuro-
science 21, 1698–1709.
Sceniak, M.P., Ringach, D.L., Hawken, M.J. & Shapley, R. (1999).
Contrast’s effect on spatial summation by macaque V1 neurons. Nature
Neuroscience 2, 733–739.
Sceniak, M.P., Hawken, M.J. & Shapley, R. (2001). Visual spatial
characterization of macaque V1 neurons. Journal of Neurophysiology
85, 1873–1887.
Jones, H.E., Grieve, K.L., Wang, W. & Sillito, A.M. (2001). Surround
suppression in primate V1. Journal of Neurophysiology 86, 2011–
2
028.
Jones, H.E., Wang, W. & Sillito, A.M. (2002). Spatial organization and
magnitude of orientation contrast interactions in primate V1. Journal of
Neurophysiology 88, 2796–2808.
Kapadia, M.K., Ito, M. & Gilbert, C.D. (1995). Improvement in visual
sensitivity by changes in local context: Parallel studies in human
observers and V1 of alert monkeys. Neuron 15, 843–856.
Kapadia, M.K., Westheimer, G. & Gilbert, C.D. (1999). Dynamics
of spatial summation in primary visual cortex of alert monkeys.
Proceedings of the National Academy of Sciences of the U.S.A. 96,
1
2073–12078.
Kastner, S., Nothdurft, H.C. & Pigarev, I.N. (1997). Neuronal corre-
lates of pop-out in cat striate cortex. Vision Research 37, 371–376.
Kennedy, H. & Bullier, J. (1985). A double-labeling investigation of the
afferent connectivity to cortical areas V1 and V2 of the macaque
monkey. Journal of Neuroscience 5, 2815–2830.
Lamme, V.A.F. (1995). The neurophysiology of figure-ground segregation
in primary visual cortex. Journal of Neuroscience 15, 1605–1615.
Lamme, V.A.F. & Spekreijse, H. (2000). Contextual modulation in pri-
mary visual cortex and scene perception. In The New Cognitive Neuro-
sciences, ed. Gazzaniga, M., pp. 279–290. Cambridge, Massachusetts:
MIT Press. .
Lamme, V.A.F. & Roelfsema, P.R. (2000). The distinct modes of vision
offered by feedforward and recurrent processing. Trends in Neurosci-
ences 23, 571–579.
Levitt, J.B. & Lund, J.S. (1997). Contrast dependence of contextual
effects in primate visual cortex. Nature 387, 73–76.
Schiller, P.H., Finlay, B.L. & Volman, S.F. (1976). Quantative studies
of single-cell properties in monkey striate cortex. I. Spatiotemporal
organization of receptive field. Journal of Neurophysiology 39, 1288–
1319.
Sclar, G., Lennie, P. & Depriest, D.D. (1989). Contrast adaptation in
striate cortex of macaque. Vision Research 7, 747–755.
Sengpiel, F., Baddeley, R., Freeman, T.C.B., Harrad, R. & Blakemore,
C. (1998). Different mechanisms underlie three inhibitory phenomena
in cat area 17. Vision Research 38, 2067–2080.
Sillito, A.M., Grieve, K.L., Jones, H.E., Cudeiro, J. & Davis, J. (1995).
Visual cortical mechanisms detecting focal orientation discontinuities.
Nature 378, 492–496.
Skottun, B.C., De Valois, R.L., Grosof, D.H. & Movshon, J.A. (1991).
Classifying simple and complex cells on the basis of response modu-
lation. Vision Research 31, 1079–1986.
Sugita, Y. (1999). Grouping of image fragments in primary visual cortex.
Nature 401, 269–272.
Toth, L.J., Rao, S.C., Kim, D., Somers, D. & Sur, M. (1996). Sub-
threshold facilitation and suppression in primary visual cortex revealed
by intrinsic signal imaging. Proceedings of the National Academy of
Sciences of the U.S.A. 93, 9869–9874.
Truchard, A.M., Ohzawa, I. & Freeman, R.D. (2000). Contrast gain
control in the visual cortex: Monocular versus binocular mechanisms.
Journal of Neuroscience 20, 3017–3032.
Levitt, J.B. & Lund, J.S. (2002). The spatial extent over which neurons in
macaque striate cortex pool visual signals. Visual Neuroscience 19,
439–452.
Li, C.Y. & Li, W. (1994). Extensive integration field beyond the classical
receptive field of cat’s striate cortical neurons—classification and
tuning properties. Vision Research 34, 2337–2355.
Livingstone, M.S. & Hubel, D.H. (1984). Specificity of intrinsic con-
nections in primate primary visual cortex. Journal of Neuroscience 4,
2
830–2835.
Walker, G.A., Ohzawa, I. & Freeman, R.D. (1999). Asymmetric sup-
pression outside the classical receptive field of the visual cortex.
Journal of Neuroscience 19, 10536–10553.
Maffei, L. & Fiorentini, A. (1976). The unresponsive regions of visual
cortical receptive fields. Vision Research 16, 1131–1139.
Malach, R., Amir, Y., Harel, M. & Grinvald, A. (1993). Relationship
between intrinsic connections and functional architecture revealed by
optical imaging and in vivo targeted biocytin injections in primate
striate cortex. Proceedings of the National Academy of Sciences of the
U.S.A. 90, 10469–10473.
Walker, G.A., Ohzawa, I. & Freeman, R.D. (2000). Suppression outside
the classical cortical receptive field. Visual Neuroscience 17, 369–379.
Webb, B.S., Tinsley, C.J., Barraclough, N.E., Easton, A., Parker, A.
& Derrington, A.M. (2002). Feedback from V1 and inhibition from
beyond the classical receptive field modulates the responses of neurons
in the primate lateral geniculate nucleus. Visual Neuroscience 19,
583–592.
Merrill, E.G. & Ainsworth, A. (1972). Glass-coated platinum-plated
tungsten microelectrodes. Medical and Biological Engineering 10,
6
62–672.
Wörgötter, F. & Eysel, U.T. (2000). Context, state and the receptive
fields of striatal cortex cells. Trends in Neurosciences 23, 497–503.
Yao, H. & Li, C.Y. (2002). Clustered organization of neurons with similar
extra-receptive field properties in the primary visual cortex. Neuron 35,
547–553.
Mizobe, K., Polat, U., Pettet, M.W. & Kasamatsu, T. (2001). Facili-
tation and suppression of single striate-cell activity by spatially discrete
pattern stimuli presented beyond the receptive field. Visual Neurosci-
ence 18, 377–391.
Nelson, J.I. & Frost, B. (1978). Orientation selective inhibition from
beyond the classical receptive field. Brain Research 139, 359–365.
Zipser, K., Lamme, V.A.F. & Schiller, P.H. (1996). Contextual modula-
tion in primary visual cortex. Journal of Neuroscience 16, 7376–7389.