Multifocal ERG Findings in Complete Type Congenital Stationary Night Blindness

Abstract

purpose. To study the multifocal electroretinogram (mfERG) in patients with the complete type of congenital stationary night blindness (cCSNB), which is thought to be due to a defect in neurotransmission from the photoreceptors to the ON-bipolar cells.

methods. mfERGs were recorded with the VERIS recording system from four patients with cCSNB, none of whom had nystagmus. The stimulus array consisted of 61 hexagons, and the total recording time was approximately 4 minutes. The amplitudes and implicit times of the first- and second-order kernels of the local responses were compared with those from 20 myopic controls. Waveforms of the summed response from all locations were also compared between the two groups.

results. The first-order kernels of the mfERGs of cCSNB patients had normal amplitudes but delayed implicit times for nearly the whole field tested. The second-order kernel was severely attenuated in amplitude in cCSNB patients. The ratios of the second- to first-order kernel amplitudes were significantly reduced in cCSNB and clearly separated the cCSNB group from the control group without any overlap of the values.

conclusions. The second-order kernel, which is involved in adaptative mechanism of the retina to repeated flashes, is selectively reduced in cCSNB. The delay of the implicit times of the first-order kernel in patients with cCSNB may be related to the severe amplitude reduction of the second-order kernel.

The Schubert-Bornschein type of congenital stationary night blindness (CSNB) is a nonprogressive retinal disorder characterized by night blindness, moderately decreased visual acuity, and myopia. ¹ The electroretinograms (ERGs) in patients with CSNB are quite characteristic; when elicited by a bright stimulus after dark adaptation, the ERGs are the negative-type with an a-wave of normal amplitude and a b-wave that is smaller than the a-wave. Rhodopsin density and kinetics have been shown to be normal in patients with CSNB. ² Thus, the defect in this disorder is thought to lie not in the photoactivity in rod photoreceptors but in the neurotransmission from the rods to the rod bipolar cells. 

CSNB has been subdivided into two clinical entities: the complete type (cCSNB), which has no detectable rod function, and the incomplete type (iCSNB), which has a small but detectable rod function. ^{3 4} Recent genetic analyses have shown that these two types have separate genetic loci on the chromosome, supporting the idea that they are distinct clinical entities. ^{5 6 7 8}  

The cone pathway is also thought to be affected in CSNB. The amplitude of the photopic ERG b-wave elicited by a brief flash is reduced in cCSNB. ^{9 10} In addition, when a long-duration photopic stimulus is used, cCSNB patients show severely reduced ON-response b-wave, whereas the OFF-response d-wave is preserved. ^{11 12 13} This waveform can be simulated in the monkey photopic ERG after treatment with 2-amino-4-phosphonobutyric acid, which blocks neurotransmission from photoreceptors to the ON-bipolar cells. ^{14 15 16} This implies that the defect in the cone pathway of cCSNB also lies in the signal transmission from the cone photoreceptors to the depolarizing ON-bipolar cells. ^{12 13}  

The multifocal ERG (mfERG) is a relatively new objective test designed to study local retinal function. ^{17 18} This technique allows the simultaneous recording of focal cone ERGs from multiple retinal locations in a single recording session of approximately 4 to 8 minutes. This technique is particularly useful when one wants to detect local retinal damage or to assess retinal function topographically in both normal subjects and patients with retinal diseases. ^{19 20 21 22 23 24 25 26 27 28 29}  

Although numerous researchers ^{17 18 19 20 21 22 23 24 25 26 27 28 29} have been assessing retinal function in various retinal diseases using the mfERG, the exact origin of each component of the mfERG is still under investigation. There is recent evidence that the negative and positive components of the first-order kernel of mfERG behave as do the a-wave and the positive peaks of the conventional full-field flash cone ERG. ^{21 23} In addition, by use of glutamate analogs in rabbits, Horiguchi et al. ³⁰ demonstrated that the first-order kernel of the mfERG contains significant contribution from postreceptoral ON- and OFF-components as do the conventional flash cone ERG, ^{16 31} supporting the results of Hood and coworkers. The second-order kernel, the temporal nonlinear component of the mfERGs, is interpreted to be involved in short-term adaptational mechanisms of the retina to successive flashes and contains a greater contribution from the proximal retina. ^{20 23 30 32 33 34}  

The aim of the present investigation was to extend these previous studies in establishing the retinal origins of each component of the mfERG. We examined how the waveform of the first- and second-order kernels of the mfERGs is changed in patients with cCSNB. 

Methods

Subjects

From the patients with cCSNB seen in our clinic (Department of Ophthalmology, Nagoya University School of Medicine), four cooperative patients were recruited and examined. None of them had nystagmus or irregular eye movements before and during the mfERG recordings. The clinical characteristics of the four patients are summarized in Table 1 . Patients 1 and 2 are siblings and were classified as autosomal recessive, patient 3 as an X-linked, and patient 4 as an autosomal recessive cCSNB. Only the eye with the better visual acuity was tested. The corrected visual acuity ranged from 0.3 (20/67) to 0.6 (20/33), and the refractive error from −4.00 to −9.50 diopters (D) with a mean of− 6.50 D. 

Each patient was diagnosed as having cCSNB, based on ophthalmological, psychophysical, and electrophysiological examinations. ^{3 4 11} All patients had poor night vision, and no fundus abnormalities were seen except for myopic changes. The rod branch of the dark adaptation curve was missing on psychophysical dark adaptometry. The full-field ERGs recorded from the four cCSNB patients are shown in Figure 1 . The rod responses were undetectable. The rod and cone mixed maximal ERGs had a negative-shaped ERG with no oscillatory potentials. The brief-flash cone responses presented a wide a-wave trough, and the cone b-wave was reduced by 30% to 45%. The amplitude ratio of the cone b-wave to the a-wave was reduced (range of cCSNB, 1.10–1.35; lower limit of normals, 1.41). The amplitude of the 30-Hz flicker ERGs was reduced by 10% to 35%. The photopic long-flash ERG showed severely reduced ON-response b-wave and normal OFF-response d-wave. 

For control, 20 normal subjects (age, 19–38 years; mean, 27.1 years) with myopia (refractive error, −3.00 to −8.50 D; mean, −5.15 D) were examined. None had known abnormalities of the visual system except for myopia, and visual acuity was 1.0 (20/20) or better. Informed consent was obtained after a full explanation of the procedures. All studies were conducted in accordance with the principles embodied in the Declaration of Helsinki. 

Multifocal ERGs

The method of recording the mfERG has been reported in detail previously. ^{17 18 19 20 21 22 23 24 25 26 27 28 29} In brief, mfERGs were recorded with a VERIS recording system (EDI, San Mateo, CA). The visual stimulus consisted of 61 hexagonal elements scaled in size to give approximately equal amplitude mfERGs with eccentricity. The stimulus array was displayed on a high resolution CRT monitor (SONY GDM, Tokyo, Japan) driven at a 75-Hz frame rate. At a viewing distance of 27 cm, the diameter of the stimulus array subtended approximately 60°. The luminance of each hexagon was independently modulated between black (3.5 cd/m²) and white (138.0 cd/m²) according to a binary m-sequence at 75 Hz. A small red fixation spot was placed at the center of the stimulus matrix. The luminance of the surround was set at 70.8 cd/m². 

Before the recording, the subject’s pupil was fully dilated with a combination of 0.5% tropicamide and 0.5% phenylephrine hydrochloride, and the cornea was anesthetized with proparacaine hydrochloride. ERGs were recorded with a Burian-Allen bipolar contact lens electrode (Hansen Ophthalmic Laboratories, Iowa City, IA), and a ground electrode was attached to the earlobe. After insertion of the contact lens electrode, the subject was optically corrected for the viewing distance. The opposite eye was occluded. 

The signals were amplified (100,000×), and the band-pass filter was set at 10 to 300 Hz (Grass, Quincy, MA). The sampling rate was 1200 Hz (interval, 0.833 msec). The m-sequence used in this study had 2¹⁴ elements and required a total recording time of approximately 4 minutes. For the comfort of the subjects, the recording time was divided into eight segments. The first- and second-order kernels were analyzed with VERIS 2.05 software (EDI). 

An artifact reduction technique was used once to improve the signal-to-noise ratio for all subjects. ¹⁷ A small amount of spatial filtering was also applied in one patient (P3) because his local responses were relatively noisy and the exact amplitude and implicit time at each location could not be made; each individual response was averaged with 6% of its six neighboring responses. 

Results

First-Order Kernel of Local Responses

Figure 2 shows the 61 first-order kernels of the local responses from a representative myopic control (age, 20 years; refractive error, −6.00 D) and four patients with cCSNB. At first glance, it is difficult to distinguish the patients from the myopic control; both the amplitudes and waveforms of each local response appear similar. The amplitudes of the local responses of the myopic controls and the patients with cCSNB were often slightly smaller than nonmyopic normals. Because there is electrophysiological evidence that moderate-to-high myopia causes depression of cone function in the posterior pole of the eye, ^{35 36} the slight amplitude reduction was considered to be due to the myopic changes. 

We next assessed the local cone function quantitatively for two parameters, viz., the amplitude and timing for the four cCSNB patients. The amplitudes and implicit times of the positive component were measured at all 61 locations as shown at the top of Figure 3 . As limits of normality, the 5 percentile and 95 percentile values were obtained for the 20 myopic controls at each location. In Figure 3 , the white areas indicate that the values of the amplitude or implicit time were within the 5 percentile to 95 percentile range. The black areas indicate low amplitudes or delayed implicit times greater than this normal range. 

Three of the four CSNB patients (P2–P4) had normal amplitudes at all locations tested. Only one patient (P1) had reduced amplitudes in 11 areas of the nasal field. In contrast, all four CSNB patients had delayed implicit times in many areas. Note that one patient (P4) had normal amplitudes with delayed implicit times at all 61 locations. Of the total 244 retinal areas tested (61 locations × 4 patients), 210 areas (86%) had significantly delayed implicit times, whereas only 11 areas (5%) had reduced amplitudes. These findings indicate that the pathology of cCSNB affects the mfERG implicit times in preference to reducing the amplitude. 

We also examined whether there were any topographical variations in the degree of implicit time delays with retinal eccentricity, between the upper/lower or nasal/temporal retina, but no significant variations were found in cCSNB. 

Waveform Change of First-Order Kernel

To compare the waveforms of the cCSNB patients and the myopic controls, the 61 local responses were summed. The superimposed and averaged response waveforms for the 20 myopic controls are shown on the left of Figure 4 , and the waveforms for the four cCSNB patients are shown on the right of Figure 4 . The vertical dashed lines are drawn at 30 msec. The results of statistical comparisons for the two groups are presented in Table 2 . The amplitudes of the initial negative (N1) and following positive component (P1) were not significantly different in the two groups. The amplitude ratio of P1 to N1 also did not differ significantly between the two groups. It was surprising that the mean amplitude of P1 was equal to or even slightly larger in the cCSNB patients than in myopic controls (Table 2 and Fig. 4 ), which was in contrast to the results of conventional full-field cone ERGs (Fig. 1) . The implicit times of N1 and P1 were significantly delayed (P < 0.05, nonparametric Mann–Whitney test). 

In addition to the timing delays for the N1 and P1 components, we also noticed other minor differences in the later components. First, the second negative component (N2) was less prominent in the responses from the cCSNB patients (arrowhead in lower right panel of Fig. 4 ). Second, the two or three oscillations (asterisks) followed by N2 component were diminished or essentially absent in cCSNB. Similar minor waveform changes in the late components of the mfERGs have been reported in some patients with diabetic retinopathy. ²⁸ It is known that the later portion of the first-order kernel contains contributions from second- and higher-order kernels and thus reflects the nonlinear temporal interactions between flashes, as do the higher order kernels. ²³  

Second-Order Kernel

Figure 5 shows the 61 local responses of the first slice of the second-order kernel for the same myopic control and four cCSNB patients shown in Figure 2 . Because the second-order kernel is relatively small when compared with the first-order kernel, it is often difficult to identify each positive or negative component at local areas even for the controls. It is evident, however, that the local responses of the second-order kernel are severely reduced or virtually absent in the four cCSNB patients. Only one patient (P4) had detectable local responses in some regions in the upper and nasal fields, but these responses were still smaller than those of the controls. 

To compare the waveforms, all the 61 second-order kernels were summed and the summed kernels are presented in Figure 6A . As in Figure 4 , the superimposed and averaged response waveforms for 20 myopic controls are shown on the left, and the waveforms for the four cCSNB patients are shown on the right. The amplitude of the summed second-order kernel in cCSNB patients was markedly reduced as opposed to the normal amplitudes of the first-order kernel. The amplitudes of the second negative component (N2), the most prominent component in the second-order kernel, were measured as shown in the left side of Figure 6B , and the mean value was found to be significantly reduced by more than half of the control amplitude (Table 2) . 

Because it is known that the amplitudes of the mfERGs have fairly large intersubject variation, one of the effective ways to measure the relative amplitude of the second-order kernel was to calculate a ratio of the amplitudes of the second- to first-order kernels for each subject. ³⁷ We calculated this ratio and found that it was significantly reduced in all cCSNB patients (Table 2) , and a plot of the ratios separated the two groups clearly without any overlap (Fig. 6B) . 

Discussion

Waveform Changes of First- and Second-Order Kernels in cCSNB

The results demonstrated that the cCSNB patients had two distinct mfERG waveform changes: (1) normal amplitude with delayed implicit times for the first-order kernel and (2) severe amplitude reduction for the second-order kernel. These waveform changes were present for nearly the whole field tested for all four cCSNB patients. However, they do not appear to be specific for cCSNB because a recent study ²³ reported that some patients with retinitis pigmentosa and diabetic retinopathy have similar mfERG waveform changes. In addition, we have observed that some patients with X-linked retinoschisis also have similar waveform changes (unpublished data, 2000). It is unlikely that all these patients with different diseases have a common defect in the retinal ON-pathway as do the patients with cCSNB. 

There are two findings showing that the two waveform changes, delayed first-order kernel and reduced second-order kernel, are related. First, the second-order kernel contributes to the later portion of the first-order kernel, and its contribution begins earlier than the peak of the positive component (P1). ²³ Thus, the change in the second-order kernel can affect the timing of the first-order kernel. Second, it has been reported that large, but delayed, first-order kernels seen in some patients are always associated with reduced second-order kernels. ²³ Hood ²³ demonstrated that this holds true consistently both within patients and for patients with various diseases. Therefore, the most plausible interpretation of the waveform changes in cCSNB is that the second-order kernel, which is involved in adaptative mechanisms of the retina to successive flashes, is reduced in cCSNB, presumably because of the abnormality in postsynaptic ON pathway, and this reduced second-order kernel may cause delayed implicit times for the first-order kernel. 

Comparison with Conventional Full-Field ERG

It is currently accepted that the N1 and P1 components of the mfERGs are comprised of the same components as the a- and the positive waves (b-wave and OPs) of the conventional full-field cone ERGs. ^{21 23 30} However, it is still unknown to what extent the N1 and P1 components of the mfERGs are correlated with the a- and b-waves of the conventional cone ERGs in clinical diseases. In our cCSNB patients, the b-wave of the full-field cone ERGs was reduced by 30% to 45% (Fig. 1) , as has been previously reported, ^{3 9 10} but the P1 component of the first-order kernel was not reduced. Another discrepancy between the two waveforms is in the ratio of the amplitudes of the positive to negative components. Although the amplitude ratio of the b-wave to the a-wave for full-field cone ERGs is clearly reduced in CSNB (Fig. 1) , the amplitude ratio of the P1 to N1 of the mfERGs was equal or even slightly larger than in the myopic controls (Table 1 and Fig. 3 ). These results suggest that, although the N1 and P1 of the mfERG may originate from same retinal elements as the negative and positive components of conventional cone ERG, the two waveforms are not necessarily correlated quantitatively in retinal diseases. This disparity is thought to be mainly due to different stimulus intensities and adaptational states between the two stimulus conditions. ²¹  

Our present results of delayed implicit times without an amplitude reduction for the first-order kernel are reminiscent of a previous report of 30-Hz flicker ERG analysis in cCSNB. Kim et al. ³⁸ reported that cCSNB patients had a phase delay in the fundamental component of strobe-flash 30-Hz flicker ERG without significant amplitude reduction. Although the fundamental component of the 30-Hz flicker ERG was not analyzed in our patients because of our recording conditions, this agreement suggests that the defect in cCSNB causes similar waveform changes of both the fundamental component of 30-Hz flicker ERG and first-order kernel of mfERG. This implies that there is a possibility that an interaction of the ON- and OFF-components, which is observed in primate flicker ERG, ^{38 39 40} may be involved in shaping the first-order kernel of the mfERGs. 

Clinical Implications

Finally, we would like to emphasize again the importance of measuring the implicit times of the mfERG. ^{22 25 26 28 29} cCSNB is a nonprogressive retinal diseases characterized by night blindness, decreased visual acuity and myopia. Fundus findings are usually normal except for myopic changes. The diagnosis of cCSNB is not difficult if patients visit ophthalmologists with complaints of night blindness. However, it is known that cCSNB patients often visit ophthalmologist only with complaints of low visual acuity or myopia. Actually, in our hospital, the most common complaint for cCSNB at the initial visit is not night blindness but decreased visual acuity. If ophthalmologists do not realize the possibility of cCSNB, they may order mfERG testing to assess central retinal function in these patients. Then if they are not aware of the timing delay of the mfERGs, they may conclude that cCSNB patients have normal retinal function across whole field tested and may misdiagnose them as having amblyopia, optic nerve or central nervous system disease, or a psychological visual loss. 

 Supported by Grant-in Aid 11470363 from the Ministry of Education, Science, Sports and Culture, Japan.

 Submitted for publication July 11, 2000; revised October 30, 2000 and January 18, 2001; accepted January 26, 2001.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked“ advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Clinical Characteristics of Examined Patients

Case	Age/Sex	Inheritance Pattern	Tested Eye	Refractive Error	Visual Acuity
P1	13/M	Autosomal recessive	OS	−6.50 D	0.3 (20/67)
P2	11/F	Autosomal recessive	OD	−6.50 D	0.6 (20/33)
P3	14/M	X-linked	OD	−9.50 D	0.4 (20/50)
P4	51/M	Autosomal recessive	OS	−4.00 D	0.4 (20/50)

Figure 1.

 

Full-field ERGs recorded from a myopic control and four patients with cCSNB (P1–P4). After 30 minutes of dark adaptation, a rod ERG was recorded with a blue light at an intensity of 5.2 × 10⁻³ cd-s/m². A cone-rod mixed maximum ERG was recorded with a white flash at an intensity of 44.2 cd-s/m². A cone ERG and a 30-Hz flicker ERG were recorded with a white stimulus of 4 cd-s/m² and 0.9 cd-s/m², respectively, on a background illumination of 68 cd-s/m². A photopic long-flash (200 msec) ERG was recorded with a light-emitting diode (peak wavelength, 566 nm) built-in contact lens electrode at an intensity of 250 cd/m² under a blue background illumination of 34 cd/m². ⁴¹

Figure 2.

 

Sixty-one local first-order kernels of the mfERGs recorded from a myopic control (age, 20 years; refractive error, −6.00 D) and four patients with cCSNB (P1–P4).

Figure 3.

 

Topographical map of the amplitudes and implicit times for four patients with cCSNB. White areas: values are within 5 percentile to 95 percentile range; black areas: low amplitudes or delayed implicit times greater than this normal range. Note that there are regional variations on both the amplitude and implicit time of the mfERG across the retina for normal subjects. The normal ranges of the amplitude and implicit time were calculated at all locations independently.

Figure 4.

 

Summed first-order kernel responses for all 61 local responses for 20 myopic controls (left) and four cCSNB patients (right). All responses were superimposed in the upper traces and averaged waveforms are presented in the lower traces.

Table 2.

 

Parameters of the mfERG of cCSNB Group Compared with Myopic Control Group

	n	First-Order Kernel							Amplitude Ratio of P1/N1	Second-Order Kernel			Amplitude Ratio of 2nd- to 1st-Order Kernel
				N1 Amp.	P1 Amp.	N1 Time	P1 Time			N2 Amp.	N2 Time
Myopia	20	10.5 ± 3.3	24.6 ± 7.9	15.8 ± 0.4	29.3 ± 0.7	2.3 ± 0.5	8.2 ± 2.2	29.7 ± 1.1	0.28 ± 0.08
cCSNB	4	10.2 ± 2.6	27.5 ± 5.7	17.5 ± 0.7*	32.5 ± 1.2*	2.7 ± 0.2	3.4 ± 1.1*	31.2 ± 1.9*	0.10 ± 0.03*

 Values are means ± SD.

*  P < 0.05, nonparametric Mann–Whitney test.

Figure 5.

 

Sixty-one local second-order kernels of the mfERGs recorded from same subjects as shown in Figure 1 : a myopic control and four patients with cCSNB.

Figure 6.

 

(A) Summed second-order kernels across all 61 local responses for 20 myopic controls (left) and four cCSNB patients (right). All responses were superimposed in the top trace and averaged waveforms are presented in the bottom trace. (B) Amplitude ratio of the second- to first-order kernel for 20 myopic controls and four cCSNB patients. Note that the ratio is significantly lower for cCSNB than for myopic controls, and the ratio separates the two groups without any overlap.

The authors thank Donald C. Hood for valuable discussions and comments on the manuscript.