Tissue Responses of the Monkey Retina: Tuning
and Dependence on Inner Layer Integrity
William W. Dawson,*f Robert D. Srrarron,* G. Morion Hope,* Ronald Parmer,*
Harry M. Engel,* and Matt J. Kessler^:
Recent technical and conceptual advances have made it possible to experiment with models of local
human inner retinal disease and changes in very small, tissue-specific signals. Local retrograde degeneration of the ganglion cells was induced in four rhesus monkeys by 160° microdiathermy fiber layer
burns at the nasal or temporal edges of the optic disc. There were no abnormalities of the classical
electroretinograms (ERGs) during the following 210 days. With nasal lesions, pattern-evoked retinal
(PERR) and cortical responses over a range of grating contrasts and spatial frequencies were largely
normal. The authors found a cortical spatial tuning peak near 0.5 cycles/deg (cpd) and a retinal peak
at 0.25-0.3 cpd. With temporal lesions, the retinal signals to high frequency stimuli (>1.0 cpd) approached
zero between 20-60 days. The cortical evoked signal declined with a course similar to the retinal components. Histological evidence was found for extensive loss of ganglion cells and fibers in a central 3 0 40° temporal area, including the macula, 210 days after the temporal lesions. This is strong evidence
that local ganglion cell-dependent electrical potentials, bearing little relation to the ERG, can be measured
in response to selected stimuli. Invest Ophthalmol Vis Sci 27:734-745, 1986
During this century, research on the electrical action
potential recorded from the vertebrate cornea in response to retinal stimulation (electroretinogram, ERG)
has produced a body of literature which probably establishes this as one of the most thoroughly studied of
all bioelectric signals. Much of the basic effort has been
devoted to components and tissue origins of the signal,
typified by Granit's 1 pioneering investigations of the
signal changes accompanying physiological and chemical impairment of particular retinal cell classes. Early
clinical research correlated tissue-specific retinal pathology to associated ERG abnormalities, as typified
by Karpe's 2 reports from this period. The progression
of this research has been traced in recent reviews.3 The
resultant body of information on the cellular origins
of the ERG, including more recent intra- and extracellular microelectrode studies,4"7 allows an understanding of retinal neurophysiology unparalleled by
that of any comparable neural tissue. A feature of this
impressive body of evidence and theory is the assignment of the traditional ERG to specific classes of glial
and outer-retinal cells and insensitivity to inner-retinal,
spatio-temporal neural processing and/or inner-retinal
pathology. This generally and firmly held concept of
inner-retinal independence probably influenced early
interpretation of pattern-evoked coraeal potentials as
ERGs of unusually small amplitude due to reduced
intraocular light scatter and/or photopic adaptation
levels, both features of this stimulus configuration.89
Against this very structured background, Maffei and
Fiorentini10 reported loss of pattern-evoked responses
at the cornea, but normal flash-evoked ERGs, following
induced optic atrophy in cats. Dawson, Maida, and
Rubin" reported similar findings from a human patient
with unilateral optic nerve section. In this, as in all
clinical research on this subject, the selectivity of the
lesions was assumed. Subsequently, research from both
clinical (~90% of total) and basic science laboratories
has demonstrated interrelationships between patternevoked retinal signals and inner-retinal pathology,
spatial, and temporal stimulus properties (tuning) and
independence of outer-retinal disease.12"17
However, not all findings support the association of
PERR with inner retina and there is controversy on
this subject. Normal PERR production in the presence
of clinically diagnosed inner-retinal lesions has been
reported.18 Depth recordings have not indicated an inner-retinal origin for pattern elicited potentials.19 Also,
there have been theoretical arguments for outer-retinal
From the Departments of Ophthalmology,* Physiology,t and
Neuroscience.f College of Medicine, University of Florida, Gainesville, Florida, and Caribbean Primate Research Center,:): University
of Puerto Rico School of Medicine, San Juan, Puerto Rico.
This work was supported in part by NEI grant 5 R01 EY04460
and an unrestricted Departmental grant from Research to Prevent
Blindness (UF) and NIH grant RR-01293 (UPR).
Submitted for publication: May 20, 1985.
Reprint requests: Dr. William W. Dawson, Department of Ophthalmology, Box J—284, JHMHC, University of Florida, Gainesville,
FL 32610.
734
Downloaded From: http://arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933130/ on 07/31/2017
No. 5
RETINAL RESPONSES AFTER NERVE FIDER LESIONS/Dowson er ol.
origins of these signals.20 Clearly, these issues, innerretinal dependence and cellular origin, remain unresolved, while many other questions have yet to be
posed. For example, there are no data on effects of
histologically verified, characterized, local, inner-retinal
lesions in species having primate macular specializations, on central retinal/macular-specific lesions and
responses, or on differentiation of components of the
PERR. In order to address some of these problems, we
have recorded macular PERRs and ERGs from the
corneas of rhesus monkeys (Macaca mulatto) after microdiathermy lesions to a 160° arc along the temporal
margin of the optic papilla (experimental), similar lesions on the nasal side (operated control) and no lesion
(normal control), with histological verification of selective atrophy of the central retinal/macular ganglion
cell population in the experimental group.
705
30 Hz
20 Hz
15 Hz
Materials and Methods
Five juvenile (aged 2-4 yr) rhesus monkeys (Macaca
mulatto) from the Caribbean Primate Research Center
were shipped to the AAALAC accredited facility at the
University of Florida where they were housed in individual cages. All animal use in the research was consistent with the ARVO Resolution on Use of Animals
in Research. The animals were clinically screened by
ophthalmoscopy, under mydriasis and cycloplegia.
Clinical electrophysiology was applied to detect any
early degenerative outer retinal disease. All monkeys
had clinically normal eyes.
Prior to electrical recording, each animal was fitted
with a scleral ERG contact lens similar to that of Riggs.8
Special attention was given to the optical quality of the
front surface so that normal resolution could be attained with corrections of no more than 2 diopters.
There was a 4 mm artificial pupil. The cornea
was anesthetized with topical proparacaine drops
(Ophthaine; Squibb), cycloplegia and mydriasis was
achieved and, under ketamine-xylazine anesthesia, the
trachea was intubated. Atropine was administered IM
to reduce secretions and a patent IV drip was established for fluid supplementation. Subsequent to the
initial ketamine-xylazine anesthesia, tranquilization
was produced with diazipam and muscular relaxation
was achieved with pancuronium bromide. 20 Small
quantities of the ketamine-xylazine mixture were administered periodically to insure stable analgesia as indicated by EEG records. Respiration was assisted with
constant monitoring to maintain end-tidal CO 2 between 3.8 and 4.4%. Eye signals were recorded between
the active corneal electrode and an indifferent electrode
placed immediately adjacent to the eye at the lateral
canthus. Signals could not be recorded at the right (OD)
eye when the left (OS) eye was stimulated. Visual cor-
10 Hz
40.0 msec
Stimulus
Fig. 1. Rhesus pattern-evoked retinal signals with different frequency limit constraints. Two major components (PI and Nl) are
identified for amplitude analysis. A previously unreported component
(circled) depends on frequency limit. Grating pattern, 52% contrast,
2.5 Hz horizontal counterphase rate, 0.66 cpd (1.5° elements) spatial
frequency. Low frequency limit is 0.2 Hz.
tical signals were recorded by an active silver-silver
chloride disc electrode attached with conductive paste
to the scalp surface above the lunate gyrus. The cortical
reference electrode was on the adjacent earlobe.21 A
ground lead was attached to the forehead. After
ophthalmoscopy, streak retinoscopy and the addition
of corrective lenses, stimuli (2.5 Hz square wave, optically counterphased grating or 0.25 Hz flash) were
presented to each eye on a 30° tangent screen at 1
meter. The square wave gratings had an integrated luminance of 1.4 log foot Lamberts (fL) with contrast
adjustable from 17-52%. The flash stimulus originated
from the diffused output of a Grass PS22 Xenon flash
tube (Grass; Quincy, MA). Flash-integrated luminance
of the tangent screen measured by a UDT photometer
was 20 fL • sec. The tangent screen was Vi meter before
a diffusely illuminated ( ~ 1.4 log fL) rectangular white
curtain that subtended 100°. Eyes were aligned with
the screen by centration of the reflex. Stimulation,
electrical signal recording, and processing were otherwise as described earlier.22 Passband was 0.2-300 Hz
(3db points) except as specified in Figure 1.
Peripapillary lesions were made in the fiber layer.
Under sterile conditions and deep anesthesia a scle-
Downloaded From: http://arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933130/ on 07/31/2017
736
INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / May 1986
Table 1. Eye treatment—lesion status
Animal number
B46
B860
B41
A3
B42
OD
OS
Temporal 160°
Normal
Normal
Normal
Normal
Normal
Nasal 160°
•Temporal 160°
•Temporal 160°
Normal
* Post surgical vitritis.
rotomy was made at the pars plana. A bipolar diathermy electrode (Clinitex; Beverly, MA) was inserted
under operating microscope control and microdiathermy lesions were made without touching the retina.
Five or six white round spots about 0.2 disc diameters
were made in a partial crescent to be confluent at the
disc margin. Blood vessels were not involved in the
lesions. The production of these lesions, their discrete
nature, and the resulting nerve fiber bundle defects have
been fully described.23
As described in Table 1, three monkeys had lesions
placed over a 160-170° area at the temporal margin
of the disc. One had a 160° lesion placed along the
nasal margin of the disc and one remained as an unoperated control. The left eye of B-860 (nasal lesion)
and the right eye of B-46 (temporal lesion) were enucleated after 210-220 days. The two eyes were fixed
immediately following enucleation with the slow intravitreal injection of Karnovski's solution. After 5 min
the anterior segment was removed and the eye cups
and 3-mm sections of the optic nerves were immersed
in cold Karnovski's solution. The flat mounted retina
was embedded in Epon-Araldite and subsequently
buttons of tissue were sectioned at 1-2 nm, stained
with toluidine blue, and examined by interference
contrast microscopy. All monkeys were returned to the
Caribbean Primate Research Center.
Results
All eyes responded similarly to the local microdiathermy treatment. Animal B-41 and A-3 developed
mild vitritis that was expressed primarily as a vitreal
clouding which persisted for about 10 days and cleared
completely. All animals showed ophthalmoscopic evidence of edema within 1 disc diameter of the lesioned
area for approximately 20 days. In one (B-41) this was
followed by very local gliosis in the region of the fiber
layer up to half a disc diameter from the lesion. After
about 30 days, in the temporally lesioned eyes there
were changes in coloration of the arcuate area extending
from the lesion margins toward, superior and inferior
to, the macular area. These were interpreted clinically
as indicative of changes in the fiber layer. These changes
were not as evident on the nasal side (B-860), but the
Vol. 27
other sequelae were present. In B-46, B-41, and A-3
there was a clinical loss in "reflectivity" of the fovea
during the final 50 days of the study. No retina or pigment epithelium changes were ever identified at any
other retinal sites.
The electrophysiological record bandwidth is important in the resolution of the primate retinal response
to pattern stimulation (PERR). The signals in Figure
1 were recorded with a lower frequency limit (digital
filtering, exponential rolloff) at 0.2 Hz. Using the fast
Fourier analysis and inverse transform it is possible to
remove specific frequency domain information without
altering phase. Repeated digital filtration (with due attention to sampling limits and artifacts) of the wideband (0.2-1 KHz) retinal signal (upper trace) yields a
family of waveforms (lower traces) temporally related
to the visual stimulus. The character of the wideband
PERR (upper trace) provides a rationale for amplitude
analysis of two components, PI and N l . However,
amplitude measures require subjective identification
of peaks and, as inspection of this trace demonstrates,
Nl amplitude may not be independent of PI amplitude
changes. Inspection of the lower five waveforms in Figure 1 illustrates that PI is virtually eliminated if frequency information above 15 Hz is excluded, whereas
little Nl information is lost if only frequencies from
15 Hz and below are considered. Thus, power spectral
density offers a means for objective quantification of
the signal and eliminating judgmental errors in peak
detection with small signals. Isolation of the two components in the frequency domain, avoids the problem
of amplitude dependence, allowing for meaningful
component analysis. The wide band signals (upper
trace) in these eyes typically show the small negative
deflection (implicit time = 20 msec) proceeding PI and
the small component (circled, Fig. 1) on the descending
phase of N l . Modal implicit times for PI and Nl are
about 37 and 70 msec, respectively. All components
were observed to vary with stimulus spatial frequency,
contrast, focus, and luminance.
Figure 2 provides evidence for spatial selectivity in
the rhesus eye. These are signals from B-46 before lesioning. Amplitudes (A) and power spectral density (B)
of the cortical signal and retinal PI and Nl are plotted
against spatial frequency. Cortical signal amplitude was
measured peak-to-peak, and PI and Nl amplitudes
were measured as described above. In the frequency
domain, more than 80% of the total power was at the
stimulus frequency harmonics, allowing use of a 'total
power' measure with less intrinsic bias. In (B), total
cortical power includes all power from 0.2-50 Hz,
whereas retinal power was divided into totals in two
bands, 2-11 Hz and 12-50 Hz. The smooth curves
were 'best fit' (all standard deviations < 1.0, most < 0.5,
scale units) by nonlinear regression analysis and all
Downloaded From: http://arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933130/ on 07/31/2017
No. 5
737
RETINAL RESPONSES AFTER NERVE FIBER LESIONS/Dowson er ol.
functions are of the form y = Ai + A2X + A3X2. The
total power measure in the 2-11 Hz band, roughly corresponding to N l , shows improved tuning relative to
the N1 amplitude function and more nearly approximates that of the cortical signal. The PI and 12-50 Hz
tuning functions are generally less similar to the cortical
functions than are those of Nl and the 2-11 Hz band.
Although not this clear in all eyes under our stimulus
and recording conditions, the presence of spatial tuning
in all functions in Figure 2 establishes that this important property of the PERR is found in the rhesus
eye. Spatial frequency tuning is a well-established
characteristic of inner retinal function in many species,
including man.
Typical retinal and cortical signals in response to a
series of patterns (grating bar width = 0.3-3.4 degs)
recorded over an extended pre- and post-lesion period
for a lesioned-control (Nasal) animal are shown in the
upper set of records of Figure 3. These do not differ
significantly from the potentials from the normal control eyes. Day-to-day variations may be due to anesthetic level, image quality, and electrode placement.
These are more pronounced in the cortical signals. A
particularly good set of cortical records were obtained
from B-860 (Nasal lesion) at +120 days. The peak positive signal here has an implicit time between 80 and
90 msec while the preceeding negative trough occurred
between 55 and 65 msec. These values parallel retinal
counterparts described above (Fig. 1).
The differential effects of lesion site and duration on
retinal and cortical signals are illustrated in Figure 3
by comparing data from experimental animal B-46
(Temporal) to lesion-control animal B-860 (Nasal).
Retinal and cortical signals, particularly for the smaller
gratings, are sharply reduced or eliminated by 120 days
by lesions along the temporal margin of the optic papilla. Very large gratings (ie, 1-3.4°) with high (33%
and 52%) contrast continue to produce PERRs after
210 days. But these responses were not recordable at
lower contrast levels. Other features of this figure, consistent across all temporally lesioned animals (B-41, B46, A-3), are that the PI amplitude is usually smaller
than the Nl amplitude, and there is a trend toward
early, post-lesion, selective enlargement of Nl. Polarity
reversal in the cortical signal can occur from day to
day, as a consequence of dipole "straddling" and small
changes in electrode position.
Power spectral density was not routinely utilized to
evaluate effects of lesions over time because the increased sampling requirements (large numbers of individual records) lengthened the recording sessions to
a degree inconsistent with the animals' welfare. However, other parameters emerging from the presentations
in Figures 1 and 2, PI and Nl amplitudes at spatial
frequencies, change over time after temporal lesions.
14
12
10
4
CO
z
O
D
m
-o
O
z
Q.
CO
UJ
W
3
<r
<
o
»cr
O
o
2
m
I
o
o
<
%
14
12
10
o
o
'a.
cc
UJ
O
QL
1.0
0.5
0.5
LOG SPATIAL FREQUENCY (c/deg)
Fig. 2. Tuning of PERR and cortical signals of the rhesus monkey.
Grating stimuli counterphased at 2.5 Hz, 33% contrast. A, component
analysis by amplitude is identified (P|, N,) in Figure I. B, Analysis
of (A) data in the frequency domain by power spectral density. Ordinate is total power in the bands 2-11 Hz or 12-50 Hz (retina), and
2-50 Hz (cortex). Cortical absolute value is scaled X 0.5. Curves are
fitted by nonlinear regression methods.
Downloaded From: http://arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933130/ on 07/31/2017
738
Vol. 27
INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / May 1986
NASAL
+210 days
R
C
0.5
0.3
TEMPORAL
Fig. 3. Simultaneously recorded PERR (R) and cortical (C) signals from the rhesus monkey, before and after nasal (B-860) and temporal
(B-46) retina lesions at the disc margin. Days are numbered relative to surgery. Grating stimuli were 0.3° (3.33 cpd) to 3.4° (0.29 cpd) in
subtense with a 2.5 Hz counterphase rate and 33% contrast. Calibrations (temporal-10 days) refer to all signals.
Figure 4 presents mean signal ratios (lesion/control eye)
for PI (A), Nl (B), and cortical signals (C) from B-46
in response to 52% contrast gratings for the 210 days
following surgery. Since the data in Figure 3 suggest
that signal loss may vary with stimulus spatial frequency, mean ratios (n = 4) for low (<1 cycle/degree)
and high (>1 cycle/degree) frequency gratings were
plotted separately. Some SD limits have been drawn
to indicate variance. Large variances (eg, A, day 40,
low frequency) result from zero values in the sample.
All data points are relative to the - 1 0 day (pre-surgery)
value, which was set to one. Lines are estimates of best
fit. Where temporal lesions were present (B-41, B-46,
A-3) trends were the same; decline—except Nl—on
the day of surgery (while under residual anesthesia)
followed by relative hyperresponsiveness with rapid
subsequent decline. Important specific points in this
figure are that responses to high spatial frequency stimuli decay more rapidly than those to low frequencies,
while the latter are generally more resistant to effects
of temporal lesions, and that PI and cortical responses
to high frequencies vary in nearly identical fashion,
while Nl lags slightly. Notably, there is poor correspondence between the behavior of the cortical re-
Downloaded From: http://arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933130/ on 07/31/2017
No. 5
RETINAL RESPONSES AFTER NERVE FIDER LESIONS/Dowson er ol.
sponse and P1 and N1 under the low frequency stimulus conditions.
All temporally lesioned eyes produced significantly
lower (P < 0.01) responses (stimuli 52% contrast and
> 1 cycle/degree) by 120 days. Normal animal (B-42)
and nasal-lesion animal (B-860) showed only random
variation of the various response ratios (lesion/control)
at all contrasts during a similar 220-day period (not
illustrated). At 33% contrast all temporal-lesion animals
gave measurable 210-day responses (eye or cortex) only
to low frequency stimuli. At 17% contrast and +120
days there were no responses to any pattern. Flashelicited ERGs were always normal, but could have been
contributed by nonlesioned retinal areas stimulated by
intraocular light scatter. To confirm that outer retina
can support normal ERGs when stimuli are confined
to the area having inner retinal lesions, we increased
the luminance (2.0 log fL) of the surround to differentially light adapted peripheral retina and masked the
stimulator to confine the flash (< 1.0 fL • sec) to the 30°
central tangent screen (1.0 log fL). Under these conditions, the normal (OD) and temporally lesioned (OS)
eyes (A-3) produced normal light-adapted ERGs of
virtually identical amplitude, although no cortical signal was detected with stimulation of OS. When the
stimulus flash was presented in the high luminance
surround, with the central 30° screen masked, no ERGs
or cortical responses were detectable on stimulation of
either eye, whereas normal ERGs and cortical responses
(both eyes) were recorded under this condition with
reduction of the surround luminance to 1.0 log fL. The
surround adaptation, therefore, effectively eliminated
ERG production by the peripheral retina indicating
that the ERGs recorded (OS) on central stimulation
were from the same central area which was incapable
of PERR production in response to high frequency
and/or low contrast stimuli.
At the end of the 210-day period, and following tissue
preparation, the plastic-embedded, flat-mounted retinas were sampled in the same way. Two or three samples for examination were taken from three sites: perifovea, peripapilla (site of lesions), and intermediate
(midway) between these two points. The samples were
3-mm diameter retinal buttons removed by motorized
cutter from the plastic flatmount. Adjacent buttons
were immediately superior to, centered on, and inferior
to the horizontal meridian at each site. Figure 5, from
the intermediate sample site of the temporal-lesion retina, shows a section from the most superior portion of
the superior button, one from the middle of the central
button and one from the most inferior portion of the
inferior button. A few scattered cell bodies (G) appear
to indicate ganglion cells in the most distal portions of
the inner layer in the superior and inferior section.
Elsewhere there are no ganglion cells or fiber layer.
739
0.08-0.6 c/deg 9
1.0-2.0 c/dog Q
0L
g 1.6
|
c
1.2
2 0.8
a.
o
o
\
\
/
\
UJ 0.4
>
<
X
D
°
-10
0
5
D-i
C
12 18
40
.....4.
60
90 120
210
DAYS AFTER SURGERY
Fig. 4. Summary of representative electrical signal changes at the
eye and cortex for 52% contrast, 2.5 Hz counterphased grating stimuli.
Results are pooled for small (^ 1.0 cpd) and large (< 1.0 cpd) stimuli.
Component analysis is as in Figure 2A. Abcissae values are means
of ratios (lesion eye/control eye) of signal amplitudes produced 6y
stimulation over a 220-day period. All points are referred to day - 1 0 ,
where the means are set to one. Some representative SD limit marks
are shown. Monkey B-46, temporal lesion.
'Cystic-like' (C) spaces occur only in the inner-nuclear
layer of the central sections. These are presumed to
have been the sites occupied by the large displaced
ganglion cells described in the rhesus retina.24 Outer
retinal layers, receptor outer segments, and pigment
epithelium appear normal.
The only transretinal pathology occurred at the sites
of the microdiathermy burns (not illustrated) where
gliosis and fibrosis extended through the retinal thickness to the pigment epithelium. The radii of the burns
extended 300-600 ^m, where normal outer retinal layers began. Samples from the same positions in the nasal-lesion retina (lesion-control) showed no evidence
Downloaded From: http://arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933130/ on 07/31/2017
740
INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / May 1986
SUPERIOR
Vol. 27
is extensive atrophy of the B-46 nerve and consequentially a lower overall area.
Discussion
CENTRAL
INFERIOR
140 urn
Fig. 5. Retinal sections taken at the extremes and center of a 9mm vertical segment {3 buttons) centered between the disc and fovea
of monkey B-46, lesioned (temporal disc edge) eye. G and C indicate
ganglion cell and "cystic-like" space, respectively. Toluidine blue
stain, Epon-Araldite embedment, 2-fim section.
of abnormality. There was no evidence of pathological
detachment in any tissue. Foveal sections from B-46
(Temporal lesion, A) and B-860 (Nasal lesion, B) are
shown in Figure 6. A full complement of receptors,
outer nuclear and inner nuclear layer cells, may be
seen. The ganglion cell layer is 5-6 cells thick in the
region immediately adjacent to the foveal pit of B-860
(B). In B-46 (A) the prominent parafoveal mound is
missing with preservation of the inner plexiform layer,
while a few scattered cells remain in the ganglion cell
layer. Results at the intermediate sample sites were
consistent with those in Figure 6 (central site). Samples
from B-860 were normal whereas the fiber and ganglion
cell layers in B-46 were absent.
Optic nerve sections confirm the retinal inner layer
histopathology. Figure 7A shows a relatively local area
(region N-N, 20-30%) with the fasciculated, normal
optic nerve axons remaining in the section from B-46
(Temporal lesion). Elsewhere there are no fasciculated
axons but moderately active gliosis is present. The results from B-860 (Fig. 7B) are almost the inverse. The
area of gliosis and extensive destruction covers approximately 20-30% of the sample. The margins of the
abnormal areas fade gradually into the regions where
the normal, fasciculated axons appear. Direct spatial
comparison of the two nerve samples is difficult. There
In these experiments we have created two types of
local retinal lesions, each directly involving approximately the same amount of retinal tissue but very different areas of influence and degrees of inner-retinal
layer involvement25; we then evaluated PERRs under
each condition. We chose the rhesus monkey because
of its similarity to human retinal structure, its fovea,
its cone distribution, and the extensive elaboration of
macular ganglion cell and fiber layer.26 The recent international activity in research on the PERR has been
reviewed by Lawwill.27 Previously PERRs have only
been recorded from humans, dogs, cats, and pigeons.8'2122'28 Current fundamental and clinical interest
in the inner and macular retina suggest the need for a
model offering, in addition to the necessary structures,
a good signal/noise ratio and spatial frequency sensitivity characteristics.
Our results from the rhesus eye establish that its
PERR has electrical properties similar to those inferred
from human data from this laboratory.11 Rhesus wide
band signals contain complexities and small components (cf, Fig. 1) absent, with two notable exceptions, 1729 in published PERR waveforms, because of
upper frequency constraints usually imposed to improve signal-to-noise ratios. Sharp band limitations can
be very useful to isolate electrical components. But information may be lost. Waveform distortion and phase
shifts can be introduced by active filters.30 We find that
the rhesus eye offers excellent signal-to-noise ratios relative to human, without extensive filtering, as well as
excellent spatial frequency sensitivity (see Figs. 1, 2).
These species characteristics, when added to those
noted above, identify special advantages for those interested in the tissue aspects of inner retinal physiology.
Some early reports on the PERR do not show strong
tuning.31"34 Recently however, evidence of spatial processing similar to Figure 2 has become more common.17'35"37 Dawson22 describes contrast and rate-dependent tuning of inhibition of the PERR by windmilllike objects moving within the grating-stimulated field.
Ringo et al38 did not find tuning of a pattern-elicited
retinal signal in cat. Schuurmans and Berninger,39
conversely, found clear tuning functions in the perfused
cat eye and a "striking similarity" of pattern ERGs and
optic nerve responses (from 42 eyes). This deserves
special notice since the traditional relationship between
the flash stimulated ERG and optic nerve response has
been very poor.40
The human psychophysical contrast sensitivity
maximum, at modest adaptation levels C — 111 candles/
Downloaded From: http://arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933130/ on 07/31/2017
No. 5
RETINAL RESPONSES AFTER NERVE FIBER LESIONS/Dowson er ol.
B
Fig. 6. Sequential sections of the rhesus parafovea.
A, 210 days after temporal
160° microdiathermy discedge lesions. B, nasal lesions
otherwise as A, one approximately mid-fovea, six approximately inferior rim.
Sections are approximately 35 ftm apart and 2 pm
thick. Interference contrast
microscopy, toluidine blue
stain, Epon-Araldite embedment.
A
tjfl0B*WV'
^
*
k >&
200/Um
Downloaded From: http://arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933130/ on 07/31/2017
•M.
742
INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / May 1986
imm
Vol. 27
B
Fig. 7. Rhesus optic nerve
sections taken approximately
4 mm from the globe. A, two
hundred ten days after 160°
lesions of the inner retina at
the temporal margin of the
disc. B, two hundred ten days
after nasal lesions (otherwise
as A). Regions where normal
(N) and atrophic tissue merge
are indicated. Interference
contrast microscopy, 3-/im
sections, Epon-Araldite
embedment, toluidine blue
stain, A and B, equal magnification.
ma
m2), is reported in the range 0.4-0.6 log cycles/degree
(2 to 4 cpd).41 Our rhesus tuning functions peak at
about 0.5 cpd for the cortical signal and about 0.35
cpd for the PERR (see Fig. 2). These peaks are somewhat lower than those reported for human PERRs (0.610 cpd) 1337 ' 42 under a variety of stimulation methods.
The tuning peak can shift toward lower spatial frequencies as average integrated luminance is reduced.29
Jacobs43 demonstrated variation in spatial frequency
sensitivity in the monkey eye with changes in average
luminance. These factors suggest that our relatively low
spatial frequency peak is due to the moderate adaptation level.
All of our temporal-lesion animals showed initial
retinal and cortical hyperresponsiveness followed by
decline beginning the second week post-lesion. Arden
et al44 found a similar hyperresponse-depression sequence in early active human optic neuritis. Our results
further indicate different decay rates for the major
PERR components, differential effects of stimulus spatial frequency, and contrast on decay and differences
in similarity of PI and N1 decays to that of the cortical
response. The observation (Fig. 4) that for low frequency/high contrast stimuli, the cortical response decays more rapidly and completely than either PERR
component suggests that PERRs to low frequency/high
contrast stimuli are less ganglion cell dependent than
their high frequency/low contrast counterparts. Of
course, not all ganglion cells were eliminated by our
lesions, as indicated by the retinal and optic nerve histopathology. Surviving cells/axons in the lesioned area
may account for a significant portion of the residual
function apparent when very low spatial frequency,
high contrast stimuli are presented. Differences in PI
and Nl post-lesion behavior and spatial frequency
tuning may suggest different generators, but current
understanding of this complex signal is inadequate to
support more than speculation on this question or on
the significance of the interesting small components
seen in the wide band response.
The delayed effects of the lesions, even in transmission to the visual cortex, are surprising. This suggests
a fiber population damaged but not immediately destroyed and that axon transmission failure may be a
graded process that proceeds for at least 2 wk. This
agrees well with the temporal decay of the high spatial
frequency PI component. Sommer45 et al provide
clinical evidence for up to 5-yr delays in field loss in
the presence of nerve fiber layer defects. James 46 cut
the optic nerve in rabbits intraorbitally and found that
most ganglion cells showed some chromatolysis at ten
days. By the twentieth day only a few nuclear and cell
body fragments could be found in the ganglion cell
layer. By 30-60 days very large ganglion cells could
only be seen rarely. These had disappeared by 120 days.
Radius and Anderson47 reported that the ganglion cells
and nerve fiber layer did not change noticeably during
the first 2 wk after photocoagulation of the peripapillary
axons of the monkey. By the end of 3 wk there was
some decrease in ganglion cell counts, especially in the
perimacula. After 4 wk the ganglion cell population
had decreased significantly, and by 6 wk the degeneration had reached a plateau and was highly significant
statistically. This time course appears to be consistent
with the decay of the PERRs and cortical responses in
our data. Overall, the results support the notion that
ganglion cell body function may not cease immediately
following damage to its axon, with the first effects appearing between 2-4 wk.
Microscopic examination of the lesioned retinas
disclosed no multi-layer pathology involving outer retina or pigment epithelium, except immediately at the
microdiathermy burn sites. Except for effects directly
attributable to retrograde degeneration of the fiber layer
Downloaded From: http://arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933130/ on 07/31/2017
No. 5
RETINAL RESPONSES AFTER NERVE FIBER LESIONS/Dawson er al.
and ganglion cells, all retina other than the burn sites
appeared completely normal after 210 days. This evidence of outer retinal integrity in the lesioned eye was
confirmed by the consistently normal flash (local and
diffuse) ERGs in the lesioned eyes. Neither ERGs nor
serial color fundus photographs gave evidence of ischemia, although the photographs showed large arcuate
areas of "highlight loss" interpreted clinically as fiber
bundle defects. Because of the optimal pigmentation
of the rhesus fundus no special fundus photographic
techniques48 were needed.
Interesting 'cystic-like' spaces, about the size of one
or two large cell bodies, found in the inner nuclear
layer in the central retina of temporal-lesion eyes probably indicate degeneration of displaced ganglion cells.24
We did not find these spaces in nasal-lesion retinas.
The only detectable effect of these lesions on outer retina was some distortion (not cell pathology) of a 100/im region of the foveal outer nuclear layer (cf, Fig.
6A). Presumably, this was a mechanical effect of redistribution of inner plexiform tissue, Henle fibers and/
or photoreceptor cell bodies subsequent to degeneration
of inner retinal tissue (ganglion cells/fibers), although
intermittant processing artifact cannot be ruled out.
These observations clearly establish that our temporal
peripapillary diathermy burns produced extensive fiber
layer defects and elimination of the ganglion cell layer
over the majority of the central retina, including the
macula and fovea, with no significant effect on outer
retinal structure. This conclusion is consistent with our
understanding of fiber layer organization.49
The inner layer pathology may indirectly affect the
fundus appearance. As the lesions developed our
ophthalmoscopic data indicated progressive loss of foveal reflexes in the temporal-lesion retinas. The primate
medial parafoveal mound, prominent in the nasal-lesion retinas, (cf., Fig. 6B), appears to be composed
mainly of Henle fiber elaboration and, to some extent,
inner nuclear layer tissue. According to Fine and Yanoff,50 the obliquely oriented Henle fibers (the foveal
cone receptor axons) and the plexiform layer just distal
to the primate inner nuclear layer form the outer plexiform layer. They do not describe a foveal mound in
their rhesus material. However, Van Buren26 shows
several examples of a medial mound in his human tissue (A57-261, A56-1, NA-25-51) but does not discuss
it. The atrophy of the mound, with loss of its shadowing
effect and loss of most of the foveal pit undoubtedly
accounts for the loss of the foveal "reflex" during the
post-temporal-lesion (+60 days) period.
In spite of the size and completeness of the inner
retinal temporal lesions, the local ERG recordings establish the ability of the rhesus outer retina to produce
normal signals in areas where the inner retina is
atrophic. This is entirely consistent with current ERG
theory (as discussed below) and suggests that major
743
damage to the glial network is absent. Additionally,
important relationships are established between our
local lesions, PERRs, ERGs, and cortical signals initiated in the primate central retina.
The prime origin of the term electroretinogram is
obscure. We have not found it used before the 1924
paper "Das Elektroretinogramm" by Kahn and Lowenstein.51 Magitot52 referred to the retinas' "courant
d'action" in 1922. And in America, Sheard and
McPeek53 applied the term "electrical response of the
eye" in 1919, which was adopted by Hartline 54 in 1925
and as late as 1930 by Chaffee and Sutliff.55 Each of
these authors clearly identified their signal with the retinal response to diffuse illumination. Until the first use
of patterned stimuli in Riggs' laboratory,8 there had
been no deviation from this convention. It is not yet
established that the signals recorded by Riggs and discussed in this paper have any close relation to the classical "ERG." Indeed, the gathering evidence presented
by authors using the "ERG" terminology to describe
responses to patterns seems to argue against the relationship. The high precision of current vision research
calls for a brief examination of this practice.
No alterations of the ERG have been found in patients with profound disease of the inner retina and
optic nerve. Following pioneer findings of Karpe 2 and
many others, subsequent influential clinical volumes
by Krill,56 Deutman, 57 and recently Fishman 58 do not
discuss inner retinal implications of the ERG a- and
b-waves, for it is well established that there are none.
In the last 10 years there have been a few papers of
clinical interest where ERG and inner retina are related. These involve the small ERG components called
"minor components," "oscillations" or "fast potentials".59"61 In stark contrast, during the last 3 years
there have been numerous papers relating the patternevoked retinal response to diseases or abnormalities of
the inner retina. A small sample includes amblyopia, 4462 optic nerve disease,63 glaucoma,14 multiple
sclerosis,64 and traumatic optic atrophy.''
The basic science background that has so strongly
fused the conceptual linkage between the ERG and the
outer retina has been described in full detail in chapters
by Riggs, Tomita, and Granit in a recent volume.3
Riggs' contribution to this volume was written in 1982;
at that time, he did not discriminate between the ERG
and the pattern-evoked retinal signal (which was first
recorded in his laboratory). Another major contributor
to the firm conceptual fusion between the "ERG" and
the outer retina was provided in the now classic paper
on potential distributions by K. T. Brown.65 Together
with the elegant cellular proofs offered by Werblin and
Dowling5 and by Miller and Dowling6 there seemed
little remaining to spark controversy about the cellular
origins of the ERG. But, Tomita and Yanagida66 provide more detail on the identification of additional
Downloaded From: http://arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933130/ on 07/31/2017
744
INVESTIGATIVE OPHTHALMOLOGY b VISUAL SCIENCE / May 1986
current sinks provided by Muller cells and stimulated
by potassium discharge from nearby retinal neurons.
Intraretinal techniques similar to those used by
Brown65 have been employed by Holden and Vaegan19
and by the group in Amsterdam.* The expected distribution for pattern-elicited potentials at the inner retina has not been found. Indeed, they report signals
more consistent with the traditional region of the ERG
b-wave; that is, approximately mid-retina. This seems
to directly contradict the larger body of accumulated
indirect data that binds the PERR to the inner retina.
But, it is not mandatory that all retinal action potentials
arise directly from active neurones. And there is a very
recent proof that some potential distributions (m-wave)
may change depth profile depending on stimulus properties.67
Tomita and Yanagida66 describe a large proximal
retinal sink for potassium currents, in addition to a
smaller distal site. Publications by Newman 68 and
Newman, Frambach, and Odette 69 provide impressive
evidence for the preeminence of the Muller cell endfoot
for clearing of potassium. K + injection there can cause
large potentials in mid-retina distal to the site of injection.66 Together these form a promising potential
mechanism for the involvement of the Muller cells in
the consequences of processing of pattern information
by the neurons of the inner retina.
Upon this basis, it is possible to develop a testable
hypothesis that may be used to reconcile the distribution of the pattern-elicited potential in mid-retina
with the inner retinal cell, pattern-signal interdependence. It is consistent to propose a mid-retinal potential
change that is driven indirectly by inner-retinal neurons. In a process analogous to the interaction between
the photoreceptors and Muller fiber elements of the
outer retina, ganglion cells (and/or other inner retinal
elements) may depolarize Muller cells indirectly by
liberating potassium near their endfeet. This mechanism does not eliminate a possibility of direct contribution of the inner-retinal neurons to the "PERR" but
encourages the consideration of a composite potential
of summed neural and glial activity. In this way the
PERR may share mechanisms with the ERG but arise
from different origin(s). The PERR may be a complex
sum of these (glial and neuronal) processes, whose regional balance and excitability is dependent upon both
species and stimulus characteristics. A combination
hypothesis involving both activated inner-retinal neuron, K + extrusion, and secondary Muller cell potentials
would serve to explain most of the existing data relating
to PERR origin(s).
The disease sensitivities, stimulus characteristics and
* Riemslag F and Heynen H: Depth profile of pattern LERG in
Macaca, (in preparation).
Vol. 27
inner retinal cell viability associated with pattern-elicited retinal responses are very different from those traditionally associated with the ERG. The term used to
describe these interesting action potentials is not important. But continued, indiscriminant use of the term
"ERG" to describe pattern-evoked potentials may
confuse students of retinal physiology for decades.
Key words: PERR, lesions, Macaca mulatta, retina, ganglion
cells, macula, optic nerve, pattern, ERG, K+
References
1. Granit R: Receptors and Sensory Perception. New Haven, Yale
University Press, 1955, pp. 151-190.
2. Karpe G: The basis of clinical electroretinography. Acta
Ophthalmol24(Suppl).i418, 1945.
3. Dawson WW and Enoch JM, editors: Foundations of Sensory
Science. Berlin, Springer-Verlag, 1984, 575 pp.
4. Brown K and Wiesel T: Localization of electroretinogram components in the intact cat eye. J Physiol (Lond) 158:257, 1961.
5. Werblin F and Dowling J: Organization of the retina of the mudpuppy, Necturus maculosus. II. Intracellular recording. J Neurophysiol 32:339, 1969.
6. Miller R and Dowling J: Intracellular responses of the Muller
(glial) cells of the mudpuppy retina: their relation to the b-wave
of the electroretinogram. J Neurophysiol 33:323, 1970.
7. Kaneko A: Physiological and morphological identification of
horizontal, bipolar and amacrine cells in goldfish retina. J Physiol
(Lond) 207:623, 1970.
8. Riggs L, Johnson E, and Shick A: Electrical responses of the
human eye to moving stimulus patterns. Science 144:567, 1964.
9. Lawwill T, Walker C, and Crockett S: The photopic and scotopic
bar-pattern ERG contributions of the central and peripheral retina. Doc Ophthalmol Proc Series 13:287, 1977.
10. Maffei L and Fiorentini A: Electroretinographic responses to alternating gratings before and after section of the optic nerve.
Science 211:953, 1981.
11. Dawson WW, Maida TM, and Rubin ML: Human patternevoked retinal responses are altered by optic atrophy. Invest
Ophthalmol Vis Sci 22:796, 1982.
12. Fiorentini A, Maffei L, Pirchio M, Spinelli D, and Porciatti V:
The ERG in response to alternating gratings in patients with
disease of the peripheral visual pathway. Invest Ophthalmol Vis
Sci 21:490, 1981.
13. Hess R and Baker C Jr: Human pattern-evoked electroretinogram. J Neurophysiol 51:939, 1984.
14. Papst N, Bopp M, and Schnaudigel O: Pattern electroretinogram
and visually evoked cortical potentials in glaucoma. Graefes Arch
Clin Exp Ophthalmol 222:29, 1984.
15. May J, Ralston J, Reed B, and Van Dyk H: Loss in patternelicited electroretinograms in optic nerve dysfunction. Am J
Ophthalmol 93:418, 1982.
16. Groneberg A and Teping C: Topodiagnostik von Sehstonungen
durch ableitung retinaler und kortikaler auworten auf UmkehnKontrastnuster. Ber Dtsch Ophthalmol Ges 77:409, 1980.
17. Korth M: Pattern evoked responses and luminance evoked responses in the human electroretinogram. J Physiol (Lond) 337:
451, 1983.
18. Sherman J and Richardson V: What is the origin of the pattern
reversal electroretinogram in human? ARVO Abstracts. Invest
Ophthalmol Vis Sci 22(Suppl):138, 1982.
19. Holden A and Vaegan: Intraretinal responses to contrast reversal
in the pigeon. J Physiol 324:69, 1982.
20. Riemslag F, Ringo J, Spekreijse H, and Verduyn Luvel H: The
Downloaded From: http://arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933130/ on 07/31/2017
No. 5
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
RETINAL RESPONSES AFTER NERVE FIBER LESIONS/Dowson er ol.
distinction between luminance and spatial contrast components
in the pattern ERG. Doc Ophthalmol Proc Series 37:255, 1983.
Odom J, Bromberg N, and Dawson W: Canine visual acuity:
retinal and cortical field potentials evoked by pattern stimulation.
Am J Physiol 245:R637, 1983.
Dawson W: Rate-dependent inhibition of pattern-evoked retinal
responses in the pigeon. Invest Ophthalmol Vis Sci 25:108, 1984.
Mae Y and Matsui M: Ophthalmoscopic model of nerve fiber
bundle defects in monkey eyes produced by intraocular retinal
coagulation. Jpn J Ophthalmol 26:274, 1982.
Bunt A and Minckler D: Displaced ganglion cells in the retina
of the monkey. Invest Ophthalmol Vis Sci 16:95, 1977.
Dawson WW and Maida TM: Relations between the human
retinal cone and ganglion cell distributions. Ophthalmologica
188:216, 1984.
Van Buren J: The Retinal Ganglion Cell Layer. Springfield, Illinois, Charles C Thomas, 1963, 143 pp.
Lawwill T: The bar pattern electroretinogram. Doc Ophthalmol
Proc Series 40:1, 1984.
Hollander H, Bisti S, MafFei L, and Hebel R: Electroretinographic
responses and retrograde changes of retinal morphology after
intracranial optic nerve section. A quantitative analysis in the
cat. Exp Brain Res 55:483, 1984.
Korth M and Rix R: Effect of stimulus intensity and contrast
on the pattern ERG. Ophthalmic Res 16:60, 1984.
Dawson WW and Doddington HW: Phase distortion of biological
signals: extraction of signal from noise without phase error. Electroencephlogr Clin Neurophysiol 34:207, 1973.
Armington J, Corwin T, and Marsetta R: Simultaneously recorded retinal and cortical responses to patterned stimuli. J Opt
Soc Am 61:1514, 1971.
Groneberg A: Simultaneously recorded retinal and cortical potentials elicited by checkerboard stimuli. Doc Ophthalmol Proc
Series 23:255, 1980.
Korth M: Human fast potentials and the spatial properties of a
visual stimulus. Vision Res 21:627, 1981.
Trick G and Wintermeyer D: Spatial and temporal frequency
tuning of pattern-reversal retinal potentials. Invest Ophthalmol
Vis Sci 23:774, 1982.
Korth M, Rix R, and Sembritzki O: Spatial contrast transfer
functions of the pattern-evoked electroretinogram. Invest
Ophthalmol Vis Sci 26:303, 1985.
Sokol S, Jones K, and Nadler D: Comparison of the spatial response properties of the human retina and cortex as measured
by simultaneously recorded pattern ERGs and VEPs. Vision Res
23:723, 1983.
Odom J, Dawson W, Romano P, and Maida T: Human patternevoked retinal response (PERR): spatial tuning and development.
Doc Ophthalmol Proc Series 37:265, 1983.
Ringo J, Van Dijk B, and Spekreijse H: Pattern ERG in the cat.
Vision Res 24:859, 1984.
Schuurmans R and Berninger T: Pattern reversal responses in
man and cat: a comparison. Ophthalmic Res 16:67, 1984.
Schmeisser E and Dawson W: Slow wave relationships in the
visual system of Necturus maculosus. Comp Biochem Physiol
67A:605, 1980.
Higgins K, Jaffe M, Coletta N, Caruso R, and de Monasterio F:
Spatial contrast sensitivity. Arch Ophthalmol 102:1035, 1984.
Odom J and Norcia A: Synchronous retinal and cortical potentials: spatial and temporal characteristics. Doc Ophthalmol Proc
Series (in press).
Jacobs G: Visual capacities of the owl monkey (Aotes trivirgatus).
II. Spatial contrast sensitivity. Vision Res 17:821, 1977.
Arden G, Vaegan, and Hogg C: Clinical and experimental evidence that the pattern electroretinogram (PERG) is generated
in more proximal retinal layers than the focal electroretinogram
(FERG). Ann NY Acad Sci 388:580, 1982.
745
45. Sommer A, Miller N, Pollack I, and Maumeuee A: The nerve
fiber layer in the diagnosis of glaucoma. Arch Ophthalmol 95:
2145, 1977.
46. James G: Degeneration of ganglion cell following axonal injury.
Arch Ophthalmol 9:338, 1933.
47. Radius R and Anderson D: Retinal ganglion cell degeneration
in experimental optic atrophy. Am J Ophthalmol 86:673, 1978.
48. Quigley H, Miller BN, and George T: Clinical evaluation of nerve
fiber layer atrophy as an indicator of glaucomatous optic nerve
damage. Arch Ophthalmol 98:1564, 1980.
49. Ogden T: Nerve fiber layer of the macaque retina: retinotopic
organization. Invest Ophthalmol Vis Sci 24:85, 1983.
50. Fine B and Yanoff M: Ocular Histology: A Text and Atlas. New
York, Harper and Row, 1972, pp. 93-107.
51. Kahn R and Lowenstein A; Das elektroretinogramm. Graefes
Arch Clin Exp Ophthalmol 114:304, 1924.
52. Magitot A: Sur le courant retinien. Annales d'Oculistique 159:
241, 1922.
53. Sheard C and McPeek C: On the electrical responses of the eye
to stimulation by light of various wave-lengths. Am J Physiol
48:45, 1919.
54. Hartline H: The electrical response to illumination of the eye in
intact animals including the human subject and in decerebrate
preparations. Am J Physiol 73:600, 1925.
55. Chaffee E and Sutcliffe E: The differences in electrical response
of the retina of the frog and horned toad according to the position
of the electrodes. Am J Physiol 95:250, 1930.
56. Krill A: Krill's Hereditary Retinal and Choroidal Diseases. Vol
2, Clinical Characteristics. New York, Harper and Row, 1977,
1339 pp.
57. Deutman A: The Hereditary Dystrophies of the Posterior Pole
of the Eye. Assen, The Netherlands, VaivGorcum, 1971, 484
pp.
58. Fishman G: The Electroretinogram and Electro-oculogram in
Retinal and Choroidal Disease. Rochester, MN, Am Acad
Ophthalmol Otolaryngol, 1975, 44 pp.
59. Simonsen S: The value of the oscillitory potential in selecting
juvenile diabetics at risk of developing proliferative retinopathy.
Acta Ophthalmol 58:865, 1980.
60. Adams CK and Dawson WW: Fast retinal potential luminosity
function. Vision Res 11:1135, 1971.
61. Dawson WW and Stewart HL: Signals within the electroretinogram. Vision Res 8:1265, 1968.
62. Persson H and Wanger P: Pattern-reversal electroretinograms in
squint amblyopia, artificial anisometropia and stimulated eccentric fixation. Acta Ophthalmol 60:123, 1982.
63. Porciatti V and von Berger G: Retinal and cortical evoked potentials in response to alternating gratings in the diagnosis of
optic nerve disease. In Evoked Potentials II, Nodar R and Barber
C, editors. Worbun, MA, Buttersworth Press (in press).
64. Bobak P, Bodis-Wollner I, Harnois C, Maffei L, Mylin L, Podos
M, and Thornton J: Pattern electroretinograms and visual-evoked
potentials in glaucoma and multiple sclerosis. Am J Ophthalmol
96:72, 1983.
65. Brown K: The electroretinogram: its components and their origins. Vision Res 8:633, 1968.
66. Tomita T and Yanigida T: Origins of the ERG waves. Vision
Res 21:1703, 1981.
67. Sieving P and Steinberg R: M-wave in cat—proximal retinal
contribution to PERG and flash ERG. ARVO Abstracts. Invest
Ophthalmol Vis Sci 26(Suppl):112, 1985.
68. Newman E: Regional specialization of retinal glial cell membrane.
Nature 309:155, 1984.
69. Newman E, Frambach D, and Odette D: Control of extracellular
potassium levels by retinal glial cell K+ siphoning. Science 225:
1174, 1984.
Downloaded From: http://arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933130/ on 07/31/2017