spatial cone activity distribution in diseases of the …vision research 38 (1998) 3817–3828...

Vision Research 38 (1998) 3817–3828

Spatial cone activity distribution in diseases of the posterior poledetermined by multifocal electroretinography

Ulf Kretschmann *, Mathias Seeliger, Klaus Ruether, Tomoaki Usui, Eberhart ZrennerUni6ersity Eye Hospital, Department of Pathophysiology of Vision and Neuro-Ophthalmology, Schleichstr. 12–16, D-72076 Tubingen, Germany

Received 5 August 1997; received in revised form 18 December 1997

Abstract

Thirty patients with a reduced central vision due to diseases of the posterior pole were examined with the VERIS systemdeveloped by Sutter and Tran (Vis Res 1992;32:433–446) to characterize the topography of electroretinographic (ERG) changesin comparison to the results in 30 normal volunteers. Diagnoses included Stargardt’s macular dystrophy (SMD, n=10),age-related macular degeneration (AMD, n=5), cone dystrophy (CD, n=5), central retinal vein occlusion (CRVO, n=5), andautosomal dominant optic atrophy (ADOA, n=5). The 61 local responses obtained from each subject were grouped byeccentricity to form five concentric rings. The foveal ERG, originating from a central area of 2° radius, was non-recordable ormarkedly diminished in all patients except those with optic atrophy, where amplitudes were found to be in the normal range. Inpatients with advanced stages of SMD, functional defects were larger and involved more peripheral areas than in patients withearly stages of SMD or with AMD. A reduction of response amplitude even in the most peripheral ring (17–30.5° eccentricity)was found in cone dystrophies and—moderately—in patients with advanced SMD and central retinal vein occlusion only.Prolonged implicit times were found in all but the patients in early stages of SMD and they were maximal in patients with CRVO.This study shows that the multifocal ERG (MFERG) can contribute to differential diagnosis of retinal diseases of the posteriorpole especially in cases with a normal photopic Ganzfeld ERG. © 1998 Elsevier Science Ltd. All rights reserved.

Keywords: Multifocal electroretinogram; Maculopathy; Cone dystrophy; Central vein occlusion; Optic atrophy

1. Introduction

Multifocal electroretinography (MFERG) based onthe m-sequence stimulation technique [1–3] allowsquick simultaneous recording of many local elec-troretinograms from the posterior pole. The responsedensity distribution of the first order kernel computedfrom the data is in good accordance with the conedensity distribution found by histological methods [4,5].Recently, a close relationship of signal generation andshape between the first order kernel response of themultifocal ERG and the photopic Ganzfeld-ERG wasdemonstrated [6].

The multifocal ERG is thus a powerful tool in pa-tients with an affected cone system, especially whendefects are localized and therefore not detectable withGanzfeld-electroretinography. First reports [7–9] haveshown the possibility to map cone function in patientswith regional defects of the outer retinal layer. In anormative value study, Parks et al. [10] have further-more demonstrated that the retest-variability of themethod is comparable to that of standard electrophysi-ological methods.

The purpose of the present study was to investigateand describe differences in regional cone dysfunctionbetween groups of patients with impaired vision due todiseases affecting predominantly the posterior pole. Toreveal which MFERG components are most susceptibleto pathological alteration, waveforms of individual nor-mal volunteers and patients are presented. Quantitativeanalysis was performed by measuring peak responsedensities and timing.

* Corresponding author. Present address: Eye Hospital, MedicalSchool Hannover, Carl-Neuberg-Str. 1, 30625 Hannover, Germany.Tel.: +49 511 5322473; fax: +49 511 5323050.

0042-6989/98/$19.00 © 1998 Elsevier Science Ltd. All rights reserved.

PII: S0042-6989(98)00071-6

U. Kretschmann et al. / Vision Research 38 (1998) 3817–38283818

2. Materials and methods

2.1. Subjects

Six groups of five patients each were formed on thebasis of clinically established diagnoses, yielding a totalstudy population of 30 patients. Group diagnoses in-cluded early Stargardt’s macular dystrophy (SMD),advanced SMD, age-related macular degeneration(AMD), cone dystrophy (CD), central retinal vein oc-clusion (CRVO), and autosomal dominant optic atro-phy (ADOA).

The diagnosis of SMD was based on history, sym-metric bilateral involvement, the typical alterations ofthe pigment epithelium layer (assessed by fluorescenceangiography if necessary), by visual field, and byGanzfeld-electroretinography according to the ISCEV-standard. Patients in the early SMD group differedfrom those of the advanced SMD group by a centralscotoma less than 10° in diameter. Visual fields weretested with a Tubingen automated perimeter (Oculus,Germany) using program No. 1 (30° visual field, 190test points, suprathreshold strategy, stimulus size 10%,maximal stimulus intensity 1000 cd/m2, background in-tensity 10 cd/m2). Patients in the AMD-group all hadthe atrophic form of the disease. In none of the patientswere hemorrhages or neovacularizations present.

In Table 1, information on age (in years), sex (m=male, f= female), and the visual acuity (VA) of theanalysed eye is given for the individual patients. If thelesions were symmetrical in both eyes the left eye wasanalyzed; in asymmetrical cases the more affected onewas investigated. If the right eye was analyzed it isindicated with ‘OD’.

Thirty normal volunteers (age 22–58, median 31)were examined with the multifocal ERG under thesame conditions, to serve as a control group. They hadfull visual acuity, refractive errors below 96.0 diopters, and no history of eye disease or otherrelevant disorders.

Informed consent was obtained from patients andcontrols after description of the procedures. Tenets ofthe declaration of Helsinki were followed and institu-tional human experimentation committee approvalobtained.

2.2. Stimulation

The stimulus, consisting of 61 hexagons covering avisual field of 30°, was presented on a monitor with aframe rate of 75 Hz at a distance of 28 cm from thesubject’s eye. The radius of the central hexagonal ele-ment was 2°. Since element areas are scaled by eccen-tricity, the most eccentric areas are 4.7 times larger thanthe central area (Fig. 1). Each element, which waseither black or white (93% contrast, mean luminance

51.8 cd/m2), changed following a binary m-sequenceindependent from other elements. In a time of 3.63 min,16383 different stimulus permutations were presented.A red central fixation point of 2 mm in diameter wasused.

2.3. Electrophysiological recordings

Both eyes were dilated with tropicamide (0.5%) andphenylephrine (5%) and the refractive errors corrected.ERG responses were recorded by means of DTL fiberelectrodes (UniMed Electrode Supplies, England),which were positioned on the conjunctiva directly be-neath the cornea and attached with its two ends at thelateral and nasal canthus. The reference and groundskin electrodes (gold cup electrodes) were attached tothe ipsilateral temple and forehead, respectively.

The signal was amplified (×200000) and bandpass-filtered (10–100 Hz; Grass-amplifier, model 12, Quincy,USA).

Each recording session was subdivided into 20recording segments of approx. 11 s duration, duringwhich the subjects were not allowed to talk or move.The total duration of a recording session was about5–7 min.

2.4. Data analysis

The VERIS Scientific software™, using a fast m-transform algorithm [1,3], was employed for the calcu-lation of the 61 local ERG responses from themeasured signal. Specifically, first order kernels wereused in this study because of their close correlation withthe function of the outer retina. The program for theMacintosh Quadra was provided by Dr Sutter, EDI,San Francisco.

Fig. 1. Geometry of the test field containing 61 stimulated areas. Forfurther analysis these areas were grouped into five concentric rings(1–5).

U. Kretschmann et al. / Vision Research 38 (1998) 3817–3828 3819

Tab

le1

For

the

indi

vidu

alpa

tien

tsof

the

six

dise

ase

grou

psag

e(i

nye

ars)

CR

VO

AD

OA

AM

DE

arly

SMD

CD

Adv

ance

dSM

D

Eye

VA

Age

Sex

Eye

VA

Age

Sex

Eye

Age

VA

Sex

Age

Sex

Eye

VA

Age

Sex

Eye

VA

Eye

VA

Age

Sex

OS

0.5

54m

OS

0.5

11m

OS

0.6

5829

ff

OS

0.2

24m

OS

0.5

0.8

OS

f24

OS

0.16

59m

OS

0.4

40m

OS

0.8

1234

mm

OS

0.2

33f

OS

0.5

OS

0.5

44m

OS

0.1

66m

OS

0.6

36m

OS

0.8

54m

m54

OS

0.25

13f

OS

0.4

21m

OS

0.8

OS

0.12

559

fO

S0.

718

mO

S0.

172

fO

D20

0.1

m20

mO

S0.

1O

S0.

127

fO

S0.

160

fO

S0.

125

23m

OS

0.5

69m

OD

f0.

0534

22m

OS

0.3

mO

S0.

829

Sex

(m=

mal

e,f=

fem

ale)

,an

dth

evi

sual

acui

ty(V

A)

ofth

ean

alys

edey

e(O

D=

righ

t,O

S=

left

)is

give

n.


Fig. 2. Multifocal electroretinograms recorded from the left eye of a normal volunteer, age 35. (a) Trace array with 61 focal ERGs. The responsemarked with a ring is smaller than the one marked with an arrow, because the stimulus area was projected partially onto the optic nerve head.(b) Three-dimensional plot of the response density measure with the scalar product method. (c) Foveal ERGs (ring 1 of Fig. 1) in response densityscaled form from five normal volunteers. These subjects represent rank 3, 14, 15, 16, 27 in foveal response density out of 30 normal subjects. Thefirst negative, the first positive, and the second negative deflections are marked as N1, P1, and N2. The asterix marks the subject used in a, b.(d) Averaged ERGs derived from an eccentricity of 17–30.5° (ring 5 in Fig. 1 from the same subjects as in c).

The resulting 61 ERG traces are shown in form of atrace array in Fig. 2a, each response positioned at itsapproximate place of origin in the visual field.

For comparison of diseases, five regional groups wereformed by averaging responses from elements of equaleccentricity (Fig. 1). Amplitudes and implicit times oftheir major components were then analyzed. Addition-ally, three-dimensional plots of the response densityusing the scalar product method [2] are shown in Fig.2b. The scalar product is a dot product of the localERG response with a normalized template. In thenormal volunteers and the patients with ADOA ringaverages served as templates, in the other patients anaverage of the entire field was used.

For statistical analysis a t-test was applied.

3. Results

3.1. Normal ERG topography

Fig. 2 illustrates the normal multifocal ERG results.In Fig. 2(a, and b) data from the left eye of one malevolunteer aged 35 are presented. Due to scaling of thestimulus (the area of the hexagons increases towardsperiphery to compensate for lower cone density, Fig. 1),the resulting 61 local ERG responses are almost equalin amplitude (Fig. 2a). In the region of the blind spot,which is located on the left side of the trace array inanalogy to a visual field plot, the evoked response issmaller than the corresponding one on the other side ofthe retina.


Table 2The mean amplitudes (response densities innV/deg2) and mean implicit times (ms) of the three major deflections N1, P1 and N2 (9S.D.) of all30 normal volunteers are presented. The values were measured from the baseline. The peak amplitude was measured between N1 and P1.

Ring 2 (1.8–7°) Ring 3 (5–13°)Ring 1 (0–2°) Ring 4 (11–22°) Ring 5 (17–30.5°)

S.D. Average S.D. Average S.D.Average Average S.D. Average S.D.

7.5 −9.9 2.5 −6.3 1.4 −4.7 1.2 −4.7N1 amplitude 1.4−20.91.9 16.7 1.1 16.0 0.917.8 16.5N1 implicit time 1.0 16.7 0.7

39.0P1 amplitude 8.4 17.3 3.6 10.5 2.3 7.3 1.6 7.8 1.932.3P1 implicit time 1.8 29.9 1.4 29.6 1.4 30.2 1.1 31.0 1.1

13.0 27.3 5.7 16.8 3.659.8 12.0Peak amplitude 2.7 12.3 3.28.3N2 amplitude −11.5−24.6 3.1 −7.3 1.4 −4.5 0.9 −4.1 1.83.5 45.9 1.5 44.1 1.449.5 44.4N2 implicit time 1.2 45.3 1.4

As expected from retinal anatomy, there was a contin-uous decrease in response density from the maximum atthe fovea towards the periphery. The scalar productmethod proposed by Sutter and Tran [2], using the scalarproduct as a measure of overall signal amplitude, is thebasis for the three-dimensional response density plot(Fig. 2b) provided by the VERIS software. Besides thedecrease in response density with eccentricity, localizedareas of low amplitude such as the blind spot are alsovisible.

In Fig. 2c the foveal ERG (ring 1), and in Fig. 2d theaveraged ERGs of 17–30.5° eccentricity (ring 5), for fivenormal volunteers are shown. The responses are densityscaled. The response density can be obtained by dividinglocal ERGs by the area they were elicited from. A typicalwaveform begins with a negative deflection (N1), fol-lowed by a positive deflection (P1), and a second nega-tive deflection (N2). In addition, there are two to threefurther components that are not as consistent as thethree ones mentioned, as they are often masked by noise.

Response densities and implicit times of the majorcomponents N1, P1, N2 of all 30 normal volunteers arepresented in Table 2 for the five eccentricity groups. Allcomponent densities decreased with eccentricity, al-though there was no further decrease from ring 4 (10.5–22°) to ring 5 (17–30.5°). The implicit times were highestin the fovea (ring 1) and lowest in ring 3 (5–13°) for allthree components.

3.2. Patients

Figs. 3–7 illustrate the multifocal ERG results in thesix patient groups. In the left upper panel (a) of eachfigure (3, 5–7) the trace array, and in the left lower panelthe response density 3D plot, of one typical patient pergroup are shown. The mean response densities in thecentral element (i.e. ring 1) are plotted in the right upperpanel (c) and in the most peripheral region (i.e. ring 5)are plotted in the right lower panel (d). Each plotconsists of five traces, corresponding to the five membersof that group (in the same order from top to bottom as

in Table 1), the asterix indicates data from the patientused in the left panel. Table 3 summarizes the data forcomparison of mean response densities and implicittimes between all six groups.

3.2.1. Early SMD (Fig. 3)Patients with juvenile (SMD) and age-related (AMD)

maculopathies exhibited patterns of extinguished ormarkedly diminished responses in the center, whichapproached normal values towards the periphery.

In the early SMD group (field defect of less than 10°in diameter), a deep functional defect was restricted tothe macula. The three-dimensional plot of the responsedensity had a crater-like appearance due to that centraldefect. Only in one patient with a visual acuity of 0.8 anda small relative macular scotoma, could a residual fovealresponse be recorded (Fig. 3c, top trace), while in theother four patients there was no foveal response (Fig. 3c,lower traces). Response densities in the peripheral ring 5(17–30.5°) were only insignificantly decreased. Nochange of implicit times could be detected (Table 3).

3.2.2. Ad6anced SMD (Fig. 4a and b)A significant macular response could not be recorded

in any of the patients with advanced SMD (field defectof more than 10° in diameter). In two cases, the responsedensity was markedly diminished up to ring 5 (Fig. 4b),but in contrast to cone dystrophy patients, all responsesin ring 5 could be clearly distinguished from noise.Compared to the normals, the response densities in theadvanced SMD patients were significantly lower in theentire test field. Implicit times, as far as they could bedetermined, were moderately but significantly prolonged(Table 3).

3.2.3. AMD (Fig. 4c and d)Similar to moderate SMD, the functional lesions in

AMD were also restricted to the macular area, withgood responses in the periphery. However, implicittimes, were prolonged (Table 3).


Tab

le3

Mea

npe

akam

plit

ude

(res

pons

ede

nsit

yin

nV/d

eg2)(9

S.D

.)an

dm

ean

impl

icit

tim

es(m

s)(9

S.D

.)of

the

peak

ampl

itud

eof

mul

tifo

cal

ER

Gs

reco

rded

from

pati

ents

and

norm

als

inth

efiv

eec

cent

rici

tygr

oups

are

pres

ente

d.

Rin

g4

(11

–22

°)R

ing

5(1

7–

30.5

°)R

ing

1(0

–2°

)R

ing

2(1

.8–

7°)

Rin

g3

(5–

13°)

PB

Ave

rage

S.D

.PB

Ave

rage

S.D

.PB

Ave

rage

S.D

.S.

D.

PB

Ave

rage

Ave

rage

PB

S.D

.

Ear

lySM

D0.

001

10.4

0.5

0.00

18.

91.

40.

052.

99.

913

.21.

8n.

s.P

eak

ampl

itud

e7.

80.

001

4.1

1.6

33.0

n.s.

30.7

1.7

n.s.

30.7

0.8

n.s.

31.3

1.3

n.s.

1.6

n.s.

30.3

P1

impl

icit

tim

eA

dvan

ced

SMD

0.00

14.

12.

10.

001

4.6

2.7

0.00

13.

97.

80.

001

3.2

0.01

6.2

Pea

kam

plit

ude

11.3

3.8

4.3

34.3

0.00

132

.02.

50.

0133

.03.

60.

001

32.7

3.1

0.05

6.2

n.s.

34.7

P1

impl

icit

tim

eA

MD

0.00

110

.22.

40.

001

9.3

2.0

0.05

4.3

11.0

16.7

2.4

n.s.

Pea

kam

plit

ude

8.0

0.00

17.

22.

535

.70.

001

33.7

3.2

0.00

132

.72.

30.

001

32.7

2.3

0.05

4.0

0.01

33.0

P1

impl

icit

tim

eC

D0.

001

3.0

0.5

0.00

12.

01.

10.

001

3.9

2.1

11.3

1.4

0.00

1P

eak

ampl

itud

e6.

20.

001

3.8

4.3

34.0

0.01

33.7

3.2

0.00

136

.75.

20.

001

34.0

3.3

0.00

14.

9n.

s.32

.7P

1im

plic

itti

me

CR

VO

0.00

16.

21.

60.

001

6.1

1.0

0.00

13.

06.

618

.70.

50.

001

Pea

kam

plit

ude

8.5

0.00

18.

61.

740

.00.

001

38.0

1.9

0.00

136

.02.

50.

001

35.3

1.9

0.00

14.

40.

001

40.7

P1

impl

icit

tim

eA

DO

An.

s.19

.61.

6n.

s.13

.61.

4n.

s.2.

713

.2n.

s.1.

4n.

s.30

.4P

eak

ampl

itud

e54

.510

.21.

335

.70.

0130

.70.

9n.

s.31

.00.

8n.

s.31

.00.

8n.

s.0.

80.

001

32.0

P1

impl

icit

tim

eN

orm

als

16.8

3.6

12.0

2.7

5.7

12.3

27.3

3.2

Pea

kam

plit

ude

59.8

13.0

P1

impl

icit

tim

e29

.632

.31.

430

.21.

131

.01.

11.

829

.91.

4

Inth

eth

ird

colu

mn

inea

chgr

oup

the

leve

lof

sign

ifica

nce

for

the

diff

eren

ceof

the

pati

ent-

and

norm

algr

oup

isgi

ven

(n.s

.=no

tsi

gnifi

cant

).


Fig. 3. Early SMD. In the left panel (a, b) data from the OS of a 21 year old male patient with a visual acuity of 0.8 are shown: (a) trace arrayof 61 focal ERGs and (b) the three-dimensional response density graph. In the right panel (c, d) data from the five individual patients of the groupare presented (asterix indicating the patient from a, b; (c) the foveal response (ring 1 up to 2° eccentricity) and (d) the response of ring 5 (17–30.5–eccentricity).

3.2.4. Cone dystrophy (Fig. 5)All patients with cone dystrophy had strongly re-

duced or non-detectable responses in the entire testfield. The patient shown in Fig. 5(a and b) had no focalresponse distinguishable from noise in any of the 61focal ERG traces. In all patients with cone dystrophy,no foveal response was discernible, and the responsedensities were diminished up to ring 5 (Fig. 5d, Table3). The implicit times were prolonged (Table 3).

3.2.5. CRVO (Fig. 6)In the patients with CRVO, the functional defect was

similar to that found in maculopathies. No foveal re-sponses were found that exceeded the noise level, but,unlike early SMD and AMD, the peripheral responseswere also significantly diminished. The smaller waveletsfollowing N2 in normals were absent. The implicit

times were markedly prolonged in the entire test field(Table 3). Frequently, a patchy appearance of the tracearray and the three-dimensional response density plotwas found (Fig. 6b) caused by areas with normal andwith markedly diminished responses in the directvicinity.

3.2.6. ADOA (Fig. 7)Normal multifocal ERG amplitudes within the entire

test field were found in all patients with autosomaldominant optic atrophy, regardless of their decreasedvisual acuity. Although foveal response densities werewithin the normal magnitude, the waveforms seemed tobe more noisy (Fig. 7c). Also, implicit times in themacular responses (ring 1 and 2) were significantly pro-longed, whereas extramacular implicit times were nor-mal (Table 3).


Fig. 4. In the left panel (a, b) data from the five individual patients with advanced SMD are presented: a) the foveal response (ring 1 up to 2°eccentricity) and b) the response of ring 5 (17–30.5° eccentricity). In the right panel (c,d) data from the five individual patients with AMD arepresented: c) the foveal response (ring 1 up to 2° eccentricity) and d) the response of ring 5 (17–30.5° eccentricity).

4. Discussion

The multifocal ERG allows mapping of the outerretinal function with a high resolution on the basis ofphotopic electrophysiological responses. The m-se-quence technique of Sutter [1,3] makes it possible toobtain a large number of focal ERGs simultaneouslywith a good signal to noise ratio. A new stimulus pictureis presented every 13.3 ms (monitor frame rate of 75 Hz),so that the recording sequence can be completed withinminutes. Since the interstimulus intervals are potentiallyshorter than the recovery time of the retina, each retinalregion is differently pre-adapted by the previous stimu-lation sequence. Therefore, the responses cannot beobtained just by averaging multiple recordings as in a

flash-ERG, but some computation is necessary to deter-mine the contribution of each stimulated area (by across-correlation with the m-sequence) and to extract thelinear portion (first order kernel) of this local response.

Sutter and Bearse [11] have also isolated ganglion cellcomponents from the non-linear second order kernel ofthe multifocal ERG. Contributions of the inner retina tothe linear first order kernel response are thought to benegligible. The group of patients with autosomal domi-nant optic atrophy (ADOA), all showing a first orderkernel response of normal amplitude (Fig. 7), reinforcesthis assumption, and may serve as a proof for thehypothesis that first order kernel data represents mainlythe outer retinal layers at least with high contraststimulation [12].


Fig. 5. Cone dystrophy. In the left panel (a, b) data from the OS of a 18 year old male patient with a visual acuity of 0.1 are shown: a) tracearray of 61 focal ERGs and b) the three-dimensional response density graph. In the right panel (c,d) data from the five individual patients of thegroup are presented (asterix indicating the patient from a, b): c) the foveal response (ring 1 up to 2° eccentricity) and d) the response of ring 5(17–30.5° eccentricity).

Hood et al. [6] have found a close relationship be-tween Ganzfeld-ERG and multifocal ERG first orderkernels. After slowing down the stimulation rate of theMFERG by the insertion of frames with backgroundintensity, responses with an a- and b-wave and anoscillatory potential (PC1 in their nomenclature), simi-lar to the Ganzfeld-ERG, were recorded. They showedthat if a fast stimulus sequence is used, PC1 and theb-wave merge to form a single wave. The same effectwas seen in the case of slow stimulation together with ahigh background intensity. The authors conclude thatthe multifocal ERG obtained with a fast stimulationsequence is equivalent to a photopic ERG under condi-tions of a high light adaptation level. The waveformsrecorded here in normal subjects show this relativelysimple shape (Fig. 2), which was described above in theresults section.

A comparable decrease in response density with ec-centricity (Table 2) was previously detected with con-

ventional focal stimulation using a Maxwellian viewsystem [13]. Variations of the implicit time with eccen-tricity were also found by these authors and furtheranalyzed by Yamamoto et al. [14] using focal laserbeam stimulation in monkeys. The spatial distributionof implicit times of multifocal ERGs was recently inves-tigated by our group [15]. The higher implicit time atthe fovea (Table 2), compared to more peripheral areas(with a minimum in the parafoveal region), complieswith these results.

Given that the multifocal ERG is in principle aphotopic ERG under special conditions of light adapta-tion, the most impressive alterations in the multifocalERG can be expected in patients with general receptordystrophies leading to reduced and delayed Ganzfeldresponses. Consequently, the most prominent and wide-spread changes of amplitudes were found in the conedystrophy group. Since there were barely any clearresponses in the macular region it was difficult to


Fig. 6. CRVO. In the left panel (a, b) data from the OS of a 58 year old female patient with a visual acuity of 0.2 are shown: a) trace array of61 focal ERGs and b) the three-dimensional response density graph. In the right panel (c,d) data from the five individual patients of the groupare presented (asterix indicating the patient from a, b); c) the foveal response (ring 1 up to 2° eccentricity) and d) the response of ring 5 (17–30.5°eccentricity).

reliably measure implicit times. For the averaged re-sponses of the three more peripheral rings implicit timeswere highly significantly prolonged, as expected from theGanzfeld-ERG.

Juvenile (Stargardt’s) and age-related maculopathieswere considered diseases in which a primary defect in theretinal pigment epithelium leads to a secondary, localizeddegeneration of photoreceptors. Recently, a defect in theABCR-gene which is expressed in rods was found inpatients with Stargardt’s macular dystrophy [16]. There-fore, changes in the pigment epithelium and cones mightbe secondary to a primary defect in the rod photorecep-tors at least in some of the patients. Similar mechanismswere proposed for AMD [17]. The following thoughtsmay illustrate why Ganzfeld-ERG cannot detect thephotoreceptor loss in these diseases. The area stimulatedby the multifocal ERG setup used here covers roughly35% of the entire cone population. The contribution ofthe macular area (ring 1 and 2) to the overall sum ofmultifocal ERG responses is only about 12%. Under the

simplifying assumption that all cones are reached by thephotopic Ganzfeld-ERG and each cone contributesequally, the estimated loss caused by an isolated defectof the macula would be about 4%, which is certainlywithin the range of inter-individual variability in nor-mals.

The data presented show that the multifocal ERGdetected the central functional defect in every patientwith a maculopathy, regardless of the stage of the disease.Even in the SMD patient with nearly full visual acuity,the foveal responses were diminished. The differences inspatial activity distribution between the early and theadvanced SMD group followed largely the extension ofvisual field defects; in the early SMD group (field defectdiameter B10°), electroretinographic lesions were muchsmaller than in the advanced SMD group with a largerdefect in perimetry. In many cases, multifocal ERG canbe used to verify unclear visual field defects of retinalorigin, although it is not possible to predict visual fielddefects from a reduced or extinguished MFERG.


Fig. 7. ADOA. In the left panel (a, b) data from the OS of a 20 year old male patient with a visual acuity of 0.1 are shown: a) trace array of61 focal ERGs and b) the three-dimensional response density graph. In the right panel (c,d) data from the five individual patients of the groupare presented (asterix indicating the patient from a, b); c) the foveal response (ring 1 up to 2° eccentricity) and d) the response of ring 5 (17–30.5°eccentricity)

The implicit times were normal in the early SMDgroup and significantly prolonged in advanced SMDand AMD. Since the size of this increase is moderate,the patients with maculopathies can be better distin-guished from normals by the loss of the central re-sponses than by the changes of implicit times. Incontrast, increases in implicit times in the affectedregion in MFERG are much more pronounced in Re-tinitis pigmentosa [15].

The overall outcome of the multifocal ERG in pa-tients with CRVO was a decrease in amplitude andan increase in implicit times as in flicker Ganzfeld-ERG. The effect on the fovea was more pronouncedthan the effect on the extramacular area. One reasonfor this may be the greater susceptibility of the ma-cula to develop oedema [18]. Compared to the find-ings in maculopathies, the defect had a patchy appear-ance.

There was no negative waveform in the ERG of thepatients with CRVO. The filter settings may influencethe shape of the waveform, as addressed by Keating etal. [19]. Since the high pass filter was set to 10 Hz,negative values cannot be expected as in Ganzfeld-flicker-ERG recordings following the ISCEV standard[20]. Secondly, the fast stimulation technique used inthis study with an average stimulation frequency of37.5 Hz might elicit waveforms similar to Ganzfeld-Flicker-ERG [6].

The smaller wavelets following N2 in normals wereabsent in CRVO. They were present in the patients withmaculopathies with similar decrease of peak-to-peakamplitude and in the patients with ADOA. Therefore,these potentials might be generated in inner retinallayers but distal from the ganglion cells. Similarchanges were found by Palmowski et al. [21] in diabeticsubjects in the second order kernel.


The shape of the waveform with its major compo-nents N1, P1, and N2 was not affected by the differ-ent diseases beyond amplitude reduction and implicittime increase. Again, this is similar to the Ganzfeld-Flicker-ERG.

The patients with autosomal dominant optic atro-phy could be clearly distinguished from any of ourother patients since amplitudes in all areas were nor-mal. Multifocal ERG can therefore separate patientswith maculopathies from patients with optic nervediseases, which is important for differential diagnosis.Foveal implicit times were significantly prolonged inthese patients compared to normal volunteers. Al-though the responses were more noisy in the ADOA-group than in normals, the prolonged implicit timescan be considered to be due to the pathophysiologyof the disease, since noise would bias measurementsin all 61 elements in the same direction. In the fivepatients with a decreased visual acuity and a centralscotoma, only the macular implicit times were in-creased, while extramacular implicit times were nor-mal. An abnormality in the ganglion cell layer mayinfluence the first order kernel response or in retinalelements other than the ganglion cell may be affectedin ADOA.

5. Conclusion

The multifocal electroretinography provides detailedinformation of local activity of the cone system. Itwas possible to detect macular cone dysfunction inpatients with early and advanced SMD, AMD, conedystrophies, and retinal vein occlusions. In most ofthe patients with maculopathies, the photopicGanzfeld ERG was normal, so that multifocal ERGcan be valuable in diagnosing these diseases. Also,affections of the retina could be distinguished fromoptic nerve diseases, since all patients with ADOAshowed normal multifocal ERG amplitudes in the en-tire test field.

Acknowledgements

Software and hardware were provided by Dr E.Sutter, EDI, San Francisco, and Tomey, Nagoya,Japan. We thank J. Isensee and S. Scholz for theirtechnical assistance, R. Hofer for help with thegraphics, and Dr A. Kurtenbach for commenting onthe manuscript. The work was supported by theDeutsche Forschungsgemeinschaft, Bonn, Germany(grant Zr 1/10-1 and Zr 1/10-2).

References

[1] Sutter EE. The fast m-transform: a fast computation of cross-correlations with binary m-sequences. Soc Ind Appl Math JComput 1991;20:686–94.

[2] Sutter EE, Tran D. The field topography of ERG components inman-I: the photopic luminance response. Vis Res 1992;32:433–46.

[3] Sutter EE. A deterministic approach to nonlinear systems analy-sis. In: Pinter RB, Nabet B, editors. Nonlinear Vision: Determi-nation of Neural Receptive Fields, Function, and Networks.Boca Raton: CRC Press, 1992:171–220.

[4] Østerberg G. Topography of the layer of rods and cones in thehuman retina. Acta Ophthalmol Suppl 1935;6:1–103.

[5] Curcio CA, Sloan KR, Kalina RE, Hendrickson AE. Humanphotoreceptor topography. J Comp Neurol 1990;292:497–523.

[6] Hood DC, Seiple W, Holopigian K, Greenstein V. A comparisonof the components of the multifocal and full-field ERGs. VisNeurosci 1997;14:533–44.

[7] Kondo M, Miyake Y, Horiguchi M, Suzuki S, Tanikawa A.Clinical evaluation of multifocal electroretinogram. Invest Oph-thalmol Vis Sci 1995;36:2146–50.

[8] Bearse MA, Sutter EE. Imaging localized retinal dysfunctionwith the multifocal electroretinogram. J Opt Soc Am A1996;13:634–40.

[9] Kretschmann U, Ruther K, Usui T, Zrenner E. ERG campime-try using a multi-input stimulation technique for mapping ofretinal function in the central visual field. Ophthalmic Res1996;28:303–11.

[10] Parks S, Keating D, Williamson TH, Evans AL, Elliot AT, JayJL. Functional imaging of the retina using the multifocal elec-troretinograph: a control study. Br J Ophthalmol 1996;80:831–4.

[11] Sutter EE, Bearse MA. Extraction of a ganglion cell componentfrom the human electroretinogram. Invest Ophthalmol Vis SciSuppl 1995;36:444.

[12] Bearse MA, Sutter EE. The different contrast dependencies ofmultifocal electroretinogram components. Invest Opthalmol VisSci 1998;39:973.

[13] Sandberg MA, Effron MH, Berson EL. Focal cone electroretino-grams in dominant retinitis pigmentosa with reduced penetrance.Invest Ophthalmol Vis Sci 1978;17:1096–101.

[14] Yamamoto S, Gouras P, Lopez R. The focal cone electroretino-gram. Vis Res 1995;35:1641–9.

[15] Seeliger M, Kretschmann U, Apfelstedt-Sylla E, Zrenner E.Implicit time topography of multifocal electroretinograms. InvestOphthalmol Vis Sci 1998;39:718–23.

[16] Allikmets R, Singh N, Sun H, et al. A photoreceptor cell-specificATP-binding transporter gene (ABCR) is mutated in recessiveStargardt macular dystrophy. Nat Genet 1997;15:236–46.

[17] Allikmets R, Shroyer NF, Singh N, et al. Mutation of theStargardt disease gene (ABCR) in age related macular degenera-tion. Science 1997;277:1805–7.

[18] Cunha-Vaz JG, Travassos A. Breakdown of the blood-retinalbarriers and cystoid macular edema. Surv Ophthalmol1984;28:485–92.

[19] Keating D, Parks S, Evans AL, Williamson TH, Elliott AT, JayJL. The effect of filter bandwidth on the multifocal electroretino-gram. Doc Ophthalmol 1997;92:291–300.

[20] Marmor MF, Zrenner E. Standard for clinical electroretinogra-phy (1994 upgrade). Doc Ophthalmol 1996;89:199–210.

[21] Palmowski AM, Sutter EE, Bearse MA, Fung W. Mapping ofretinal function in diabetic retinopathy using the multifocalelectroretinogram. Invest Ophthalmol Vis Sci 1997;38:2586–96.

spatial cone activity distribution in diseases of the …vision research 38 (1998) 3817–3828...

Documents