Download Saccade-related activity of periaqueductal gray matter of the

Saccade-Related Activity of Periaqueductal
Gray Matter of the Monkey
Manabu Kase, Renpei Nagara, and Masamichi Koto
Single unit activity was recorded extracellularly from the periaqueductal gray matter (PAG) and the
superior colliculus of three monkeys during spontaneous saccades and fixation. Most saccade-related
cells were found in the dorsal and lateral regions of the PAG and they paused with saccades. The pause
preceded the onset of saccades by 34.5 ± 6.8 msec (n = 31). In 90% of the 31 PAG cells, the saccaderelated modulation in activity occurred in all directions. There was a linear relationship between pause
duration and saccade duration with correlation coefficients larger than 0.70 in most cells. Superior
colliculus cells showed bursts preceding onset of saccades. The lead times averaged 25.3 msec (n = 35).
There was no linear relationship between burst duration and saccade duration. These results suggest
that the dorsal and lateral regions of PAG play an important role in the saccadic system, probably
through long lead burst units in the deep layer of the superior colliculus and/or pontine reticular formation.
Invest Ophthalmol Vis Sci 27:1165-1169, 1986
Lesions of the periaqueductal gray matter (PAG) are
known to induce abnormal eye movements (periaqueductal gray syndrome1). Although the PAG is surrounded by important structures related to eye movements, such as the superior colliculus, interstitial nucleus of Cajal, and nucleus of Darkschewitsch, it is not
known whether the PAG itself takes part in eye movement control.
Anatomical data concerning the connections between the PAG and eye movement-related structures
are still obscure, but degeneration 23 and autoradiographic transport4'5'6 experiments indicate that the
dorsal and lateral regions of PAG below the pretectum
and superior colliculus send axons to the ipsilateral
superior colliculus. Also, the PAG receives inputs from
the fastigal nuclei7 which are related to oculomotor
function.8
In this study, we have analyzed the activity of PAG
cells in association with eye movements and compared
it with the activity of neurons in the superior colliculus.
The data show a clear temporal relation between PAG
neural activity and saccades. A preliminary report of
these results has been presented elsewhere.9
Materials and Methods
The experiments were performed on three alert
monkeys (macaca fascicularis). First, under general
anesthesia EOG-electrodes were implanted bilaterally,
and above and below the left orbit for recording eye
movements in the horizontal and vertical planes, respectively. A stainless steel recording chamber was
placed over the parietal bone at anterior plane 5 mm
according to the stereotaxic atlas by Kusama and Mabuchi.10 Transverse tubes were fixed to the skull with
dental acrylic. Neural activity and ocular movements
were then investigated for about 3 months. During recording sessions, the monkey sat in a specially designed
primate chair. The head was painlessly immobilized.
The animal viewed a rear projection screen subtending
60° of the visual field horizontally and 45° vertically.
Stainless steel microelectrodes (Elgiloy 0.25 mm)
insulated with Isonel 31 were introduced into the brain
through the recording chamber. They were guided to
just above the superior colliculus through a 23-gauge
cannula and were driven from the cannula into the
PAG by a hydraulic microdrive.
Extracellular action potentials and horizontal and
vertical EOGs were displayed on a polygraph, recorded
on magnetic tape, and photographed. Lead times of at
least 20 saccades were measured from film records with
errors of less than half a millisecond. Spontaneous
mean discharge rates were calculated during 10 fixation
periods which lasted for more than 2 sec. This investigation was carried out in adherence to the principles
From the Department of Ophthalmology and the Department of
Physiology, Hokkaido University, School of Medicine, Sapporo 060,
Japan.
Submitted for publication: June 11, 1984.
Reprint requests: Manabu Kase, MD, Department of Ophthalmology, Hokkaido University, School of Medicine, Sapporo 060,
Japan.
1165
Downloaded From: http://iovs.arvojournals.org/ on 04/30/2017
Vol. 27
INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / July 1986
1166
MLF
2mm
Fig. 1. Anatomical location of recording sites of
pause cells and burst cells
(A-C) and microphotograph
of transverse sections showing an electrolytic lesion in
the PAG (arrow) (D). Open
circles and filled circles in A,
B, and C indicate pause cells
and burst cells, respectively.
Most pause cells were located
in the dorsal and lateral regions of the PAG underlying
the superior colliculus and
pretectum. No pause cells
were recorded from the ventral region and the PAG below the level of the trochlear
nucleus. All the burst cells
were observed in the deep
layer of the superior colliculus and the dorsal mesencephalic reticular formation.
The interrupted lines in D
denote microelectrode tracks.
The sections of the brains
were cut at an angle of 30°
to the coronal plane of the
stereotaxic atlas by Snider
and Lee," so that the ventral
portions of sections were located more rostrally than in the atlas. PAG = periaqueductal gray matter, SC = superior colliculus, MLF = medial
longitudinal fasciculus, III = oculomotor nucleus.
set forth in the ARVO Resolution on the Use of Animals in Research.
Electrolytic lesions were made at the site of representative units by passing 20 juA cathodal current for
20 sec through the recording electrode. At the end of
experiments, the monkey was deeply anesthetized with
ketamine and nembutal and perfused with saline, potassium ferrocyanide, and 10% formalin. Recording
sites were anatomically reconstructed with reference
to a microelectrode track at the center of an active
zone, which was histologically identified by the spot
resulting from the Prussian btue reaction to deposited
iron.
Results
A total of 130 units exhibiting phasic activity changes
in association with spontaneous saccades, were recorded from the PAG and the superior colliculus. All
had predominantly negative spikes, often with an inflection on their rising phase and amplitudes ranging
from 200 /xV to 2.5 mV. This indicates that we recorded
from cell bodies rather than axons.
Sixty-six saccade-related cells were quantitatively
analyzed. They were classified into two classes: 31 cells
Downloaded From: http://iovs.arvojournals.org/ on 04/30/2017
were pause units which ceased firing in close temporal
relation to saccadic onset; 35 cells were burst units
which fired prior to saccadic onset.
Localization of Saccade-Related Units
Histological reconstruction of the brain stem showed
that recording sites for the two classes of saccade-related
units were different: most pause units were clustered
in a limited dorsolateral region of the PAG underlying
the pretectum and the superior colliculus. None were
in the ventral PAG region or in the mesencephalic reticular formation (Fig. 1). Antero-lateral locations corresponded to the anterior 5 mm-7 mm stereotaxic
planes of Kusama and Mabuchi atlas10 (corresponding
to the posterior 0 mm to 2 mm planes of the stereotaxic
Snider and Lee atlas"). No burst units were found in
the PAG. They were located more superficially and
laterally in the deep layer of the superior colliculus and
dorsal mesencephalic reticular formation.
Saccade-Related Cells in the PAG
All the saccade-related units in the dorsal and lateral
PAG regions showed a pause that was time-locked to
the onset of saccades, and they had tonic discharges
1167
SACCADE-RELATED ACTIVITY OF PERIAQUEDUCTAL GRAY CELLS / Kase er ol.
No. 7
N=31
X=34.8msec
B
cc 5-
JU
-100
-400
-200
200
400 (msec)
20
Fig. 2. An example of PAG cells exhibiting pauses in relation to
saccadic onset. A, original record of a pause cell. Time calibration
dots at 4 Hz. Upper trace of EOGs were horizontal component and
lower trace vertical component. Upward deflections in upper and
lower traces denote rightward and upward eye movements, respectively. Calibration bars of EOGs mean 10°. B, raster generated from
16 spontaneous saccades. C, perisaccadic time histogram of the same
cell. Zero in B and C denotes the onset of saccades.
during steady fixation. The average spontaneous discharge rates in 29 PAG cells were calculated during
the periods of fixation, which lasted for more than 2
sec before the saccades. Tonic discharge rates varied
from cell to cell (mean values from 30-79 spikes/sec,
with an average of 57.8 spikes/sec). One cell showed
fluctuating discharge frequencies (Fig. 2). This cell exhibited presaccadic discharges at 50-190 spikes/sec.
The discharge rates did not change in relation to eye
position or to slow eye movements. There was no correlation between discharge rate and period of fixation.
Visual inputs, such as light on and off, proved to be
ineffective to alter the discharge rates. The pause onset
preceded the saccades. The cessation of firing preceded
the onset of saccades with an average lead time of 34.4
msec (Fig. 2B, C), varying from 20-100 msec (Fig. 2B).
The average lead time of the 31 pause cells was 34
± 6.8 msec (mean ± SD), ranging from 20.4-53 msec
(Fig. 3 A). Twenty-eight of the 31 pause cells (90%) exhibited pauses for all saccades, regardless of their directions. In three cells (10%), pauses of activity occurred
during ipsilateral saccades. No cells had saccade-related
modulation in activity during eye movements in the
other directions. The pause durations were linearly related to saccade durations, since the correlation coefficients between them were larger than 0.70 in 29 pause
cells. The regression lines of typical ten cells are shown
in Figure 4. Their coefficients ranged from 0.71 to 0.91.
Downloaded From: http://iovs.arvojournals.org/ on 04/30/2017
ll
N = 35
X=25.3msec
40
LEAD TIME
60
(msec)
80
Fig. 3. Distribution of lead times in pause cells (A) and burst cells
(B). Time bin of the columns was 5 msec in A and 2 msec in B. The
lead time was measured in individual pause cell from the last spike
of the tonic discharges and in burst cell from the first spike of the
bursts to the onset of 20 saccades.
Pause durations tended to be longer than saccade durations.
Saccade-Related Cells in the Superior Colliculus
Thirty-five cells were histologically identified in the
deep layer of the superior colliculus or in the dorsal
300-,
200-
03
O
O
100-
CO
100
Pause
200
duration
300
(msec)
Fig. 4. Pause durations as a function of saccade duration for ten
typical pause cells. All regression lines had correlation coefficients
ranging from 0.71-0.91. The pause durations tended to be longer
than the saccade durations.
1168
INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / July 1986
mesencephalic reticular formation. These cells showed
bursts starting clearly before the onset of saccades. The
lead times of the bursts (n = 35) ranged from 16.734.1 msec with an average of 25.3 msec (Fig. 3B). Although there was an intimate temporal relationship
between burst onset and saccade onset, burst durations
and saccade durations were not related in the majority
of these cells. Twenty-eight of the 35 burst cells (80%)
were direction selective. The directional preference was
oblique for 15 cells, contralateral for 11 cells, and vertical for 2 cells. Seven showed no directional preference.
Their saccade-related modulations were not analyzed
in detail any further because they had the same characteristics of activity as described in previous reports.12-13
Discussion
The present study has shown the existence of cells
in the PAG which pause in association with spontaneous saccades. The cells were located in the dorsal
and lateral PAG in a limited region at the level of the
pretectum and the superior colliculus. This is in a contrast with the previous report by Matsunami. 14 According to him, the PAG units in the diencephalon of
the monkey had bursts preceding saccadic onset by 8.4
msec. We found no burst units in the PAG of the mesencephalon, and all our burst cells were located in the
deep layer of the superior colliculus and the dorsal
mesencephalic reticular formation. The discrepancy
between the two sets of data might result from a difference in recording sites. From the published figures
of Matsunami, the majority of his burst units recorded
seemed to be located in more rostral part of PAG than
our recording sites.
Wurtz et al12 have described pause units that lie deep
in the colliculus bordering on the deep layer, whose
relation to eye movements was similar to that of pause
units in the pontine reticular formation. In contrast to
the pause units in the pontine reticular formation which
have an average lead time of 16 msec, 1516 pause cells
of the PAG in the present study had significantly longer
lead times (34.8 ± 6.8 msec). This value was also substantially larger than that of pause units in the flocculus17 and vestibular nuclei.18
The present study has also shown that the PAG cells
stopped firing before cells in the ipsilateral superior
colliculus began to fire. This suggests that there may
be an inverse relationship between activity of PAG and
superior colliculus cells. Anatomical data show that
the efferent fibers from the dorsal and lateral regions
of PAG project to ipsilateral superior colliculus.3'4'5 It
is possible that the pause PAG cells may inhibit the
Downloaded From: http://iovs.arvojournals.org/ on 04/30/2017
Vol. 27
superior colliculus cells during spontaneous saccades,
in the same way as the pause cells in the pontine reticular formation inhibit the burst cells in pontine reticular formation,16 and substantia nigra cells exhibiting
a pause during visually guided saccades inhibit the superior colliculus cells.19 However, it is anatomically
known that the PAG also projects fibers to the pontine
reticular formation in the monkey,6 where saccade-related long lead burst units are located. Therefore, the
PAG cells may also inhibit these units.
Besides lead times, there is another difference between our PAG cells and pontine reticular formation
units. The pause cells in the PAG had relatively irregular tronic discharges during intersaccadic intervals,
whereas pause units in the pontine reticular formation,
as well as in the flocculus and vestibular nucleus, had
high and regular tonic discharges ranging from 50-200
spikes/sec. Taking lead time and discharge frequency
into consideration, it seems likely that the PAG cells
may regulate presaccadic activity influencing superior
colliculus cells and/or long lead bursts units in the
pontine reticular formation.
Key words: periaqueductal gray matter, superior colliculus,
saccade, pause units, saccade-related modulation
Acknowledgments
We thank Dr. H. Matsuda, the Department of Ophthalmology, Hokkaido University School of Medicine, Dr. H.
Noda, the Department of Physiology, Indiana University
School of Optometry, and Dr. J. Schlag, the Department of
Anatomy, UCLA School of Medicine for their valuable criticism during the preparation of this manuscript.
References
1. Hoyt WF and Daroff RB: Supranuclear disorders of ocular control
system in man. In The Control of Eye Movements, Bach-y-Rita
P, Collins CC, and Hyde JE, editors, New York, Academic Press,
1971, pp. 175-235.
2. Hamilton BI: Efferent connections of the periaqueductal gray
matter in the cat. J Comp Neurol 139:105, 1970.
3. Hamilton BI: Projections of the nuclei of the periaqueductal
gray matter in the cat. J Comp Neurol 152:45, 1973.
4. Grofova I, Ottersen OP, and Rinvik E: Mesencephalic and.diencephalic afferents to the superior colliculus and periaqueductal
gray substance demonstrated by retrograde axonal transport of
horseradish peroxidase in the cat. Brain Res 146:205, 1978.
5. Mantyh PW: Connections of midbrain periaqueductal gray in
the monkey. I. Ascending efferent projections. J Neurophysiol
49:567, 1983.
6. Mantyh PW: Connections of midbrain periaqueductal gray in
the monkey. II. Descending efferent projections. J Neurophysiol
49:582, 1983.
7. Hirai T, Onodera S, and Kawamura K: Cerebellotectal projections studied in cat with horseradish peroxidase or tritiated amino
acids axonal transport. Exp Brain Res 48:1, 1982.
8. Hepp K, Henn V, and Jaeger J: Eye movement related neurons
No. 7
9.
10.
11.
12.
13.
SACCADE-RELATED ACTIVITY OF PERIAQUEDUCTAL GRAY CELLS / Kose er ol.
in the cerebellar nuclei of the alert monkey. Exp Brain Res 45:
253, 1982.
Nagata R, Kase M, and Kato M: Eye movement-related neurons
at the central gray matter of the monkey. Neurosci Letters Supp
13:82, 1983.
Kusama T and Mabuchi M: Stereotaxic Atlas of the Brain of
Macaca Fuscata. Kusama T and Mabuchi M, editors, Tokyo,
University of Tokyo Press, 1970.
Snider RS and Lee JC: A Stereotaxic Atlas of the Monkey Brain
(Macaca Mulatta). Snider RS and Lee JC, editors, Chicago, The
University of Chicago Press, 1961.
Schiller PH and Stryker M: Single unit recording and stimulation
in superior colliculus of the alert monkey. J Neurophysiol 35:
915, 1972.
Wurtz RH and Goldberg ME: Activity of superior colliculus in
behaving monkey. II. Cells discharging before eye movements.
J Neurophysiol 35:575, 1972.
Downloaded From: http://iovs.arvojournals.org/ on 04/30/2017
1169
14. Matsunami K: Saccadic eye movement and neurons in the central
gray area in awake monkeys. Brain Res 38:217, 1972.
15. Luschei ES and Fuchs AF: Activity of brainstem neurons during
eye movements of alert monkeys. J Neurophysiol 35:445, 1972.
16. Keller EL: Participation of medial pontine reticular formation
in eye movement generation in monkey. J Neurophysiol 37:316,
1974.
17. Noda H and Suzuki DA: The role of the flocculus of the monkey
in saccadic eye movements. J Physiol 294:317, 1979.
18. Miles FA: Single unit firing patterns in the vestibular nuclei related
to voluntary eye movements and passive body rotation in conscious monkeys. Brain Res 71:215, 1974.
19. Hikosaka O and Wurtz RH: Visual and oculomotor functions
of monkey substantia nigra pars reticulata. IV. Relation of substantia nigra to superior colliculus. J Neurophysiol 49:1285, 1983.