The Multifocal Electroretinogram : Journal of Neuro-Ophthalmology

Secondary Logo

Journal Logo


The Multifocal Electroretinogram

Hood, Donald C. PhD; Odel, Jeffrey G. MD; Chen, Candice S. MD; Winn, Bryan J. BA

Author Information
Journal of Neuro-Ophthalmology 23(3):p 225-235, September 2003.
  • Free



The electroretinogram (ERG) is a mass potential, the result of the summed electrical activity of the cells of the retina. Typically, the clinical ERG is elicited by full-field (Ganzfeld) flashes of light. With an appropriate selection of test and background lights, rod and cone function can be assessed separately (1). As the ganglion cells contribute relatively little to the full-field flash ERG, the ERG has helped neuro-ophthalmologists to distinguish between diseases of the outer retina (affecting photoreceptors and/or bipolar cells) and diseases of the inner retina (ganglion cells) and optic nerve. However, because the ERG is the sum of all retinal activity, relatively large retinal defects may not be detected by standard full-field ERG testing. Although the pattern ERG and focal ERG can both provide information about visual loss from lesions in the foveal region (2,3), these techniques do not provide topographical information or assessment of nonfoveal lesions.

The multifocal ERG (mfERG) technique was developed by Sutter et al. (4–6) to provide a topographical measure of retinal activity. With the multifocal technique, 61 or 103 focal ERG responses can be recorded from the cone-driven retina within minutes. Although the technique is relatively new, hundreds of centers around the world have the equipment necessary to record mfERG responses. The equipment developed by Sutter (Electro-Diagnostic Imaging [EDI], San Mateo, CA) dominates the market, but equipment from other companies (such as Roland Instruments, Germany) can be found outside the United States.

The mfERG has been used widely to diagnose and study retinal diseases (7). In fact, over 200 articles on the mfERG, most dealing with clinical topics, have been published within the past 5 years. In this review, we discuss the mfERG technique and its applications in neuro-ophthalmology. When combined with automated perimetry, the mfERG is a valuable tool for localization and differential diagnosis.


The Display

Figure 1A,B shows the mfERG display used in the work summarized here. Similar to the display originally described by Sutter and Tran (5), it is a standard part of the VERIS software (EDI, San Mateo, CA) developed by Sutter (4–6,8). It consists of 103-scaled hexagons that subtend approximately 50° in diameter when viewed at 32 cm. The scaling of the hexagons is selected to produce approximately equal-sized mfERG responses from individuals with normal retinal function (5). In Figure 1A, the width of the central sector is about 3°, whereas the width of the outermost sector exceeds 7°. For clinical purposes, some investigators have used a display of 61 hexagons, which produces larger responses but has poorer spatial resolution.

The multifocal electroretinogram (mfERG) display. A, Top: The mfERG display with circles drawn to indicate radii of 5°(thick dark gray), 15°(thin black) and 25° (dashed light gray). Middle: A schematic of the eye to illustrate where the image of the display falls. Bottom: The 3D mfERG density plot of the responses (E) from a normal subject's OD. B, The mfERG display at one moment in time. C, The stimulation sequence of 2 sectors in (B). D, The single continuous ERG record generated by the display. E, The 103 mfERG responses (first-order kernel) extracted by correlating the stimulus sequence (C) with the continuous ERG record (D).

A wide range of stimulus intensities has been used. However, the International Society for Clinical Electrophysiology in Vision (ISCEV) guidelines for the mfERG (9) suggest that the luminance of the white hexagons be set to a value between 100 and 200 cd/m2. For the work described here, the luminance of the white hexagons is 200 cd/m2 and the luminance of the black hexagons is about 2 cd/m2, the darkest the screen allowed. The luminance of the area surrounding the array of hexagons is set at 100 cd/m2 and a central cross is used for fixation. All recordings are performed with the room lights on to help assure a constant state of light adaptation. Because of the light levels used and the rapid rate of stimulation, the mfERG is a response of the cone system.

Recording the Signal

In general, the mfERG signal is recorded with the same electrodes and amplifiers used for conventional ERG recording. The critical differences are the display, the method of stimulation, and the analysis of the raw records. For the records shown here, a single continuous ERG record is obtained with a Burian-Allen contact lens electrode, although a variety of electrode types can be used. The most commonly used noncontact lens electrode is the DTL (10). The contact lens electrode is less comfortable to wear but yields superior records. Although it is possible to analyze records offline (11–13), all the analyses shown here are done with the VERIS software that is part of the system from EDI.

From a single, continuous ERG signal (Fig. 1D), the software extracts 103 mfERG responses, each associated with one of the sectors of the display (Fig. 1E). That is, 103 responses are obtained from a single record. To get a sense of how this is possible, the nature of the local stimulation of each sector must be examined.

Stimulation and Extraction of the Local Responses

To understand the mfERG technique, it is essential to understand how each of the sectors is varied during the test. Each sector is an independent stimulus. Every 13.3 milliseconds, the frame of the monitor changes and each sector has a 50% chance of appearing “white” (briefly flashed) or “black” (no flash). Figure 1C shows a series of frame changes for 2 of the locations. Each of the 103 sectors of the display in Figure 1A,B goes through its own pseudorandom sequence. In fact, the 103 pseudorandom sequences are the same series of “white” or “black” presentations, but the sequence for each of the hexagons starts at a different point in the series. The rationale for this technique is complicated (4–8); it is sufficient to know that these pseudorandom sequences allow the software to rapidly extract the response associated with each of the 103 hexagons.

Figure 2A provides a nontechnical explanation of how it is possible to extract 103 responses from a single record. Were one to sum the first 60 milliseconds of the records following the frames during which a particular hexagon appeared white, the response would look like R in Figure 2A. Likewise, were one to sum the first 60 milliseconds of the records following the frames during which the same hexagon appeared black, the response would look something like NR in Figure 2A. Response R contains the responses to all the hexagons that flashed (appeared white), including the hexagon in question. Response NR, on the other hand, contains the responses to all the hexagons except the hexagon in question. The difference between R and NR is the response to the hexagon in question. These responses (Fig. 1E) are called first-order kernels. (For a more realistic picture of the waveforms underlying the first-order kernel, see Fig. 27 in reference 7.) Although the software could, in principle, calculate the 103 mfERG kernels or responses in this way, it does not. Technically, each mfERG response (Fig. 1E) is the result of a serial correlation between the stimulation sequence of a particular hexagon (Fig. 1C) and the single continuous ERG record (Fig. 1D). The pseudorandom sequence (an m-sequence) is chosen such that, when coupled with a special algorithm, the software can make these calculations very quickly (4). Because of the patent held by EDI on the m-sequence, other manufacturers use different methods of extracting the multifocal responses.

The first-order (A) and second-order (B) kernels. See text and appendix for details.

Displaying the Responses

Figure 1E shows the 103 mfERG responses for stimulation of the OD of a control subject. These responses are positioned so that they do not overlap and thus the scaling is arbitrary, as a comparison of the circles in Figures 1A and 1E indicates. We find the trace array in Figure 1E to be the most useful presentation of the data. In some cases, it is helpful to sum or average the responses within various regions of the display. Often responses are summed within rings around fixation. In Figure 3B, the responses from Figure 1E are grouped by rings around fixation (Fig. 3A) and summed. The responses become larger with increased eccentricity from the fovea because progressively larger areas of the retina are being summed. To take area into consideration, the amplitude of the summed response is divided by the total area of the hexagons in that ring. The resulting responses (Fig. 3C) are expressed in a measure of response amplitude per unit area or response density (nV/deg2). As expected, the response per unit area is highest in the fovea (5). Although this analysis by rings is useful for many purposes, it is not an appropriate display for summarizing the effects of retinal diseases that have nasotemporal asymmetries. It also obscures the nasotemporal differences that are present in the normal mfERG (7,13). However, the software available to analyze the mfERG allows for the combination of responses from any arbitrary grouping of hexagons (Figs. 8–10).

The multifocal electroretinogram (mfERG) in relation to retinal anatomy. A, The mfERG stimulus overlaid on a fundus photograph. The sectors are grouped by rings about the center and marked by arrows. B, The mfERG responses summed for each of the rings marked in (A). C, The same responses as in (B) but expressed in units of response density (nV/deg2). That is, the summed responses in (B) are divided by the total area of stimulation in the associated rings. The principal mfERG components N1, P1, and N2 are labeled for one of the responses.
Patient 3: Branch retinal artery occlusion. A, The visual field OD shows a defect in the superior temporal field (circled in light gray). A healthy, spatially equivalent area in the inferior temporal field is circled in black. B, The multifocal electroretinogram (mfERG) responses OD. C, The mfERG second-order (2K) response OD. D, The mfERG second-order (2K) 2D plot OD. The mfERG responses show reduced amplitudes in the region corresponding to the visual field defect.
Patient 4: Retina versus optic nerve/chiasm lesion. A, The visual fields show a nasal scotoma OD (solid oval). The dashed oval encircles an equivalent, but normally functioning, area in the temporal field. B, The OS (black) and OD (gray) multifocal electroretinogram (mfERG) responses. The responses for the OS are flipped about the vertical axis so that nasal field responses are presented on the left and temporal field responses on the right for both eyes. The ovals indicate spatially equivalent areas to those circumscribed in (A). The sum of the responses inside the encircled areas is shown in the inset. The mfERG responses show reduced amplitudes in the region corresponding to the visual field defect, indicating that the problem is retinal and not from optic nerves or chiasm (third ventricular tumor).
Patient 5: Branch retinal vein occlusion (BRVO) versus glaucoma. The OD 24-2 Humphrey visual fields. The dashed rectangle indicates an inferonasal scotoma OS. The black circle circumscribes the paramacular region OS. B, The OS (black) and OD (gray) multifocal electroretinogram (mfERG) responses. The responses for the OD are flipped about the vertical axis so that for both eyes, the nasal field responses are presented on the right and the temporal field responses on the left. The areas inside the dashed rectangle and black circle indicate spatially equivalent areas to those marked in (A). The sum of the responses for these areas is shown in the inset. The mfERG amplitudes are not reduced in the region corresponding to the visual field defect, indicating that the defect is due to glaucoma and not BRVO.

The mfERG results are often displayed in a 3D plot. Figure 1A (bottom) shows the results from Figure 1E as a 3D plot. To obtain this display, the response amplitude is divided by the area of the hexagon (the response density is obtained for each hexagon). In the control subject, there is a depression and a peak associated with the optic disc and the fovea, respectively. Although the 3D plot is pleasing to the eye, it can be very misleading (see below). The 3D plot should never be analyzed, or published, without the associated trace array.

Normative Values and Reproducibility

Several investigators have published normative values for the mfERG (14–17). As expected from full-field ERG studies, the amplitude decreases and the implicit time increases with decreasing luminance. However, as is the case for all electrophysiological tests, it is important that each clinic establish its own age-related norms. The intraindividual reproducibility of the mfERG is good (18). In fact, one group reports it is better than that of static automated perimetry (19).


As typically recorded, the mfERG signal originates from the central 25° (radius) of the retina (Fig. 1A). The display covers only about 25% of the cone photoreceptor cells (20). Furthermore, the high rate of stimulation, combined with the light levels used, assures that the rods do not contribute except under very unusual circumstances (21,22). Like the traditional photopic, or cone-driven, ERG, the mfERG shows an initial negative component (N1) followed by a positive component (P1) (Fig. 3C). These components bear a superficial resemblance to the a- and b-waves of the photopic flash ERG. However, the waveform of the mfERG differs from that of the typical photopic ERG (7,20). This is not surprising as both the stimulus and the analysis are different. Recall that the standard, full-field ERG is the sum of one or more responses to single flashes. In contrast, the mfERG is not a response at all, but rather a mathematical extraction (Fig. 2A). Thus, the components of the mfERG should never be referred to as the a-wave and b-wave (9).

In spite of these differences, it appears that N1 comprises the same components as the a-wave of the full-field ERG and P1 comprises the same components as the positive waves (b-wave and oscillatory potentials) (7,20). As in the case of the full-field ERG (23,24), the mfERG waveform is largely shaped by bipolar cell activity, together with small contributions from the photoreceptor cells and inner (amacrine and ganglion) retinal cells (25). Figure 4 displays the model of Hood et al. (25) explaining how the cells of the outer retina contribute to produce the mfERG waveform, shown as the solid black curve. The N1, P1, and N2 components are influenced in different ways by the onset and offset of the bipolar cells and, to a much lesser extent, of the photoreceptors. The inner retina exerts a subtle influence on the waveform. In the monkey, for example, the “ledge” on the trailing edge of P1 is removed by blocking the action potentials from the amacrine and/or ganglion cells (25–27).

The retinal origin of the multifocal electroretinogram signal. (Modified from ref. 25).

Analysis of the components of the mfERG (Fig. 4) reveals an important message. Damage at, or before, the bipolar cells will substantially decrease the amplitude of the mfERG. Inner retinal damage to amacrine and/or ganglion cells does not affect mfERG amplitude, although it may have a small effect on its waveform (26,28–31). As an example, Figure 5 shows the mfERG responses recorded 8 months after the occurrence of ischemic optic neuropathy. The responses have been summed within each quadrant for OU. Although there is extensive ganglion cell damage, as imputed from the Humphrey visual fields (HVF) (upper panels of Fig. 5), the mfERGs from the affected eye are similar to those from the unaffected eye.

Normal multifocal electroretinogram (mfERG) in ischemic optic neuropathy OS. Top: 24-2 Humphrey visual fields (total deviation probability plots) OU. Bottom: The OS (thick black) and OD (thin gray) mfERG responses, summed by quadrants, are identical and normal.


Over the past 4 years, we have routinely recorded mfERGs from patients evaluated by two neuro-ophthalmologists, Drs. Myles M. Behrens and Jeffrey G. Odel. The following examples illustrate our experience with the mfERG as a diagnostic tool.

Excluding Outer Retinal Disease

The neuro-ophthalmologist is routinely faced with deciding whether a visual defect is due to damage to the outer retina (before the ganglion cells) or damage to the ganglion cells and/or optic nerve. The mfERG can be very helpful, especially in situations where standard tests provide ambiguous information. Because damage to the ganglion cells or optic nerve does not decrease the amplitude of the mfERG, an abnormal mfERG provides strong evidence for an outer retinal lesion.

Patient 1 is a 16-year-old girl with a 1-year history of difficulty reading, especially with her OS. Her visual acuity was 20/25-2 OD and 20/60-1 OS. A full-field ERG was normal and the diagnosis of optic neuritis was entertained. However, the mfERG (Fig. 6B) clearly suggests an outer retinal lesion, particularly when compared with the visual fields. Iso-degree contours have been added to both the mfERG responses (Fig. 6B) and the 24-2 HVF (Fig. 6A) to aid in this comparison. More sophisticated procedures for comparing the HVF to the mfERG exist (32), but these contours are sufficient for most clinical purposes. The agreement between the depressed amplitude of the mfERG (Fig. 6B) and the regions of the HVF defects (Fig. 6A) confirms that the lesion lies in the outer retina. The mfERG, reduced in amplitude but relatively unchanged in implicit time, resembles that seen in Stargardt's disease (33).

Patient 1: Stargardt's disease. A. 24-2 Humphrey visual fields (total deviation probability plots). B. The mfERG responses show reduced central amplitudes, but relatively unchanged implicit times, typical of early Stargardt's disease (33). (The vertical and horizontal calibration bars indicate 100 nV and 60 milliseconds, respectively; the thick dark gray, thin black, and dashed light gray circles indicate radii of 5°, 15°, and 25°, respectively.)

Patient 2 is a 41-year-old man with a 4-month complaint that the vision in his OS resembled a “smudge” on his glasses. He had no complaints about the vision in his OD. His fundus appeared normal and his visual acuity was 20/20 in both eyes. The 24-2 HVFs (Fig. 7A) showed paracentral ring scotomas OU. Glaucoma and retinopathy were possible diagnoses. As shown in Figure 7B, the diagnosis is not glaucoma, as the mfERG is depressed in regions corresponding to the field defects. He was later shown to have abnormal antibody activity suggestive of melanoma-associated retinopathy.

Patient 2: Possible melanoma-associated retinopathy. A, Humphrey visual fields. B, The multifocal electroretinogram (mfERG) responses show reduced amplitudes in ring distribution paracentrally, corresponding to the visual field defects.

Patient 3 is a 58-year-old woman who had a 10-year history of episodic flashing in both eyes, most often in the OD, suspected of being migrainous in nature. Several weeks prior to her visit, the flashes seemed different and appeared only in the superior temporal field of the OD, in the region corresponding to the defect seen on her HVF (Fig. 8A). Her fundus appeared normal except for mild irregular narrowing of the branch retinal artery in the region of the defect. The subtle change in the amplitude of the mfERG responses in the region of the defect (Fig. 8B), combined with the narrowed artery, suggested branch retinal artery occlusion (BRAO), a condition known to affect both the inner and outer (bipolar cells) retina. Localized defects are sometimes easier to visualize in the 3D plot (Fig. 8D) and in the second-order kernel array (Fig. 8C) (34). (See the Appendix for a discussion of the secondorder kernel.)

The neuro-ophthalmologist is occasionally faced with a patient with two comorbid ophthalmologic conditions, each of which could, in principle, explain the patient's visual loss. Two examples are worth considering here as they illustrate additional points. Patient 4, a 43-year-old man, was a former Olympic soccer player whose MRI showed a third ventricular tumor adjacent to the anterior visual pathway. A nasal scotoma OD was found on visual field testing (Fig. 9A). Visual acuity was 20/20 OU and his fundus examination OD revealed a region of hyperpigmentation in the temporal retina. A mfERG was obtained to determine whether the tumor or a retinal problem was the cause of the defect. There was a local decrease in the mfERG OD in the region of the scotoma. This is best seen by comparing the responses from the two eyes (Fig. 9B). Since there are nasotemporal variations in the mfERG that can be quite large in some individuals, it is important to compare responses from corresponding regions of the retina. The responses from the OS (black) are left–right reversed in Figure 9B so that nasal and temporal regions are aligned. The mfERG is smaller in the region of the defect. Therefore, the retinal lesion rather than the tumor is the cause of the scotoma.

Patient 5 is a 61-year-old woman with a history of a OS branch retinal vein occlusion (BRVO) occurring 3 years prior to presenting for investigation of a visual field defect in her OS (Fig. 10A). Her visual acuity was slightly reduced in the OS (20/30) and intraocular pressure was normal at 16 mm Hg. There was probable disc excavation in the OS. The mfERG was normal in the region of the visual field defect (see dashed rectangles in Fig. 10). Thus, normal tension glaucoma is the likely cause of this defect. Notice, however, that there is a small decrease in the mfERG amplitudes in the macula corresponding to macular edema secondary to the old BRVO (see black circles in Fig. 10).

Differentiating Among Retinal Diseases

The mfERG can help to differentiate among retinal diseases based on changes in the relative amplitudes and latencies of N1 and P1 (See Table 1 in reference 7 for a review.) A large delay in the timing of the mfERG is associated with damage to the photoreceptors/outer plexiform layer (7). Damage to bipolar, amacrine, or ganglion cells yields relatively small changes in the implicit time of P1 and may even shorten it.

Patient 6, a 62-year-old man, had a long history of difficulty with reading, but only mildly reduced acuity (20/20 OD; 20/40 OS). His retinal evaluations, including angiograms, were normal. Visual fields showed bilateral central scotomas. The left panel of Figure 11A shows the 24-2 HVF of his OS. Because of the nature of his visual fields and a normal-appearing fundus, he underwent repeated and extensive neuroradiologic, neurosurgical, neurologic, and serological examinations, all normal. The mfERG (Fig. 11A, right panel) showed reduced amplitudes in the central 5°and delayed implicit times at all locations, including those where the Humphrey visual fields were normal. The delays can be seen more easily in Figure 11B, where the first 40 milliseconds of the responses are shown for this patient (right panel) and for a normal subject (left panel). Delays such as these are seen in cone dystrophy (35) and retinitis pigmentosa (36,37) and are associated with damage to the receptors/outer plexiform layer (7). The findings for this patient fit the definition of occult macular dystrophy proposed by Miyake and colleagues (38). Such patients present with normal to moderately reduced visual acuity, normal peripheral fields, small central or paracentral defects, and normal, or only mildly abnormal, fullfield ERGs. They are best diagnosed with focal ERGs or mfERGs.

Patient 6: Occult macular dystrophy. A, The OS visual field and multifocal electroretinogram (mfERG) responses. B, The first 40 milliseconds of the OS mfERG responses for a normal subject and the patient. The patient's mfERG responses are only modestly reduced in amplitude, but they are markedly delayed in timing.

Following Disease Progression

Because the mfERG shows good repeat reliability (18,19), it can be used to follow the progression of a disease. This is particularly helpful in the case of patients who are poor field takers, but it can be of value in other cases as well.

Patient 7 provides an example where following the mfERG changed the diagnosis. The mfERGs in Figure 12B are delayed in all regions of the field except the central region, as is often seen with diseases of the receptors and pigment epithelium such as retinitis pigmentosa (RP) and cone dystrophy (7,35–37). Although we initially thought that the patient's problem was RP, her visual field continued to deteriorate at a rate far too fast for RP. Figure 12C and D show her HVF and mfERG responses obtained 15 months later. The field has become more constricted and the mfERG amplitudes have markedly decreased in the region previously showing the delays. We subsequently found evidence of retinal antibodies, and an autoimmune process is suspected.

Patient 7: Possible immune-mediated outer retinopathy. The baseline OS visual field (A) and multifocal electroretinogram (mfERG) (B). Recorded 15 months later, the OS visual field (C) and mfERG (D) show substantial progression. Arrows point to geographically equivalent areas.

Differentiating Organic from Nonorganic Disorders

The mfERG can be used to diagnose nonorganic disorders. The advantage of the mfERG over the conventional ERG is that it provides a topographical representation that can be compared with the patient's visual fields. Patient 8 is a 30-year-old policeman who had complained of decreased vision in his OS for the past few years, especially during the day. A nonorganic disorder was in the differential diagnosis. However, the mfERG shows depressed responses in the macula (Fig. 13, right panel). This prompted indocyanine green angiography and optical coherence tomography, which suggested a resolved central serous chorioretinopathy. A normal mfERG does not, by itself, establish a visual deficit as nonorganic. If the mfERG is normal, then a multifocal VEP should be performed as well to rule out damage to the optic nerve/ganglion cells.

Patient 8: Resolved central serous retinopathy in suspected nonorganic visual loss. Visual field (left) and multifocal electroretinogram (mfERG) (right). Reduced foveal mfERG amplitudes suggested that the depressed foveal thresholds had a retinal origin, which prompted indocyanine green angiography and optical coherence tomography that suggested that resolved central serous retinopathy is the cause of visual loss.


Eccentric Fixation

As in the case of visual field testing, it is important to monitor the patient's eye position. An eye camera or direct visualization of the eye should be used to assure that fixation is steady. Unsteady fixation can cause diminished responses in the center of the field. Monitoring the eye, however, will not assure that the fixation is accurate. Some patients with central visual problems can have eccentric fixation, which will produce mfERGs that appear to have central and paracentral defects. Figure 14, which illustrates this point, contains the mfERG from a normal individual who was instructed to fixate down and to the left by 8.5° from the center (Fig 14C). Notice the smaller responses in the central part of the field. These are due to the fact that with eccentric fixation, the small central hexagons are now falling outside the fovea. Eccentric fixation can be detected by the pattern of results in both the trace array (Fig. 14C) and 3D plot (Fig. 14E). Since the fovea is now stimulated by larger hexagons, the responses in the region of fixation will be abnormally large. In the 3D plot, the blind spot also has been shifted (Fig. 14E). [The latencies of the mfERG responses can also be used to identify the foveal and blind spot regions (see ref. 13 and Fig. 13 in ref. 7)]. The mfERG from Patient 9 provides a clinical illustration of eccentric fixation. Patient 9 is a 64-year-old woman with psoriatic arthritis who had been treated with both hydroxychloroquine and infliximab. She complained of a loss of vision in her OD, and her visual field demonstrated a marked generalized depression (Fig. 14B). Her mfERG showed a decrease in response amplitude in the central 5° or so (Fig. 14D). However, her field depression extended at least to 25° (Fig. 14B). Before one can conclude that the retina is the site of the damage, eccentric fixation must be ruled out. The patient's visual acuity was 20/400 OD and she could not see the fixation target. Her fixation was monitored during the test and it was steady. However, her mfERG (Fig. 14D, F) indicates that she is clearly fixating off center. The pattern of her mfERG resembles that in Fig. 14C, E. The location of the foveal peak and optic disc depression are both consistent with fixation up and to the left of the fixation target. Thus, her retinal function appears grossly normal and the damage must be at or beyond the ganglion cells. The message is clear: if care is not taken in the recording and interpretation of mfERGs, depressed central responses due to fixation errors can be misinterpreted as a retinal problem.

The problem of eccentric fixation. A, Visual field for a normal subject. B, Visual field for Patient 9. C, The multifocal electroretinogram (mfERG) for a normal subject instructed to fixate at a target 8.5° down and to the left of center. D, The mfERG for Patient 9. The black circle indicates an area of apparently decreased mfERG response due to a fixation error. E, The 3D plot for the mfERG in (C). F, The 3D plot for the mfERG of Patient 9. The reduced amplitudes in the patient's mfERG are likely due to eccentric fixation and not to retinal damage.

The 3D Plot Can Be Misleading

The 3D plot can be misleading and should never be presented without the associated trace array. Consider the records in Figure 15A. These responses were obtained from an electrode in saline. There are no responses here, only noise. However, since the 3D plot (Fig. 15B) is obtained by dividing the response (which, in this case, is the same everywhere) by the area of the local hexagon (which is smaller for the central hexagons), a “foveal peak” is present. Figure 15C–F provides a clinical example of where the 3D plot was misleading. Patient 10, a 66-year-old woman with multifocal choroiditis and panuveitis, had a visual acuity of counting fingers in her OD. Her OD mfERG was performed in June 2000 (Fig. 15C, E) and again 18 months later (Fig. 15D, F) to see if any recovery had taken place. The 3D plot in Figure 15D suggested an improvement. However, as her trace arrays and ring averages (Figs. 15E–H) indicate, there is little or no response on either day. The noisier records on the second day produced the artifactual peak at the fovea. Although 3D plots can be useful for visualizing small scotomas (Fig. 8B) and the blind spot (Fig. 14E, F), they should be used sparingly, and always in conjunction with the trace arrays.

The misleading 3D plot. A, The multifocal electroretinogram (mfERG) responses for a Burian-Allen electrode placed in a jar of saline. B, The 3D plot for the mfERG responses in (A). C, The mfERG 3D plot for patient 10. D, The mfERG 3D plot for patient 10 tested 18 months later. E, The mfERG responses corresponding to the 3D plot in (C). F, The mfERG responses corresponding to the 3D plot in (D). G, The mfERG responses, expressed as response density, for the area within 5° (thick dark gray), between 5° and 15° (thin black), and between 15° and 25° (dashed light gray) of (E). H, Same as (G) for responses in (F). Thick dark gray, thin black, and dashed light gray circles indicate radii 5°, 15°, and 25°, respectively. The 3D plot can give the appearance that there are responses in the central retina when in fact none are present.

The Spatial Resolution of the mfERG

Since the mfERG is a topographical map of retinal function, it is important to know its spatial resolution. Spatial resolution will be influenced by a variety of factors, including the degree of light scatter in the eye, the size of the hexagons in the test display, and the stability of fixation. A systematic consideration of these factors has yet to be published. In our experience, scotomas at least as small as 4° can be detected (see Figs. 8 and 9 and Fig. 7 in ref. 7).


The mfERG has become a useful test to detect regional outer retinal disorders that are not sufficiently extensive to significantly reduce the full-field ERG. However, to optimize its value, it is important to compare mfERG results to those of automated visual fields obtained concurrently. Seeing abnormal mfERGs in the same regions of the visual field that are abnormal on automatic perimetry provides a high degree of reassurance of the retinal origin of the defect.

Recording mfERGs requires skill and experience. The challenges in recording and analyzing mfERG responses are greater than those involved in full-field ERG testing. Even so, those who are already recording high-quality ERGs should be able to master the skill. mfERG testing is otherwise best left to centers with an electrophysiologist familiar with the technique.


The authors gratefully acknowledge the support and encouragement of Myles M. Behrens, MD.


The Second-Order Kernel

It is important to have some understanding of the second-order kernel or response (2K), because some investigators have claimed diseases of the ganglion cell/optic nerve affect it differentially. Like all the multifocal responses, it is not technically a “response” but a mathematical extraction. Figure 2B provides a simple way to understand the meaning of the 2K. The flashes in Figure 2A that make up the first-order response (1K) can be divided into two groups. Half of the times that a particular hexagon appeared white, a flash preceded it (small white hexagon in Fig. 2B). On the other half of the times, a flash did not precede it (small black hexagon in Fig. 2B). If the responses under the two conditions are the same, then there is no 2K. If these two responses differ, then there is a 2K and it is the difference between the 1K responses. Thus, the presence of a 2K indicates that there is an effect of short-term adaptation. In the normal mfERG, the presence of the flash on the preceding frame makes the response slightly smaller and slightly faster. Thus, the shape of the 2K is complex as indicated in Figure 2B.

Some investigators speak as if the 2K is an actual response generated in the inner retina. Some even claim that it is generated by the ganglion cells. As the discussion above indicates, the 2K is not a response and thus strictly speaking cannot be generated anywhere. We do know that blocking action potentials generated by the ganglion cells and amacrine cells in monkeys markedly reduces, but does not eliminate, the 2K (25). On the other hand, although ganglion cell damage can reduce the 2K in humans (28), a large 2K can be present even with extensive damage (29). Therefore, it appears that inner retinal damage, but not necessarily ganglion cell damage, can decrease the 2K in humans. However, outer plexiform damage can completely eliminate the 2K in patients with degenerative diseases of the receptors (7). Consequently, it is a mistake to associate a diminished 2K with damage to a particular set of cells. A diminished 2K indicates an abnormality in the circuits and connections involved in adaptation rather than a missing component or cellular response (7).

The 2K can be useful, however, in identifying local lesions of the inner nuclear layer and/or receptors. As mentioned above, the 2K is reduced more than the 1K by BRAO (34), as can be seen in Figure 8. Although the selective loss of the 2K has been interpreted as an indication of inner retinal damage (34), it can be caused by damage at the outer plexiform layer as well (7). Furthermore, it is likely that the relatively larger loss of the 2K response as compared with the 1K in BRAO is due to a reduced effect of stray light (7,39). For example, Shimada and Horiguchi (39) have shown that the spatial resolution of the 2K is better than the 1K because the 2K has a smaller contribution from stray light.

Other Paradigms

The software available for creating displays and for modulating the temporal sequence of light presentations allows for a wide range of spatial and temporal paradigms. A number of paradigms have been developed to help detect damage to the inner retina (i.e., amacrine and ganglion cells) (e.g., [40–45]). The best developed of these is the global flash paradigm (40). This paradigm is designed to accentuate a component generated at the optic nerve head (ONH) by action potentials from ganglion cells. The existence of an ONH component has been fairly well established (27,28), and there is some evidence that glaucoma can eliminate it (40). However, the ONH component is small and its usefulness in detecting glaucomatous damage is uncertain (45). Until more evidence is presented, we do not recommend using the mfERG to study diseases of the ganglion cell/optic nerve. If an electrophysiological measure of ganglion cell/optic nerve is needed, the multifocal VEP is a better candidate (46).


1. Fishman GA. The electroretinogram. In: Fishman GA, Birch DG, Holder GE, et al., eds. Electrophysiologic Testing in Disorders of the Retina, Optic Nerve, and Visual Pathway, Second Edition (Ophthalmology Monographs). San Francisco: The Foundation of the American Academy of Ophthalmology; 2001:1–155.
2. Holder GE. The pattern electroretinogram. In: Fishman GA, Birch DG, Holder GE, et al., eds. Electrophysiologic Testing in Disorders of the Retina, Optic Nerve, and Visual Pathway, Second Edition (Ophthalmology Monographs). San Francisco: The Foundation of the American Academy of Ophthalmology; 2001:192–236.
3. Birch DG. The focal and multifocal electroretinogram. In: Fishman GA, Birch DG, Holder GE, et al., eds. Electrophysiologic Testing in Disorders of the Retina, Optic Nerve, and Visual Pathway, Second Edition (Ophthalmology Monographs). San Francisco: The Foundation of the American Academy of Ophthalmology, 2001:177–191.
4. Sutter EE. The fast m-transform: a fast computation of cross-correlations with binary m-sequences. Soc Ind Appl Math. 1991; 20:686–694.
5. Sutter EE, Tran D. The field topography of ERG components in man—I. The photopic luminance response. Vision Res. 1992; 32:433–446.
6. Bearse MA, Sutter EE. Imaging localized retinal dysfunction with the multifocal electroretinogram. J Opt Soc Am A. 1996; 13:634–640.
7. Hood DC. Assessing retinal function with the multifocal technique. Prog Ret Eye Res. 2000; 19:607–646.
8. Sutter EE. Imaging visual function with the multifocal m-sequence technique. Vision Res. 2001; 41:1241–1255.
9. Marmor MF, Hood D, Keating D, et al. Guidelines for basic multifocal electroretinography (mfERG). Doc Ophthalmol. 2003; 106;105–115.
10. Dawson WW, Trick GL, Litzkow CA. Improved electrode for electroretinography. Invest Ophthalmol Vis Sci. 1979; 18:988–991.
11. Hood DC, Li L. A technique for measuring individual multifocal ERG records. In: Yager D, ed. Non-invasive Assessment of the Visual System. Trends in Optics and Photonics. Washington D.C.: Optical Society of America; 1997;11:33–41.
12. Hood DC, Holopigian K, Seiple W, et al. Assessment of local retinal function in patients with retinitis pigmentosa using the multifocal ERG technique. Vision Res. 1998; 38:163–179.
13. Seeliger MW, Kretschmann UH, Apfelstedt-Sylla E, et al. Implicit time topography of multifocal electroretinograms. Invest Ophthal Vis Sci. 1998; 39:718–723.
14. Verdon WA, Haegerstrom-Portnoy G. Topography of the multifocal electroretinogram. Doc Ophthalmol. 1998; 95:73–90.
15. Nabeshima T, Tazawa Y, Mita M, et al. Effects of aging on the first and second-order kernels of multifocal electroretinogram. Jpn J Ophthalmol. 2002; 46:261–269.
16. Jackson GR, Ortega J, Girkin C, et al. Aging-related changes in the multifocal electroretinogram. J Opt Soc Am. 2002; 19:185–189.
17. Gerth C, Garcia SM, Ma L, et al. Multifocal electroretinogram: age-related changes for different luminance. Graefes Arch Clin Eye Ophthalmol. 2002; 240:202–208.
18. Heinemann-Vernaleken B, Palmowski A, Allgayer R. The effect of time of day and repeat reliability on the fast flicker multifocal ERG. Doc Ophthalmol. 2000; 101:247–255.
19. Clemens C, Kirzhner M, Holopigian K, et al. Test-retest reliability of psychophysical and electrophysiological perimetric measures in patients with retinitis pigmentosa. ARVO. 2002; abstract nr. 1169.
20. Hood DC, Seiple W, Holopigian K, et al. A comparison of the components of the multifocal and full-field ERGs. Vis Neurosci. 1997; 14:533–544.
21. Wu S, Sutter EE. A topographic study of oscillatory potentials in man. Vis Neurosci. 1995; 12:1013–1025.
22. Hood DC, Wladis EJ, Shady S, et al. Multifocal rod electroretinograms Invest Ophthalmol Vis Sci. 1998; 39:1152–1162.
23. Bush RA, Sieving P. A proximal retinal component in the primate photopic ERG a-wave. Invest Ophthal Vis Sci. 1994; 35:635–645.
24. Sieving PA, Murayama K, Naarendorp F. Push-pull model of the primate photopic electroretinogram: a role for hyperpolarizing neurons in shaping the b-wave. Vis Neurosci. 1994; 11:519–532.
25. Hood DC, Frishman LJ, Saszik S, et al. Retinal origins of the primate multifocal ERG: implications for the human response. Invest Ophthalmol Vis Sci. 2002; 43:1673–1685.
26. Hood DC, Greenstein V, Frishman LJ, et al. Identifying inner retinal contributions to the human multifocal ERG. Vision Res. 1999; 39:2285–2291.
27. Hood DC, Bearse MA, Sutter EE, et al. The optic nerve head component of the monkey's (Macaca mulatta) multifocal electroretinogram (mERG). Vision Res. 2001; 41:2029–2041.
28. Sutter EE, Bearse MA. The optic nerve head component of the human ERG. Vision Res. 1999; 39:419–436.
29. Hood DC, Greenstein VC, Holopigian K, et al. An attempt to detect glaucomatous damage to the inner retina with the multifocal ERG. Invest Ophthalmol Vis Sci. 2000; 41:1570–1579.
30. Fortune B, Johnson CA, Cioffi GA. The topographic relationship between multifocal electroretinographic and behavioral perimetric measures of function in glaucoma. Optom Vis Sci. 2001; 78:206–214.
31. Hasegawa S, Takagi M, Usui T, et al. Waveform changes of the first-order multifocal electroretinogram in patients with glaucoma. Invest Ophthalmol Vis Sci. 2000; 41:1597–603.
32. Hood DC, Zhang X. Multifocal ERG and VEP responses and visual fields: comparing disease-related changes. Doc Ophthal. 2000; 100:115–137.
33. Kretschmann U, Seeliger MW, Ruether K, et al. Multifocal electroretinography in patients with Stargardt's macular dystrophy. Br J Ophthalmol. 1998; 82:267–275.
34. Hasegawa S, Ohshima A, Hayakawa Y, et al. Multifocal electroretinograms in patients with branch retinal artery occlusion. Invest Ophthalmol Vis Sci. 2001; 42:298–304.
35. Holopigian K, Seiple W, Greenstein VC, et al. Local cone and rod system function in progressive cone dystrophy. Invest Ophthalmol Vis Sci. 2002; 43:2364–2373.
36. Hood DC, Holopigian K, Seiple W, et al. Assessment of local retinal function in patients with retinitis pigmentosa using the multifocal ERG technique. Vision Res. 1998; 38:163–179.
37. Seeliger M, Kretschmann U, Apfelstedt-Sylla E, et al. Multifocal electroretinography in retinitis pigmentosa. Am J Ophthalmol. 1998; 125:214–226.
38. Miyake Y, Horiguchi M, Tomita N, et al. Occult macular dystrophy. Am J Ophthalmol. 1996; 122:644–653.
39. Shimada Y, Horiguchi M. Stray light-induced multifocal electroretinograms. Invest Ophthalmol Vis Sci. 2003; 44;1245–1251.
40. Sutter EE, Bearse MA. The retinal topography of local and lateral gain control mechanisms. In:Vision Science and Its Applications, 1998 OSA Technical Digest Series. Washington D.C.: Optical Society of America; 1998:20–23.
41. Sutter EE, Shimada Y, Li Y, et al. Mapping inner retinal function through enhancement of adaptive components in the M-ERG. In:Vision Science and Its Applications, 1999 OSA Technical Digest Series. Washington D.C.: Optical Society of America; 1999;1:52–55.
42. Palmowski AM, Sutter EE, Bearse Jr., MA et al. Mapping of retinal function in diabetic retinopathy using the multifocal electroretinogram. Invest Ophthalmol Vis Sci. 1997; 38:2586–2596.
43. Bearse Jr., MA Shimada Y, Sutter EE. Distribution of oscillatory components in the central retina. Doc Ophthalmol. 2000; 100:185–205.
44. Palmowski AM, Allgayer R, Heinemann-Vernaleken B, et al. Multifocal electroretinogram with a multiflash stimulation technique in open-angle glaucoma. Ophthalmic Res. 2002; 34:83–89.
45. Fortune B, Bearse MA, Cioffi GA, et al. Selective loss of an oscillatory component from temporal retinal multifocal ERG responses. Invest Ophthalmol Vis Sci. 2002; 43:2638–2647.
46. Hood DC, Greenstein VC. The multifocal VEP and ganglion cell damage: applications and limitations for the study of glaucoma. Prog Ret Eye Res. 2002; 22:201–251.
© 2003 Lippincott Williams & Wilkins, Inc.