Tohoku J. Exper. Med., 1962, 78, 320-337
Spectral
Responses
of Single
Units
in the
Primate
Visual
Cortex
By
Koiti Motokawa,
Norio Taira
and Junji
Okuda
From the Department of Physiology, Tohoku University, Sendai
(Received for publication, October 24, 1962)
INTRODUCTION
Although
color vision
of the cat has been studied
electrophysiologically2,22,26)
as well
as behaviorally,6,9,11,18)
no positive
evidence
for color
vision
of this
animal has been obtained,
despite
a series of work by Granit,8)
Donner,7)
Lennox16)
and others1,15) which seemed to suggest the existence
of some color vision
mecha
nism.
On the other hand, in the fish, Motokawa
et al.,20,21) Wagner
et al.31) and
Svaetichin27)
have
obtained
strong
evidence
for color
vision
using
the
micro-electrode
technique.
However,
the use of the primate
as experimental
animal
is
more appropriate
for electrophysiological
approach
to the mechanism
of human
color vision,
because
its visual
system
is more
closely
similar
to that
of man.
Recently
de Valois et al.5) made experiments
in the lateral
geniculate
body of the
macaque
monkey
and Lennox-Buchthal17)
in the visual
cortex
of the mangabey
monkey,
and reported
that there are single neurons
showing
preferential
responses
to a narrow-band
of the
spectrum.
Hubel
and
Wiesel14)
showed
that
some
optic nerve fibers of the spider
monkey
responded
very actively
to colored lights ,
whereas
they
showed
only a weak response
to white
light.
In the present
experiment
the spectral
responses
of cortical
neurons
to local
illumination
in the macaque
monkey
were investigated
with a veiw to providing
further
information
about
the central
mechanism
for color vision
in the primate
.
METHODS
Six macaque monkeys; four crab-eating monkeys (Macaca cynomolgus irus)
and two Japanese monkeys (Macaca fuscata yakui) weighing 2-3.5 kg were used as
experimental animals. Surgical procedures and experimental apparatuses were
almost the same as those used in the previous experiment on the cat .22)
Cannulation
of the trachea
and craniotomy
were carried
out under
anesthesia
with thiamylal
sodium
which was given intravenously
through
polyethylene
tub
-本 川 弘 一, 平 則 夫, 奥 田惇 二
Spectral Responses of Single Units in Monkey's Cortex 321
ing inserted into the saphenous vein. Two trephine holes were made in the skull
overlying the left occipital cortex; the larger hole, 16mM in diameter, was made
over the medial calcarine sulcus, and the smaller one, 13mM in diameter,
just laterally to the larger one over the area which, according to Talbot and
Marshall,29)
should correspond to the retinal fovea. Each of the two holes was
fitted with a hollow Lucite peg or implant. The hole of the peg had been filled
with 6% agar gel made from the Ringer's solution. Thus, respiratory and
circulatory pulsations of the brain could be avoided, and a translucent agar
layer filling the implant allowed insertion of a microelectrode
into the cortex under
direct vision. By this method the recording from a single unit was possible for
several hours. The recording from one unit was, however, stopped after about
one hour of necessary observation to make an experiment on another unit.
The head of the animal was fixed to a head-holder which permitted stimula
tion of any part of the visual field. Before the experiment, the animal was
allowed to recover from general anesthesia, and during the experiment the
animal was immobilized with Flaxedil and maintained on artificial respiration.
Wound margins and pressure points by the head-holder were infiltrated with 1%
Xylocaine. Room temperature was 30•Ž and the animal was warmed by electric
heating pads. Thus, the rectal temperature of. the animal was maintained at
38-39•Ž.
A tungsten microelectrode was used, while a silver plate placed on the
skull served as a reference electrode. A 100-ƒÊ Ag-wire insulated except at the
tip was introduced onto the occipital cortex, and the EEG from this region was
recorded to be used as a monitor of the arousal state of the animal.
The animal was placed facing a screen of frosted glass, 50•‹in visual angle at
a distance of 60cm from the corneal surface. The corneas of both eyes were
covered with contact lenses to protect them from drying and to secure suitable
refraction. The right eye to be illuminated was kept open and atropinized to
paralyze the pupil and to relax the accommodation. The other eye was spared
for the pupil to serve as a monitor of the arousal state of the animal.
Stimulating or test colored lights were obtained by means of fifteen inter
ference filters with half-band widths less than 20 mƒÊ. The filters were adjusted to
equal energy with neutral tint filters. The light source was a 150-watts tungsten
filament lamp of a slide projector lit at about 2750•‹K. A stimulating light spot
was circular in shape, about 0.5° in visual angle (5m in diameter on the screen)
and the period of illumination was 0.5 sec. The transmission factor of the screen
was approximately 45% to spectral lights in the range between 400mƒÊ and 700mƒÊ.
The maximal intensity available of a white test spot was 30 lux on the screen.
When a unit was isolated, its receptive field was roughly searched by a white light
spot, then the spectral response behavior was surveyed with colored lights ranging
322 K. Motokawa et al.
7-8 sec with the automatic device described previously,22,26) because an interval
of 7-8 sec was needed to avoid an after-effect of the preceding test light. It took
about 2 minutes to complete a survey from the red end to the violet end of the
spectrum. To determine the spectral sensitivity curve of a unit similar measure
ment had to be carried out at several different intensity levels, so that it took
18-20minutes as a whole. Most experiments were carried out under dark
adaptation, but in some cases the effect of light adaptation was investigated.
For light adaptation a tungsten filament lamp of a slide projector was used.
Adapting colored lights were obtained with four interference filters of which the
peaks of transmission were at 506, 550, 584 and 621mƒÊ respectively. The
intensity of adapting lights was adjusted by neutral tint filters. The maximal
intensity available was 500 lux on the screen.
RESULTS
Identification of radiation and cortical units
Recordings of unit discharge were performed from a rather limited area
near the medial calcarine sulcus in the visual area; this area was found to respond
most actively to a searching spot light of 0.5•‹ in visual angle presented within
15•‹ from the center of gaze. No well-responding unit could be isolated from the
lateral surface of the visual cortex which, according to Talbot and Marshall,29)
should correspond to the retinal fovea.
Radiation and cortical units were distinguished according to the criteria used
in the previous experiments22,26,28)
in the cat; radiation units were judged by
their positive monophasic spikes (Fig. 1B) and spontaneous discharge rate
(10-Fig. 1. Records of two types of single spikes isolated in visual area of macaque monkey. A: Biphasic spike with inflection in rising limb of positive phase, considered to be specific for cell discharge . B: Monophasic spike considered to be recorded from single optic radiation fiber. Positivity upward. Time and voltage calibrations are 1 msec and 1mV respectively.
30/sec) which was definitely higher than that of cortical neurons. Cortical units
were recognized by their lower spontaneous firing rate (less than 10/sec) and their
Spectral Responses of Single Units in Monkey's Cortex 323
spike configuration
which is considered to be specific for cell discharge , that is,
positive-negative
biphasic spikes with an inflection
in the rising limb of the
positive phase (Fig. lA).
In addition, radiation units generally responded well
to even illumination
and had larger receptive fields than cortical neurons .
In
contrast, cortical units generally failed in responding to even illumination,
as was
the case in the cat.13,22) The receptive field of most cortical neurons was found to
be smaller than one degree of arc in diameter even in the dark-adapted
state.
However, some cortical neurons were found to have much larger receptive fields
than those of most radiation units and responded well to even illumination.
While advancing a microelectrode perpendicularly to the pial surface, cortical
units were first picked up and, then, radiation units were encountered deep in the
visual area. The receptive fields of these successively picked up units were
located en mass in a certain position in the visual field. On further advance
ment of the microelectrode, cortical units were again recorded probably because
of folding of the striate cortex in the monkey, but their receptive fields were now
found to be located in a quite different position in the visual field. This finding
seems to be in line with the anatomically established point to point projection of
the visual field on the occipital cortex.23)
Spectral response behavior of cortical neurons
Radiation and cortical units were distinguished
clearly on the basis of the
above-mentioned
criteria,
but they showed little
difference
in their spectral
response behavior.
De Valois and Jones4) distinguished two different varieties of
geniculate
cell responses which seemed to be related to color vision,
i.e. the
narrow-band "on" cells of the dorsal layers and the opponent "on"-or-"off"
cells
of the intermediate layers.
In the present experiment
similar types of responses
were distinguished on cortical as well as radiation units.
The first type of unit
gave always an "on" discharge to colored lights of the whole visible range of
wave-lengths as well as to a white light, and showed a definite dominant peak of re
sponse discharge rate in a certain limited range of wavelengths
(see Fig. 2). Such
units are designated tentatively
as "chromatic"
type (C type).
Spectral response curves were constructed by plotting the number of impulses
during illumination of 0.5 sec against the wavelengths of the most intense lights
available. The number of successful experiments covering the whole range of the
spectrum was 22, so limited that a statistical treatment of the data was not
possible. It was, however, found that the dominant peaks tended to center in
three selected regions of the spectrum, orange-red, green and blue (see Fig. 2); the
orange-red peaks were found in the region from 600mƒÊ to 640mƒÊ, the green peaks
between 520mƒÊ and 540mƒÊ, and the blue ones at about 460mƒÊ. Units with the
orange-red peaks were more frequently encountered than those with the green or
324 K. Motokawa et al.
Fig. 2. Spectral response curves of chromatic type "on" units. Ordina tes: Number of impulses occurring during illumination period of 0.5 sec. Vertical bar under peak of each curve represents five impulses. Open and half-filled circles refer to cortical units and filled circles to single optic radia tion unit respectively.
maximum
in the yellow region
of the spectrum
could
be found,
although
it may
be encountered
in further
experiments.
It is to be noted that the spectral response curves with a dominant peak in the
blue or orange-red region usually had a submaximum or hump in the range of
470-540mƒÊ. No such submaximum could be found in the response curve with
a maximum in the green region, but instead, the spectral band of such curves was
much broader than that of the other curves. It is likely that the green response
curve had in reality two side-bands which were fused with the dominant band
because of the low resolving power of our technique.
The limited
number
of successful
experiments
was not due
to difficulty
in
securing
good long-lasting
contact
of the microelectrode
with the cortical
neuron,
but due to difficulty
in maintaining
the experimental
animal
in a uniform
arousal
state.
As has been described
in the previous
paper,28)
the
responsiveness
of
cortical
neurons
depends
greatly
on the arousal
state
of the
animal.
The second
"opponent"
type or 0 type refers to such units which show "on"
discharges
in a certain
region
of wavelengths,
but
"off"
discharges
in another
region,
thus changing
discharge
type
depending
on the wavelengths.
Such units
responded
to the
white
light
with
either
"on",
"off"
or "on-off"
discharges,
depending
upon the
adaptation
state,
the
size of the
test
stimulus,
etc.
An
Spectral Responses of Single Units in Monkey's Cortex
325
Fig. 3. Spectral response curves of opponent type cortical unit. Open circles refer to number of impulses occurring in light period of 0.5 sec and filled circles to that of dark period of 0.5 sec immediately following cessation of illumination. Records are presented on right, in which Wh stands for white light and illumination period is marked by upward deflection of beam under each record.
example
is illustrated
in Fig. 3, in which the number
of "on"
impulses
is plotted
upward,
and that
of "off"
impulses
downward.
The response
curves
for "on"
and "off" discharges
were constructed
by the averages
of two series of experiments.
As can be seen in the figure, the peak of the "on" curve lies at 640 mƒÊ, while that
of the "off" curve
at 480 mƒÊ.
Thus there is complementary
relation
between
the
"on"
and "off" discharges
. Further
examples
are shown in Fig. 4A and B.
In
A the peak of the "on" curve is found at the blue-green
part, and that of the "off"
curve
at the orange
part
of the spectrum.
The unit represented
in Fig. 4B is of
the same type
as that in Fig. 3, but there is some minor difference
in that
the "on"
curve
of Fig. 4B seems to show an elevation
at the short wavelength
end of the
spectrum.
It is, however,
not clear whether
this elevation
is due to fluctuation
of
responses
or indicates
the existence
of another
"on"
peak.
Further
studies
on
326
K. Motokawa et al.
Fig. 4. Two varieties of spectral response curves of opponent type cortical units.
this relation
will be relevant,
because
it was assumed
by Hartridge12)
that
the
R-BG-R
unit
postulated
in his theory
should
show double
response
in the
red
and
in the
violet
parts
of the
spectrum.
The third type represents an assembly of various units which cannot be
classified into the two types mentioned above. An example is shown in Fig. 5.
The unit represented in this figure showed "on-off" discharges throughout the
whole range of the spectrum at all intensity levels. The peaks of the "on" and
"off" curves were located at 500mƒÊ
, and did not show any shift by light adapta
tion. The receptive field of this type of unit was extremely large as compared
with units of other types and responded well to even illumination. Judging
from all these properties, this unit corresponds to the Granit's scotopic dominator.
But it is to be noted that the photosensitivity of this unit was too low to be
regarded as a seotopie unit.
Another example is shown in Fig. 6. Such units showed no clear-cut
dominant peak, but responded almost uniformly to all parts of the spectrum.
It cannot, however, be decided whether such units may be comparable to the
Granit's photopic dominator, because the location of the maximum responsiveness
could not be identified with that of the photopic dominator, 560mƒÊ.
Spectral Responses of Single Units in Monkey's Cortex 327
Fig. 5.
Spectral response curve of single "on-off" cortical unit
(unit A relating to rod receptors as described in Discussion).
responses to each part of the spectrum ; their responsiveness was not so uniformly
distributed over the whole range of the spectrum as that of the unit just men
tioned, but showed two main peaks towards both ends of the spectrum. Examples
are shown in Fig. 7A and B. As can be seen in this figure, each band, "on" or
"off"
, is as narrow as that of the 0 type unit illustrated in Fig. 4, so that it seems
as if two different pairs of the 0 type units were combined. It is also possible
that this unit represents combination of two C type units, if there exist C type
units which respond with "on-off" discharges. When the intensity of the test
lights was progressively reduced, the peaks at both ends of the spectrum became
lower and lower, and a new peak emerged between the two decreasing peaks. It
is apparent that the new peak is related to scotopic vision, because it is located
at 500mƒÊ and appears at low intensities of illumination.
While engaged in this sort of experiments one of the authors imagined that
a scotopic unit might yield such a two peaked response curve at a sufficiently high
level of light intensity; the mechanism would be such that the most effective lights
around 500mƒÊ would inhibit the response while the less effective lights from the
328 K. Motokawa et al.
Fig. 6. Broad spectral response curve of "on" cortical unit.
Spectral Responses of Single Units in Monkey's Cortex 329
response peaks on either side of 500mƒÊ, This hypothesis was subjected to
experimental tests with units identified as a unit having a sensitivity maximum at
about 500mƒÊ by determination of its sensitivity curve . An example of sensitivity
curve is illustrated in Fig . 8. This unit showed a sensitivity maximum at about
500mƒÊ and its response area was much broader than that of the dominant peak
Fig. 8. Spectral sensitivity curve of single "on-off" cortical unit (unit A)
. O
rdinates: Wavelengths.
Abscissas: Relative intensities as expressed by transmis
sion factors of neutral tint filters.
Wh stands for white test light .
of any C type units. However, such a dominator never suffered inhibition on rais
ing the light intensity to a maximum level, but remained to respond most
actively to lights around 500mƒÊ. This experiment is, however
, not conclusive,
because inhibition might have occurred at still higher intensities .
Intensity-response relation of cortical neurons
All the above-mentioned
experiments
except
the last were carried
out at a
fixed level of intensity.
In the following
experiment
the intensity-response
rela
tion was studied
in each type of unit.
330 K. Motokawa et al.
In Fig. 9 the relation for a typical "on-off" unit with a sensitivity maximum
at 500mƒÊ is illustrated. The magnitude of "on" and "off" responses increased
Fig. 9. Intensity-response relation of the same cortical unit as in Fig. 8 obtained with light of 500mƒÊ as test stimulus. Ordinates: Number of impulses (open circles) occurring during light period of 0.5 sec and that (filled circles) of dark period of 0.5 sec. Abscissas: Log relative intensity. Numerals indicate n of the expression 0.7n.
with increasing
intensities;
the
curve of the impulse
number
versus
log relative
intensity
was generally
sigmoid
in shape,
and showed
no tendency
to decrease
in
the highest
intensity
range.
Fig. 10 refers to a C type unit having a dominant peak at the orange part of
the spectrum. The continuous line represents the intensity-response relation
obtained with the orange light corresponding to the dominant peak, and the
broken line represents the relation obtained with light of 500 mµ which is most
effective for scotopic vision. The relation represented by the continuous line is
of the usual form, though no saturation is seen at higher levels of intensity. The
broken curve is unusual, showing a slight maximum or saturation at a very low
intensity level. This curve is probably related to a scotopic dominator which is
linked with the orange C type unit. It is to be noted that this unit was very
much sensitive to light compared with the unit illustrated in Figs. 8 and 9 and
suffers suppression of a slight degree at higher intensities. From this and other
findings it may be said that the sensitivity of units showing a sensitivity maximum
at 500mƒÊ can differ greatly even when they were recorded from one and the
same animal under one and the same experimental conditions. The slight suppres
sion at higher intensities observed above may be due to the inhibitory action of
the photopic system upon the scotopic one, because the two systems obviously
Spectral Responses of Single Units in Monkey's Cortex 331
Fig. 10. Intensity-response relation of single chromatic type cortical unit. Open and filled circles refer to data obtained with test light of 600 mƒÊ, and those obtained with test light of 500mƒÊ respectively.
likely, because no such suppression was observed in the independent unit without
any linkage (see Fig. 9). But some reservation should be made with such a con
clusion, for the number of the independent units isolated so far is limited. An
other example of similar relation is illustrated in Fig. 11. This unit showed
Fig. 11. Intensity-response
relation of the same unit as repre
sented by open circles in Fig. 2.
332 K. Motokawa et al.
the usual sigmoid-shaped relation, when explored with the orange light cor
responding to its dominant peak, but when examined with light of 500mƒÊ the
relation consisted of two parts, one of which referred to the range of lower in
tensities and the other to that of higher intensities (see the broken curve in Fig.
11). In the lower intensity range the number of impulses increased with increasing
intensities, reached a maximum, and then decreased with further increases of
intensity. As can be seen in Fig. 12 this unit was most sensitive to the light of
500mƒÊ in the lower intensity range. Therefore it may be said that this part of
the curve represents the intensity-response relation of the scotopic system. The
sensitivity curve of this unit is illustrated in Fig. 12, in which there is one
dominant peak of sensitivity at 500 mµ and submaximum at about 620 mµ. As
Fig. 12. Spectral sensitivity curve of the same unit as shown in Fig. 11. Otherwise, the same as in Fig. 8.
can be seen in this figure, the scotopic system is more active at lower intensities
than the orange photopic one, but reverse is the case at higher intensities. The
light of 500mƒÊstimulates chiefly the scotopic system at lower intensities, but
also stimulate the photopic one at higher intensities so that the two different parts
of the intensity-response curve must have been produced. On the contrary, the
orange light which was most effective for stimulating the photopic system in this
case scarecely stimulated the scotopic system, because the relation explored with
the orange light was represented by a unique sigmoid curve (see the continuous
curve in Fig. 11). When a light of intermediate wavelength, e.g. 560mƒÊ was
used, both systems were moderately stimulated concurrently so that the respective
Spectral Responses of Single Units in Monkey's Cortex
333
of 500mƒÊ, but still a break can be observed between the two parts (see the dotted
curve in Fig. 11).
The intensity-response relation for an 0 type unit is illustrated in Fig . 13.
The "on" curve explored with red light of 640mƒÊ shows that "on" activity can
be seen only at higher intensities. Similarly
, the "off" curve determined with
blue-Fig. 13. Intensity-response relation of the same cortical unit of opponent type as shown in Fig. 3.
green light of 480 mƒÊ indicates that the "off" activity is limited to a range of
higher intensities. Thus the 0 type unit obviously represents a photopic
system.
This unit
showed
almost
no "off"
response
to the red light,
although
the
continuous
curve
marked
with
filled circles seems
to represent
"off"
responses;
in reality,
however,
these apparent
"off"
responses
indicate
in part
spontaneous
discharges
and
also the
activity
of the
scotopic
system
which
is likely to be
connected
with this unit.
The highest
activity
at a low intensity
level suggests
the latter
possibility.
The same can be said about
the curve
marked
with open
circles.
DISCUSSION
In
the
present
experiment
it was
found
to be easier
to isolate
cortical
neurons
than
optic
radiation
units
so that
the
results
presented
above
refer
334 K. Motokawa et at.
mostly
to cortical neurons.
On the contrary,
Smith et al.25) found
it very difficult
to isolate cortical
neurons responsive
to photic
stimulation
in their
experiment
on
the macaque
monkey.
This difference
may be attributable
to the mode of stimula
tion and the experimental
conditions
of the animal.
In the present
experiment
localized
illumination
was used, because it has proved
to be more suitable
for
elicit-ing responses
in cortical
neurons
than
even illumination.
Hubel
and Wiesel13)
reported
that most cortical neurons
in the cat did not respond
to even illumination.
In the
macaque
monkey
the number
of cortical
neurons
responding
to even
illumination
was also limited.
If even illumination
had been used it would not be
possible
to isolate
cortical
neurons
so easily
as in the present
experiment.
Another
factor
to make
isolation
of cortical
neurons
difficult
is general
anesthesia;
the latter
tends
to suppress
the
activity
of the
cortical
neurons
in
the monkey.
Therefore
only local anesthesia
was used in the present
experiment.
In the experiment by de Valois et al.5) the narrow-band geniculate units
responded with an "on" discharge to monochromatic as well as white lights.,
and their spectral response curves showed a dominant peak in a certain limited
region of wavelengths. These authors distinguished five different varieties
of the spectral response curves with dominant peaks at about 620, 580, 550, 510
and 450mƒÊ.
The chromatic or C type cortical units isolated in the present experiment
were all "on" units and tended to show dominant peaks of their spectral response
curves in three selected regions of the spectrum, orange-red (600-640 mƒÊ), green
(520-540mƒÊ) and blue (about 460mƒÊ) respectively.
Despite our deep interest in finding traits having a definite dominant peak
in the yellow region of the spectrum, no such unit has not been encountered yet. As
a matter of fact, the orange-red group included different subgroups with
dominant peaks at about 600, 620 and 640 mƒÊ, and the one with a response
maximum at about 600mƒÊ seemed to be hardly distinguishable from the yellow
unit by de Valois3) in view of the resolving power of the present technique.
Lennox-Buchthal17)
used five filters
with
maximum
transmissions
at
five
different
wavelength
regions
and
found
that
some
cortical
neurons
responded
selectively
to one of the
five filtered
light
stimuli,
but
that
others
showed
responses
to a few kinds
of stimuli.
The
number
of spikes
in response
to a
test flash of 20 msec was only one or two and depended
little on the intensity
of
the test light.
Our
C type
units
showed,
however,
striking
intensity
dependence,
as
stated
above, and was more or less responsive
to the whole range
of the spectrum,
when the intensity
of light was sufficiently
high.
This type of unit was character
ized by its highest
rate of discharge
to lights from a limited
spectral
region.
The
difference
in mode
of response
which
can be found
between
the experiment
of
Lennox-Buchtha117)
and the present
one is probably
due
to the
difference
in
Spectrol Responses of Ringle Units in Monkefi's Cortex
335
experimental
conditions
such
as adaptation
states
of the
retina,
the
duration
and size of illumination,
cortical
loci from
which
responses
were recorded,
etc.
The opponent
or 0 type
cortical
units
giving
an "on"
discharge
to
orange-red light and an "off"
discharge
to blue-green
one, or vice versa,
represent
the
cortical
counterpart
of similar
units recorded
in the macaque
lateral
geniculate3,4)
and in the optic nerve
fiber of the spider
monkey.14)
Retinal
ganglion
cells of
similar type were found
in the carp
by Motokawa
et al.21) and in the goldfish by
Wagner
et al.31)
At a more
distal
level
in the
fish retina
Svaetichin27)
and
Motokawa
et al.20) recorded
S-potentials
which
responded
in the
direction
of
hyperpolarization
to lights
from
a certain
region
of the
spectrum,
but in the
direction
of depolarization
to lights
of complementary
color.
Okuda
et al.22) made
similar
experiments
in the
cat's
cortex
under
com-parable
experimental
conditions,
but could
not find any units
of C or 0 type.
This negative
result seems to be relevant
for interpretation
of the results obtained
from the primate
cortex,
because
it is said that the cat is behaviorally
color-blind.
6,9,11,18) It has been behaviorally
demonstrated
by Trendelenburg
and Schmidt,30)
and by Grether10)
that the macaque
monkeys
are trichromats.
Therefore
it may
be concluded
that
the C and 0 type
units
subserve
color vision
at least
in the
macaque
monkey.
When
the
result
obtained
by Lennox-Buchthal17)
at the
cortical
level
and
that
by de Valois3) at the geniculate
level are compared,
the spectral
bands
of
cortical
units
seem to be much
narrower
than
those of geniculate
neurons.
One
may
be tempted
to conclude
from this difference
that some neuronal
interaction
would
occur
in the cortex
so as to promote
color discrimination.
The response
characteristics
of cortical
units
as studied
in the present
experiment
are more
similar
to those of geniculate
neurons
as studied
by de Valois than
to those of
cortical
units
by Lennox-Buchthal.
Therefore
one
should
be careful
not
to
draw
such
a conclusion
from
comparison
of the
two experiments
carried
out
under
so different
experimental
conditions.
To answer
the question
as to whe
ther or not the power of color discrimination
is higher towards
the cortex
it will
be
needed
to
carry
out
similar
experiments
at
different
levels
of the
visual
system
of the same animal
under
comparable
experimental
conditions.
As has been shown above, there are cortical units which respond to a
comparatively broad-band of the spectrum and whose sensitivity maximum lies
at about 500mƒÊ. They belong obviously to the rod-system. There can be
distinguished two kinds of units relating to rod receptors. They will be denoted
by A and B. The unit A has a extremely large receptive field and its discharge
type is always "on-off". The stimulus threshold of the unit A is much higher
than that of the unit B, and the discharge rate is generally low. The peak of
the response curve and also that of the sensitivity curve remain at about
336
K. Motokawa et al.
other unit or units.
The unit B is characterized by its linkage with one or more photopic units;
this feature is based on the fact that in a low intensity range a sensitivity curve
with a maximum at about 500mƒÊ is obtained, whereas in a high intensity range
there appears a response maximum at a spectral region other than 500mƒÊ (see Fig.
12). In such units the intensity-response relation consists of two parts, scotopic
and photopic (see Figs. 10 and 11). The unit B has a small receptive field and
its stimulus threshold is decidedly lower than that of the unit A.
It seems that the activity of the unit B is suppressed, when the linked photopic
unit is activated by strong illumination (see Figs. 10 and 11). On the contrary, the
unit A did not suffer suppression by raising the stimulus intensity to the maximal
level available (see Fig. 9). The unit B may be linked with any type of photopic
units, C, 0 or others such as those shown in Fig. 7. The discharge type of the
unit B is either pure "on" "on-off", or pure "off".
Finally,
the physiological
significance
of each type
of unit
will be touched
upon in the following:
The C type
units
may be considered
to act in the sense
of the Young-Helmholtz
theory,
and
the 0 type
units
in the sense
of Hering
theory.
Von Kries, Sloan24) and Motokawa and Aizawa19) propounded that there
are two kinds of rod-system, based upon experiments on total color blindness.
The working ranges of these hypothetical rod-systems are different; the one
works in the scotopic range, and the other in the mesopic. The common feature
of these systems is that their sensitivity maximum lies at about 500mƒÊ (Moto
kawa and Aizawa19))
The
characteristics
of the
two
units
A and
B isolated
in the
present
experiments
are in agreement
with those of the two rod-systems
assumed
by the
previous
authors.
SUMMARY
Spectral response behavior of single units was investigated in the visual
cortex of the unanesthetized macaque monkeys with test stimuli of 0.5•‹ in visual
angle.
1.
Cortical
and optic radiation
units
were distinguished
by the configura
tion of spikes and spontaneous
discharge
rate.
2. "On" units responding to a narrow-band of the spectrum were designated
as "chromatic" or C type units. Their response maxima were found in three
selected regions of the spectrum, orange-red (600-640mƒÊ), green (520-540mƒÊ)
and blue (460mƒÊ) respectively. Most units showed one or two submaxima besides
the dominant peak in their spectral response curves.
3.
Some units
responded
to a certain
limited
region
of the
spectrum
with
"on"
or "off"
Spectral Responses of Single Units in Monkey's Cortex 337
depending
on the wavelengths.
These units
were
designated
as "opponent"
or
0 type units.
4. Some units responded to the whole parts of the spectrum, but showed no
clear-cut response maxima anywhere in the visible parts of the spectrum.
5. Some "on-off" units and pure "off" units showed two response maxima
towards both ends of the spectrum.
6. Two kinds of units, A and B, relating to rod receptors were distinguished.
The common feature of these units was that the sensitivity maximum was found
at about 500mƒÊ. The unit A had a large receptive field and gave "on-off"
discharges. Its photosensitivity is low compared with that of B. The unit B
was always linked with photopic units and worked in a range of low intensities
in which the linked photopic unit was almost inactive. Its receptive field was
very small as compared with that of A.
References
1) Chang, H.-T., Res. Publ. Ass. nerv. ment. Dis., 1952, 30, 430.
2)
Cohn, R., J. Neurophysiol.,
1956, 19, 416.
3)
De Valois, R. L., J. gen. Physiol., 1960, 43, pt. 2, 115.
4)
De Valois, R. L. & Jones, A. E., in the Visual System (Jung, R. & Kornhuber,
H.,
eds.). 1961, Berlin-Gottingen-Heidelberg,
Springer, p. 178.
5)
De Valois, R. L., Smith, C. J. & Kitai, S. T., J. comp. physiol. Psychol., 1959, 52, 635.
6)
DeVoss, J. C. & Ganson, R., J. Anim. Behav., 1915, 5, 115.
7) Donner, K. 0., Acts, physiol. scand., 1950, 21, Suppl., 72.
8) Granit, R., Sensory Mechanism of the Retina, 1947, London,
Oxford University
Press.
9)
Gregg, F. M., Jamison, E., Wilkie, R. & Radinsky, T., J. comp. Psychol., 1929, 9, 379.
10)
Grether, W. F., Comp. psychol. Monogr., 1913, 15, No. 4, 38.
11)
Gunter, R., J. comp. physiol. Psychol., 1954, 47, 169.
12) Hartridge,
H., Recent Advances in the Physiology of Vision, 1950, London, Churchill.
13) Hubel, D. H. & Wiesel, T. N., J. Physiol., 1959, 147, 226.
14) Hubel, D. H. & Wiesel, T. N., ibid., 1960, 154, 572.
15) Ingvar, D. H., Acta physiol. scand., 1959, 46, Suppi., 159.
16)
Lennox, M. A., J. Neurophysiol.,
1958, 21, 70.
17) Lennox-Buchthal,
M. A., Vision Res., 1962, 2, 1.
18) Meyer, D. R., Miles, R. C. & Ratooth,
P., J. Neurophysiol.,
1954, 17, 289.
19) Motokawa, K. & Aizawa, T., Tohoku J. Exp. Med., 1957, 66, 187.
20)
Motokawa, K., Oikawa, T. & Tasaki, K., J. Neurophysiol.,
1957, 20, 186.
21) Motokawa, K., Yamashita, E. & Ogawa, T., Tohoku J. Exp. Med., 1960, 71, 261.
22)
Okuda, J., Taira, N. & Motokawa, K., ibid., 1962, 78, 147.
23)
Polylak, S., The Vertebrate Visual System, 1957, University Chicago Press.
24)
Sloan, L. L., J. opt. Soc. Amer., 1954, 44. 117.
25)
Smith, C. J., De Valois, R. L. & Kitai, S. T., Amer. Psychol., 1958, 13, 387.
26)
Suzuki, H., Taira, N. & Motokawa, K., Tohoku J. Exp. Med., 1960, 71, 401.
27)
Svaetichin,
G., Acta physiol. scand., 1956, 39, Suppl., 134, 17.
28)
Taira, N. & Okuda, J., Tohoku J. Exp. Med., 1962, 78, 76.
29)
Talbot, S. A. & Marshall, W. H., Amer. J. Ophthal., 1941, 24, 1255.
30)
Trendelenburg,
W. & Schmidt, I., Z. vergl. Physiol., 1930, 12, 249.
31)
Wagner, H. G., MacNichol, E. F. Jr. & Walbarsht, M. L., J. gen. Physiol.,
1960,
43, pt. 2, 45.
