Download Response characteristics of neurons in the pulvinar of awake cats to

Exp Brain Res (2000) 133:209–218
Digital Object Identifier (DOI) 10.1007/s002210000374
R E S E A R C H A RT I C L E
Sabine Sudkamp · Matthias Schmidt
Response characteristics of neurons in the pulvinar of awake cats
to saccades and to visual stimulation
Received: 6 September 1999 / Accepted: 10 February 2000 / Published online: 31 March 2000
© Springer-Verlag 2000
Abstract We studied responses of pulvinar neurons in
awake cats that were allowed to execute spontaneous eye
movements. Extracellular cell activity during saccades,
saccade-like image shifts, and various stationary visual
stimuli was recorded together with the animals’ eye positions. All neurons analyzed had receptive fields that covered most of the central 80×80° of the animals’ visual
field and did only respond to large (>20°) visual stimuli.
According to their response properties, recorded neurons
were divided into three populations. The first group,
termed “S neurons” (16%), responded when the animals
performed saccades but were unresponsive to any of the
visual stimuli tested. These neurons do not seem to receive a visual input that is strong enough to drive them.
The second group, termed “V neurons” (51%), responded to various visual stimuli including saccade-like image
motion when the eyes were stationary, but not when the
animals executed saccades. V neurons therefore distinguish retinal image movements that are generated externally from internally generated image motion. Finally,
“SV neurons” (31%) responded when the animals made
saccades as well as to saccade-like image motion or to
stationary stimuli. Although these neurons do not distinguish self-induced retinal image motion from motion
generated by external stimulus movements, they must receive non-retinal motion-related input, because responses elicited by saccades had shorter latencies than responses to saccade-like stimulus movements. Only SV
neurons resemble response properties of pretectal neurons that project to the pulvinar and that comprise the
major subcortical visual input. The functional significance of pulvinar neuronal populations for visual and
visuomotor information processing is discussed.
S. Sudkamp · M. Schmidt (✉)
Allgemeine Zoologie and Neurobiologie,
Ruhr-Universität Bochum, 44780 Bochum, Germany
e-mail: [email protected]
Tel.: +49-234-3227709, Fax: +49-234-3214278
Key words Extrageniculate visual thalamus · Pretectal
nuclear complex · Saccadic eye movements · Visuomotor
integration
Introduction
In higher mammals visual information from the retina
can reach the visual cortex either directly, via the dorsal
lateral geniculate nucleus (LGNd), or by a more indirect
pathway, via the superior colliculus and pretectal nuclear
complex to the lateral posterior-pulvinar complex (LP-P)
which in turn is reciprocally connected with many visual
cortical areas. While the functional role of the geniculostriate pathway has been elucidated by many investigators, knowledge about the function of the extrageniculostriate pathway is far from complete. In cats, the LP-P
consists of three different subnuclei. Each of them is
characterized by specific afferents from visual structures
that are almost restricted to it. Tectal afferents terminate
mainly in the medial division of the lateral posterior thalamic nucleus (LPm), the lateral division of LP (LPl) is
reciprocally connected with area 17 and 18, and a pretectal projection mainly ends in the pulvinar (Pul; Graybiel
and Berson 1980; Abramson and Chalupa 1988). In addition to these specific afferents, all three subnuclei
share multiple cortical areas as common input sources, in
particular various extrastriate visual areas (Updyke 1977,
1981; Berson and Graybiel 1978; Graybiel and Berson
1980; Abramson and Chalupa 1985).
Physiological studies of visual responses of LP-P
neurons in cat, monkey, and rabbit have almost exclusively focused on the integration of tectal and cortical information. Thus, a number of studies have addressed response properties of neurons in cat and rabbit LPm and
LPl, and in tecto-recipient portions of the Pul in monkey
(cat: Fish and Chalupa 1979; Berson and Graybiel 1983;
Abramson and Chalupa 1985, 1988; Chalupa and
Abramson 1988, 1989; Chalupa 1991; Casanova et al.
1997; Merabet et al. 1998; behaving monkey: Robinson
and McClurkin 1989; Robinson et al. 1991; rabbit:
210
Casanova and Molotchnikoff 1990). In contrast, physiological data from neurons in the pretecto-recipient Pul is
rather limited (cat: Godfraind et al. 1972; Mason 1981).
Receptive fields in the LPl and LPm mostly have welldefined borders and LP neurons often respond, directionally or orientationally selective, to motion of small visual
stimuli. In contrast, receptive fields in the Pul are often
very large, do not have clear borders, and Pul neurons respond to diffuse illumination changes rather than to
small visual stimuli (Godfraind et al. 1972; Mason 1981;
Chalupa and Abramson 1988, 1989; Casanova et al.
1989).
As noted above, the primary subcortical visual input
to the Pul originates from the pretectal nuclear complex.
Following the nomenclature introduced by Guillery et al.
(1980), we regard the narrow band of retinal terminations shown by several authors (see, for example, Itoh
et al. 1979; Leventhal et al. 1980) at the caudoventral
border of the Pul as part of the medial interlaminar nucleus and not as part of the Pul. Pretectal neurons that
project to the Pul are distributed throughout the whole
superficial pretectum (Weber et al. 1986; Kubota et al.
1988). Studies in anesthetized cats have demonstrated
that these neurons have large receptive fields (diameter
>40°) and that the majority of them respond with short
bursts to rapid displacements of a large random square
pattern and to brief light flashes (Sudkamp and Schmidt
1995). They are thought to be identical with so-called
“jerk” neurons that respond, as known from awake cats,
with short bursts not only to rapid image shifts while the
eyes are stationary but also when saccadic eye movements are executed in the light (Schweigart and Hoffmann 1992). It is not known yet how this pretectal input
is further processed in the Pul because response properties of Pul neurons in awake cats so far have not been
described. However, Pul neurons that respond after saccadic eye movements have been described in encéphale
isolé cats (Infante and Leiva 1984).
In primates, one function discussed for Pul neurons
is the spatial control of visual attention (Ungerleider
and Christensen 1979; Desimone et al. 1990) possibly
by enhancing responses to new visual stimuli that appear in a complex environment (Petersen et al. 1985;
Robinson et al. 1986; Desimone et al. 1990; Desimone
and Duncan 1995). The signal necessary for this enhancement could be generated by a saccade-related activation of Pul neurons that are involved in the processing of oculomotor information rather than in object
analysis tasks.
The analysis of saccade-related responses of single
neurons therefore is a first step in order to elucidate the
possible role of the cat Pul in visual and visuomotor processing. Here we report on response properties of single
units to visual stimuli and during saccadic eye movements in awake cats.
Materials and methods
Animals
Experiments were conducted on six young adult male and female
cats with body weights between 2.7 and 4.7 kg. All experimental
procedures were done in strict compliance with governmental regulations and in accordance with the Guidelines for the Use of Animals in Neuroscience Research of the Society for Neuroscience.
Animal preparation
For initial surgery, cats were anesthetized with ketamine (20 mg/kg
body weight) and thiazine hydrochloride (1 mg/kg body weight).
Under additional local anesthesia by lidocaine hydrochloride, the
trachea was intubated to allow artificial respiration with a mixture
of 66% nitrous oxide, 32% oxygen, and 2% carbon dioxide, supplemented by 0.4–1% halothane. Additional doses of ketamine
were administered if necessary. Respiration volume was regulated
to an end tidal volume of 4% carbon dioxide. The body temperature was automatically maintained at 37.5°C. Carbon dioxide level
in expired air, endotracheal pressure, and body temperature were
continuously monitored during surgery. Throughout anesthesia, animals were given an intravenous infusion of 5% glucose in 0.9%
NaCl. A craniotomy was performed above the Pul at HorseleyClarke stereotaxic coordinates taken from Reinoso-Suarez (1961).
A recording cylinder centered at A6.5/L6 and three anchor screws
were fixed to the scull with dental acrylic cement. In order to measure eye movements, a scleral search coil was implanted in one eye
according to the method of Judge et al. (1980). During the 1st week
after surgery, animals were given 60 mg chloramphenicol per day
to prevent infections.
Recording
Daily recording sessions started 7 days after surgery and typically
lasted for 1–2 h. Cats were placed in a wooden box to limit body
movements and their head was restrained by a head holder fixed to
the box. Horizontal and vertical eye movements were monitored
using the phase-detection principle in a magnetic field (Kasper
and Hess 1991). Eye movements were measured relative to the
stationary head. Extracellular single unit activity in the LP-P was
recorded with tungsten-in-glass electrodes fixed to a holder that
was mounted on top of the recording cylinder. A micromanipulator attached to the electrode holder allowed electrode penetrations
in a 10×10-mm area around the center of the recording cylinder.
The electrodes were advanced through the intact dura by a hydraulic microdrive. Signals were conventionally amplified, band-pass
filtered, and digitally stored for off-line analysis.
Extracellular responses were collected either while the cat
executed spontaneous saccades in front of a stationary random
square pattern and in total darkness, or during visual stimulation
when the eye was stationary. Visual stimuli were projected onto a
tangent screen 30 cm in front of the cat that covered the central
80×80° of the visual field. The most common visual stimulus was
a random square pattern (pixel size 2.5°, luminance of light squares 20 cd/m2, luminance of dark squares 3 cd/m2) that was either
moved in a saccade-like fashion to elicit saccade-like retinal image shifts or that was flashed on and off at 0.5 Hz. Because the
size of the stimulus exceeded that of the tangent screen, no moving edges occurred when the stimulus was moved. Small stimuli,
such as single light or dark spots or bars were also tested, however, such stimuli were not effective in eliciting neuronal responses
(see Results).
At the end of the final recording session of each individual animal, the recording sites were marked by electrolytic microlesions
in three penetrations where neurons had been successfully recorded in prior session. Recording sites of all recorded neurons were
reconstructed from the lesions in 50-µm-thick coronal sections
stained with cresyl violet (Fig. 1).
211
Fig. 1 Photomicrograph of a
cresyl violet-stained frontal
section through the left cat
thalamus. Two electrolytic
microlesions (black arrows)
mark one of the electrode penetrations in the pulvinar (Pul);
another electrode track is seen
more lateral (white arrows).
LGNd Dorsal lateral geniculate
nucleus, LPl lateral posterior
nucleus, lateral part, LPm lateral
posterior nucleus, medial part.
Scale bar 2 mm
Data analysis
For saccade detection, eye velocity was calculated off-line from
the eye position data. A saccade was detected when the eye velocity exceeded 50°/s and its amplitude exceeded 2°. Saccade onset
and saccade end were defined as the time when velocity reached
10% or dropped below 10% of the peak velocity during the saccade. Average saccade velocity, duration, amplitude, and direction
were calculated from the horizontal and vertical eye position values between saccade onset and saccade end.
Neuronal responses were analyzed on the basis of peristimulus
time histograms triggered to either saccade or stimulus onset. A
neuronal response was regarded as significant if cell activity was
significantly different (P<0.05) from average spontaneous activity
in a time window of at least 200 ms prior to saccade or stimulus
onset. If responses to at least 25 stimulus sweeps or saccades had
been collected, a t-test was used to determine significance; for less
than 25 sweeps a Mann-Whitney rank sum test was used. Responses were characterized by the latency to half of the peak activity, response duration, and average activity.
Results
In the LP-P of six awake cats, 169 units could be characterized by their responses to visual stimuli or to saccades. Reconstruction of electrolytic lesions showed that
the great majority of them were located in the Pul
(n=125). The remaining 44 neurons were located in adjacent nuclei [LPl, LPm, and lateral intermediate nucleus,
caudal part].
Because the animals executed voluntary eye movements, saccades to which responses were analyzed covered all possible directions and varied considerably in
size (2.0–48.7°, mean 8.0±6.0°, n=1499). To allow comparison of saccade-evoked with stimulus-evoked neuronal responses, we used a standard 8° horizontal stimulus
shift with a duration of 100 ms, referred to as ‘saccadelike stimulus’, to achieve an image shift on the retina
that was similar to an average saccade.
Recorded neurons could generally be divided into
three response groups: first, neurons that responded dur-
ing saccades but not to visual stimulation (n=27, 16% of
characterized neurons; S neurons), second, neurons that
responded to visual stimuli but not during saccades
(n=86, 51% of characterized neurons; V neurons), and
third, neurons that responded during saccades and to visual stimuli (n=53, 31% of characterized neurons; SV
neurons). Three neurons (2%) remained unclassified because no saccades occurred when they were recorded.
Neurons of either group were distributed throughout the
whole Pul. While no significant difference was found in
the distribution of V and SV neurons, S neurons tended
to be more frequent in the rostral half of the Pul. Neurons recorded outside the Pul could belong to either response group (Fig. 2).
S neurons
The response of a typical S neuron is shown in Fig. 3.
This neuron was activated whenever the animal executed
a saccade in light and in total darkness but it did not respond to any of the visual stimuli used. The response
characteristics to saccades in the light of 22 S neurons
located inside the Pul are summarized in Table 1. Five
of the 8 neurons that were also tested for saccade responses in total darkness showed saccadic responses that
were not significantly different from responses in light
(Table 1). Responses of S neurons found outside the Pul
were not significantly different from those located inside
the Pul (Mann-Whitney rank sum test, P>0.05). Thus,
S neurons both inside and outside the Pul most likely receive saccade-related input of either motor, premotor, or
proprioceptive origin. If they do also receive visual input, that input is not strong enough to drive them. In all
cases tested, the S neurons’ saccade response latency,
duration, or amplitude did neither correlate to saccade
direction (n=4), nor to saccade amplitude, duration, or
eye velocity (n=12).
212
Fig. 2 Distributions of S neurons (A), V neurons (B), and SV neurons (C) in cat thalamus. Two series of reconstructed frontal sections, arranged from caudal to rostral (bottom left to top right) are
shown for each population. For better identification, the Pul is indicated by the gray shading. Each black dot represents a single recorded neuron. LGNd Dorsal lateral geniculate nucleus, LD lateral
dorsal nucleus, LIc lateral intermediate nucleus, caudal part, LPl
lateral posterior nucleus, lateral part, LPm lateral posterior nucleus,
medial part, MGB medial geniculate body, SN substantia nigra
Fig. 3 Responses of a representative S neuron to saccades in light
(A), saccades in darkness (B), saccade-like stimulus shifts (C),
light on (D), and light off (E) shown as peristimulus time histograms. Eye position traces above histograms in A and B were calculated from recorded horizontal and vertical eye positions, stimulus position is indicated above histogram in C, and stimulus luminance is indicated above histograms in D and E. Saccade or stimulus onset is always at time ‘0’. The calibration bar at eye position
traces in A and B represents 20°, binwidth of PSTHs, 5 ms
V neurons
The response of a typical V neuron is shown in Fig. 4.
This neuron was not activated by saccades but it responded to saccade-like image shifts with a response latency of 40 ms, and to On-Off stimuli with a latency of
35 ms. Responses to saccade-like image shifts of large
visual stimuli or to whole-field On-Off stimuli were a
common feature of all V neurons; small stimuli such as
dots or bars had no effect. However, only 19 of 64 V
neurons (30%) tested responded to both saccade-like image shifts and On-Off stimuli, 32 V neurons (50%) responded only to On-Off, and 13 V neurons (20%) responded only to saccade-like image shifts. Twenty-three
V neurons that responded to saccade-like image shifts
were not tested with On-Off stimuli. We did not further
differentiate between these subgroups of V neurons be-
213
Table 1 Response properties of pulvinar (Pul) neurons to saccades and to visual stimuli. (n.r. No response, by definition)
Responses to:
S neurons
V neurons
SV neurons
Inside Pul
Outside Pul
Range
Median
Range
Median
Range
Median
Range
Saccades in light
Latency (ms)
Duration (ms)
Amplitude (spk/s)
0–140
35–240
25–142
63
105
56
n.r.
n.r.
n.r.
n.r.
n.r.
n.r.
n.r.
n.r.
n.r.
n.r.
n.r.
n.r.
0–110
15–320
15–313
55
95
74
Saccades in darkness
Latency (ms)
Duration (ms)
Amplitude (spk/s)
50–195
25–170
38–151
75
105
66
n.r.
n.r.
n.r.
n.r.
n.r.
n.r.
n.r.
n.r.
n.r.
n.r.
n.r.
n.r.
50–145
60–400
44–200
100
95
84
Saccade-like stimulus
Latency (ms)
Duration (ms)
Amplitude (spk/s)
n.r.
n.r.
n.r.
n.r.
n.r.
n.r.
30–190
15–270
30–246
75
55
60
35–80
10–100
18–226
50
45
99
30–200
15–185
24–280
45
60
81
On stimulus
Latency (ms)
Duration (ms)
Amplitude (spk/s)
n.r.
n.r.
n.r.
n.r.
n.r.
n.r.
20–115
10–150
19–314
48
35
86
20–75
10–100
28–320
40
25
182
25–90
20–100
22–222
43
50
114
Off stimulus
Latency (ms)
Duration (ms)
Amplitude (spk/s)
n.r.
n.r.
n.r.
n.r.
n.r.
n.r.
25–120
15–200
11–198
58
35
54
20–100
10–70
22–245
30
30
162
25–80
10–170
51–215
40
43
105
cause no differences in their response characteristics
were encountered (Table 1).
However, a significant difference between V neurons
inside and outside the Pul was noticed. Response latencies
of V neurons outside the Pul to saccade-like stimulus shifts
(35–80 ms, median 50 ms), to On (20–75 ms, median
40 ms), and to Off stimuli (20–100 ms, median 30 ms)
were significantly shorter than for neurons inside the Pul
(Mann-Whitney rank sum test, P<0.05). Also, response activities to saccade-like stimulus shifts (18–226 spk/s, median 99 spk/s), to On (28–320 spk/s, median 182 spk/s), and
to Off stimuli (22–245 spk/s, median 162 spk/s) of V neurons outside the Pul were significantly higher than activities of neurons inside the Pul (Mann-Whitney rank sum
test, P<0.05).
SV neurons
The response of a typical SV neuron is shown in Fig. 5.
The neuron responded to saccades in light with a latency of 30 ms, but it did not respond to saccades in total
darkness. Saccade-like image shifts elicited a response
with a latency of 40 ms and On-Off stimulus responses
had latencies of 35 ms and 40 ms, respectively. Responses to saccades and large visual stimuli, such as
saccade-like image shifts or whole-field On-Off stimuli,
were a common feature of all SV neurons. Small stimuli
such as dots or bars were not effective in driving SV
cells. Most SV neurons (n=53, 51%) responded to saccades and to saccade-like image shifts as well as to OnOff stimuli, 5 SV neurons (10%) responded only to sac-
Median
cades and saccade-like image shifts, and 12 SV neurons
(23%) responded only to saccades and On-Off stimuli.
The remaining 9 neurons (16%) responded during saccades and saccade-like image shifts but were not tested
to On-Off. We did not further differentiate between
these subgroups, because no differences in their response characteristics were encountered. There was also
no significant difference between SV neurons inside the
Pul and outside the Pul (Mann-Whitney rank sum test,
P>0.05).
Response parameters of SV neurons to saccades are
summarized in Table 1. In 28 neurons tested, no significant correlations were found between response parameters and saccade amplitude, duration, and velocity. Furthermore, similar to what we found for S neurons, saccade responses of SV neurons (n=14) were independent
from saccade direction. Five out of 26 SV neurons tested
exhibited significant responses to saccades in darkness.
While response latencies to saccades in darkness were
significantly longer than those to saccades in light, both
response amplitudes and response durations were similar
to those observed in light (Table 1). Thirty-nine SV neurons (30 inside Pul, 9 outside) responded to saccade-like
stimulus displacements. Thirty-five SV neurons responded to On (30 in Pul, 5 outside) and 32 SV neurons responded to Off stimuli (27 in Pul, 5 outside).
Saccade responses vs visual responses
From the response properties of SV neurons the question
arises, whether the saccade-related responses observed
214
Fig. 4 Responses of a representative V neuron to saccades in light
(A), saccade-like stimulus shifts (B), light on (C), and light off
(D) shown as peristimulus time histograms. Eye position traces
above histograms in A were calculated from recorded horizontal
and vertical eye positions, stimulus position is indicated above
histogram in B, and stimulus luminance is indicated above histograms in C and D. Saccade or stimulus onset is always at time ‘0’.
The calibration bar at eye position traces in A represents 30°, binwidth of PSTHs, 5 ms
are entirely elicited by the retinal image shift that occurs
during saccades, or whether an additional non-visual
component is present. A comparison of the parameters of
responses to saccades with responses to saccade-like image displacements (Table 1) revealed no significant difference (Mann-Whitney rank sum test). Because the animals executed spontaneous saccades of variable amplitudes and durations, responses to saccades and responses
to saccade-like image displacements were normalized to
constant duration and accumulated (Fig. 6) to allow a
better comparison. In the histograms shown in Fig. 6,
Fig. 5 Responses of a representative SV neuron to saccades in
light (A), saccades in darkness (B), saccade-like stimulus shifts
(C), light on (D), and light off (E) shown as peristimulus time histograms. Eye position traces above histograms in A and B were
calculated from recorded horizontal and vertical eye positions,
stimulus position is indicated above histogram in C, and stimulus
luminance is indicated above histograms in D and E. Saccade or
stimulus onset is always at time ‘0’. The calibration bar at eye position traces in A and B represents 30°, binwidth of PSTHs, 5 ms
saccade or stimulus onset is set at bin 20, saccade or
stimulus end is set at bin 40. The bottom histogram
shows the difference between the cumulative normalized
histograms of responses to saccade and to image shifts.
Because normalized saccade responses start at saccade
onset while responses to image displacement begin at bin
215
significant (Mann-Whitney rank sum test, P<0.01).
These differences indicate that saccade responses contain
a non-visual component in addition to the visual component.
Discussion
Fig. 6 Population response of SV neurons to saccades (A) and to
saccade-like stimulus shifts (B) normalized to equal durations.
Saccade or stimulus onset is always at bin 20, saccade and stimulus end is always at bin 40 (both marked by stars), regardless of
individual saccade duration. Eye position traces in A were calculated from recorded horizontal and vertical eye positions and normalized to equal durations, and stimulus position is indicated
above histogram in B. The histogram in C shows the difference
between A and B. The calibration bar at eye position traces in A
represents 50°
30, an initial positive peak appears in the difference histogram. Later, the response to the visual stimulus is higher in amplitude, so the difference becomes negative. Furthermore, saccade responses end soon after saccade offset at bin 45, whereas responses to saccade-like image
shifts continue until bin 53. The observed differences between bin 20 and 30 and between bin 45 and 53 were
Neurons in the Pul of awake cats showed three different
activation patterns in response to spontaneous saccades
or to moving and stationary visual stimuli. A first group
was only activated when the animal performed a saccadic eye movement, and this group of neurons was termed
S neurons. A second group, V neurons, was activated by
visual stimulation but did not respond to eye movements.
A third group responded to both saccadic eye movements and visual stimulation and was termed SV neurons. A common feature of all Pul neurons that were responsive to visual stimulation, however, was the ineffectiveness of small stimuli (diameter <5°), such as light or
dark spots or bars, to elicit any response.
In earlier studies of the cat LP-P, both neurons with
well-defined visual receptive fields that respond to relatively small stimuli and neurons with large receptive
fields without clearly defined borders that respond to diffuse illumination have been described (Godfraind et al.
1972; Mason 1981; Chalupa and Abramson 1988, 1989;
Casanova et al. 1989). However, neurons of the two
groups are not distributed homogeneously throughout the
LP-P. Instead, neurons with well-defined receptive field
boundaries are predominantly found in the LP, while
neurons with diffuse receptive fields are most prominent
in the Pul (Godfraind et al. 1972; Mason 1981). Our results confirm these findings insofar as we recorded neurons with diffuse receptive fields that respond to very
large stimuli inside the Pul. However, in contrast to both
Godfraind et al. (1972) and Mason (1981), we did not
find any neuron with clearly defined receptive field borders inside the Pul that responded to small bars or spots.
It cannot be ruled out that this discrepancy is due to
methodological differences. We tested visual responses
in the Pul of awake cats that had not been trained to fixate. Under this condition it is impossible to test the entire
visual field with small stimuli because cats make saccades to such stimuli and we therefore may have missed
neurons that respond to small visual stimuli.
Comparison of SV with S and V neurons
Saccade responses were a common feature of S and SV
neurons. For both cell classes no saccade parameters
were encoded in saccade-related responses. Response latencies, response durations, or average response amplitudes were not significantly different between S and SV
neurons. However, saccade responses of S neurons
did not have a visually evoked component whereas
in SV neurons the visual response component was always stronger than the non-visual component. Therefore,
216
S neurons and SV neurons most likely comprise neuronal populations that are independent from each other and
that are also different functionally.
V neurons and SV neurons may seem more similar
because both neuronal populations responded to identical visual stimuli. However, in addition to the distinction in their responses to saccades, there are two further
significant differences between V neurons inside the Pul
and SV neurons. Firstly, response latencies of SV neurons to saccade-like image shifts were significantly
shorter than those of V neurons. Secondly, average response amplitudes of SV neurons to On-Off stimuli are
significantly larger than in V neurons (Mann-Whitney
rank sum test, P<0.05). Therefore, it seems likely that
SV and V neurons are also separate neuronal populations in cat Pul.
Responses to saccades
Saccade-related responses of cat LP-P neurons have
been described in encéphale isolé cats that had their eyes
covered by dark contact lenses (Infante and Leiva 1984).
Under this condition, response latencies varied between
50 and 250 ms after saccade onset (mean 100 ms). We
found similar response latencies for S and SV neurons to
saccade onset in darkness. However, we observed shorter
response durations and higher average response rates
for both SV and S neurons (Table 1) than those described
by Infante and Leiva (1984), i.e., 200–500 ms and
8–87 spk/s, respectively. Even more significant, in contrast to Infante and Leiva (1984) who reported saccade
responses in LP-P to depend on saccade direction in
their experimental condition, we did not detect any directional selectivity of saccade responses. It seems likely
that these differences result from the different types of
preparations.
Comparison with monkey data
We have grouped neurons in cat Pul into S, V, and SV
responses according to their specific responses to saccades and to various visual stimuli. Similar response
types have been described in monkey Pul nucleus: approximately half of the neurons that show excitatory responses during saccades also respond to visual stimuli
and thus correspond to SV neurons. The remaining half
of neurons does not show visual responses and can thus
be classified as S neurons (Robinson et al. 1991). Some
but not all S neurons as well as some but not all SV neurons in monkey Pul do also respond to saccades in total
darkness, which is another similarity to our S and SV
neurons. Finally, Robinson et al. (1991) found neurons
that respond to bars that are moved when the eye is stationary, but that do not respond to the same retinal image
movement when a saccade is made across a stationary
bar. Such neurons may correspond to our V neurons in
cat Pul.
Comparison with pretectal neurons that project
to the Pul
Because the main visual input to the Pul in cat originates
from the ipsilateral and, to lesser extent, contralateral
pretectal nuclear complex, the question arises whether
response properties of Pul neurons reflect this input. Pretectal neurons that project to the LP-P have large receptive fields and, in anesthetized cats, respond to On-Off
stimuli and during saccade-like shifts of large patterns
(Sudkamp and Schmidt 1995). These neurons show
mean response latencies to saccade-like image shifts of
57 ms, mean response durations of 47 ms, and mean response rates of 169 spk/s. In behaving cats, pretectal
cells have been described, that show similar response
properties to visual stimuli and that display remarkable
saccade-related responses (Schweigart and Hoffmann
1992). The similarities between the two neuronal groups
recorded in anesthetized and awake cats have led to the
conclusion that they are identical (Sudkamp and Schmidt
1995). Thus, as pretectal neurons that project to the Pul
respond to both visual stimulation and to saccades, it
seems reasonable to assume that the response properties
of SV neurons in the Pul are generated by pretectal input. Since the mean conduction time of electrically
evoked spikes from the pretectum to LP-P was 1.7 ms
(Sudkamp and Schmidt 1995), the response latencies of
SV neurons are compatible to a direct pretectal input, as
are response durations and response amplitudes.
If the pretectal input to the Pul provides the predominant visual input to SV cells, the underlying projection
must be excitatory in nature. This is reasonable to assume, even though pretectal cells that project to the
LGNd have been shown to be GABAergic (Cucchiaro
et al. 1991; Wahle et al. 1994), because pretectal neurons
that project to the Pul are independent from those that
project to the LGNd (Sudkamp and Schmidt 1995). Interestingly, both projections, though anatomically independent, seem to share the functional property of transferring saccade-related information to their target structures (Schmidt 1996; Fischer et al. 1998). All the similarities in response properties between pretectal and SV
Pul cells, of course, do not rule out the possibility that
cortical inputs also participate the generation of SV Pul
cell response properties.
In contrast to SV neurons, the response properties of
S and V neurons are more specific in that they distinguish between retinal image motion generated by saccades and motion in the external world. It seems unreasonable that such response properties result from pretectal input, because pretectal neurons do not show this differentiation in their responses between externally and internally generated image motion. More likely sources for
the generation of such elaborate response properties are
cortical areas that also provide visual input to the Pul.
For example, neurons that respond to visual stimuli but
not to saccades and vice versa have been described in area PMLS (Yin and Greenwood 1992a,b) from which another direct projection to the LP-P arises (Updyke 1981).
217
Functional considerations
In cat, no direct evidence that would support any welldefined Pul function is available. Although Pul neurons
showed significant differences in their responses to the
stimuli used in our study, all three neuronal populations
identified could be involved in the modulation of visual
information flow when saccadic eye movements are executed. The different Pul cell response properties may reflect the fact that neuronal activity in different Pul target
structures has to be modulated specifically according to
their own functional roles in the visual pathway. Thus, S
neurons could provide a strong signal that indicates the
appearance of a new visual surround as a result of saccadic eye movements to cortical areas, or they could
generate saccadic suppression in cortical areas if their
activity is fed into appropriate inhibitory circuits. However, because we must assume that these cells do not receive significant visual input, we can only speculate on
the origin of saccade-related input to S neurons. As their
responses did not show any obvious selectivity for saccade parameters, providing information about the occurrence of a saccadic eye movement may be more important than coding any saccade parameter.
The activation of V neurons, which are able to discriminate between either externally or internally generated retinal image motion, could lead to facilitation of visual responses after saccades in visual cortical areas. SV
neurons, which do not distinguish between external image motion and saccades, could more globally induce
arousal in visual cortical areas after changes of visual
surround following both self-induced eye movements
or external motion. In monkey, a function in the guidance of visual attention has been proposed for the Pul
(Robinson et al. 1991). Because of evident analogies in
response properties between neurons in the Pul of cats
and monkeys, the cat Pul may also integrate visuomotor
information in an attentional context.
Acknowledgements We wish to thank K.P. Hoffmann for valuable support and discussions during experiments and preparation
of the manuscript and Ms M. Schmidt for her excellent animal
care. This work was supported by a grant from the Deutsche Forschungsgemeinschaft (SFB 509, ‘Neurovision’, TP A3).
References
Abramson BP, Chalupa LM (1985) The laminar distribution of
cortical connections with the tecto- and corticorecipient zones
in the cat’s lateral posterior nucleus. Neuroscience 15:81–95
Abramson BP, Chalupa LM(1988) Multiple pathways from the superior colliculus to the extrageniculate visual thalamus of the
cat. J Comp Neurol 271:397–418
Berson DM, Graybriel AM (1978) Parallel thalamic zones in the
LP-pulvinar complex of the cat identified by their afferent and
efferent connections. Brain Res 147:139–148
Berson DM, Graybriel AM (1983) Organisation of the striate recipient zone of the cat’s lateralis posterior-pulvinar complex and its
relation to the geniculostriate system. Neuroscience 9:337–372
Casanova C, Molotchnikoff S (1990) Influence of the superior colliculus on visual responses of cells in the rabbit’s lateral posterior nucleus. Exp Brain Res 80:387–396
Casanova C, Freeman RD, Nordmann JP (1989) Monocular and
binocular response properties of cells in the striate-recipient
zone of the cat’s lateral posterior-pulvinar complex. J Neurophysiol 62:544–557
Casanova C, Savard T, Darveau S (1997) Contribution of area 17
to cell responses in the striate-recipient zone of the cat’s lateral
posterior-pulvinar complex. Eur J Neurosci 9:1026–1036
Chalupa LM (1991) Visual function of the pulvinar. In: Leventhal
AG (ed) The neural basis of visual function. Macmillan,
Basingstoke, pp 140–159
Chalupa LM, Abramson BP (1988) Receptive field properties in
tecto- and striate-recipient zones of the cat’s lateral posterior
nucleus. In: Hicks TP, Benedeck G (eds) Progress in brain research, vol 75. Elsevier, Amsterdam, pp 85–94
Chalupa LM, Abramson BP (1989) Visual receptive fields in the
striate-recipient zone of the lateral posterior-pulvinar complex.
J Neurosci 9:347–357
Cucchiaro JB, Bickford ME, Sherman SM (1991) A GABAergic
projection from the pretectum to the dorsal lateral geniculate
nucleus in the cat. Neuroscience 41:213–226
Desimone R, Duncan J (1995) Neural mechanisms of selective visual attention. Annu Rev Neurosci 18:193–222
Desimone R, Wessinger M, Thomas L, Schneider W (1990) Attentional control of visual perception: cortical and subcortical
mechanisms. Cold Spring Harb Symp Quant Biol 55:963–971
Fischer WH, Schmidt M, Hoffmann K-P (1998) Saccade induced
activity of dorsal lateral geniculate nucleus X and Y cells during pharmacological inactivation of the cat pretectum. Vis
Neurosci 15:197–210
Fish SE, Chalupa LM (1979) Functional properties of pulvinarlateral posterior neurons which receive input from the superior
colliculus. Exp Brain Res 36:245–257
Godfraind J-M, Meulders M, Veraart C (1972) Visual properties of
neurons in the pulvinar, nucleus lateralis posterior and nucleus
suprageniculatus thalami in the cat. I. Quantitative investigation. Brain Res 44:503–526
Graybiel AM, Berson DM (1980) Histochemical identification and
afferent connections of subdivisions in the lateralis posterior–pulvinar complex and related thalamic nuclei in the cat.
Neuroscience 5:1175–1238
Guillery RW, Geisert EE Jr, Polley EH, Mason CA (1980) An
analysis of the retinal afferents to the cat’s medial interlaminar
nucleus and to its rostral thalamic extension, the “geniculate
wing”. J Comp Neurol 194:117–142
Infante C, Leiva J (1984) Correlation between pulvinar-lateralis
posterior complex unit activity and eye movements in the cat.
Exp Neurol 85:453–460
Itoh K, Mizuno N, Sugimoto T, Nomura S, Nakamura Y, Konishi
A (1979) A cerebello-pulvino-cortical and a retino-pulvinocortical pathways in the cat as revealed by the use of the anterograde and retrograde transport of horseradish peroxidase.
J Comp Neurol 187:349–357
Judge SG, Richmond BJ, Chu FC (1980) Implantation of magnetic
search coils for measurement of eye position: an improved
method. Vision Res 20:535–538
Kasper H, Hess JM (1991) Magnetic search coil system for
linear detection of three-dimensional angular movements.
IEEE Trans Biomed Eng 38:466–475
Kubota T, Morimoto M, Kanaseki T, Inomata H (1988) Visual pretectal neurons projecting to the dorsal lateral geniculate nucleus and pulvinar nucleus in the cat. Brain Res Bull 20:573–579
Leventhal AG, Keens J, Tork I (1980) The afferent ganglion cells
and cortical projections of the retinal recipient zone (RRZ) of
the cat’s pulvinar complex. J Comp Neurol 194:535–554
Mason R (1981) Differential responsiveness of cells in the visual
zones of the cat’s LP-pulvinar complex to visual stimulation.
Exp Brain Res 43:25–33
Merabet L, Desautels A, Minville K, Casanova C (1998) Motion
integration in a thalamic visual nucleus. Nature 396:265–268
Petersen SE, Robinson DL, Keys W (1985) Pulvinar nuclei of the
behaving rhesus monkey: visual responses and their modulation. J Neurophysiol 54:867–886
218
Reinoso-Suarez F (1961) Topographischer Hirnatlas der Katze für
experimental-physiologische Untersuchungen. Merk, Darmstadt
Robinson DL, McClurkin JW (1989) The visual superior colliculus and pulvinar. In: Wurtz RH, Goldberg ME (eds) The neurobiology of saccadic eye movements. Elsevier, Amsterdam,
pp 337–360
Robinson DL, Petersen SE, Keys W (1986) Saccade-related and
visual activities in the pulvinar nuclei of the behaving rhesus
monkey. Exp Brain Res 62:625–634
Robinson DL, McClurkin JW, Kertzman C, Petersen SE (1991)
Visual responses of pulvinar and collicular neurons during eye
movements of awake, trained monkeys. J Neurophysiol 66:485–
496
Schmidt M (1996) Neurons in the cat pretectum that project to the
dorsal lateral geniculate nucleus are activated during saccades.
J Neurophysiol 76:2907–2918
Schweigart G, Hoffmann K-P (1992) Pretectal jerk neuron activity
during saccadic eye movements and visual stimulation in the
cat. Exp Brain Res 91:273–283
Sudkamp S, Schmidt M (1995) Physiological characterization of
pretectal neurons projecting to the lateral posterior-pulvinar
complex in the cat. Eur J Neurosci 7:881–888
Ungerleider LG, Christensen CA (1979) Pulvinar lesions in monkeys produce abnormal scanning of a complex visual array.
Neuropsychologia 17:493–501
Updyke BV (1977) Topographic organisation of the projection
from cortical areas 17, 18 and 19 onto the thalamus, pretectum
and superior colliculus in the cat. J Comp Neurol 173:81–
122
Updyke BV (1981) Projections from visual areas of the middle suprasylvian sulcus onto the lateral posterior complex and adjacent thalamic nuclei in cat. J Comp Neurol 201:477–506
Wahle P, Stuphorn V, Schmidt M, Hoffmann KP (1994) LGNprojecting neurons of the cat’s pretectum express glutamic
acid decarboxylase mRNA. Eur J Neurosci 6:454–460
Weber JT, Chen I-L, Hutchins B (1986) The pretectal complex of
the cat: cells of origin projecting to the pulvinar nucleus. Brain
Res 397:389–394
Yin TCT, Greenwood M (1992a) Visuomotor interactions in responses of neurons in the middle and lateral suprasylvian cortices of the behaving cat. Exp Brain Res 88:15–32
Yin TCT, Greenwood M (1992b) Visual response properties of
neurons in the middle and lateral suprasylvian cortices of the
behaving cat. Exp Brain Res 88:1–14